(tf_gpu112) urbanjo3@dgx-1:/dev/shm/mirek2$ time python premsel_network_enigma-01-2020-my1-no_symbols_tst3_10.py (tf_gpu112) urbanjo3@dgx-1:/dev/shm/mirek2$ [B  time python premsel_network_enigma-01-2020-my1-no_symbols_tst3_10.py (tf_gpu112) urbanjo3@dgx-1:/dev/shm/mirek2$ time python premsel_network_enigma-01-2020-my1-no_symbols_tst3_10.py (tf_gpu112) urbanjo3@dgx-1:/dev/shm/mirek2$ time python premsel_network_enigma-01-2020-my1-no_symbols_tst3_10.py Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_15.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-solo___out1.pkl.gz Training 0: 0 / 114355: Premsel loss 1.0000, acc 0.0000 (0.0000 / 0.0000) Training 0: 7000 / 114355: Premsel loss 0.5833, acc 0.6815 (0.6982 / 0.6649) Training 0: 14000 / 114355: Premsel loss 0.4722, acc 0.7566 (0.7691 / 0.7441) Training 0: 21000 / 114355: Premsel loss 0.5067, acc 0.7533 (0.8016 / 0.7050) Training 0: 28000 / 114355: Premsel loss 0.4813, acc 0.7522 (0.7656 / 0.7387) Training 0: 35000 / 114355: Premsel loss 0.5557, acc 0.7800 (0.8088 / 0.7513) Training 0: 42000 / 114355: Premsel loss 0.4334, acc 0.7865 (0.8066 / 0.7664) Training 0: 49000 / 114355: Premsel loss 0.3980, acc 0.8068 (0.8290 / 0.7847) Training 0: 56000 / 114355: Premsel loss 0.4327, acc 0.7840 (0.8039 / 0.7641) Training 0: 63000 / 114355: Premsel loss 0.3762, acc 0.8190 (0.8278 / 0.8102) Training 0: 70000 / 114355: Premsel loss 0.3920, acc 0.8116 (0.8368 / 0.7863) Training 0: 77000 / 114355: Premsel loss 0.3668, acc 0.8255 (0.8407 / 0.8103) Training 0: 84000 / 114355: Premsel loss 0.4094, acc 0.8011 (0.8191 / 0.7830) Training 0: 91000 / 114355: Premsel loss 0.3963, acc 0.8073 (0.8216 / 0.7930) Training 0: 98000 / 114355: Premsel loss 0.4034, acc 0.8064 (0.8203 / 0.7924) Training 0: 105000 / 114355: Premsel loss 0.3621, acc 0.8275 (0.8434 / 0.8117) Training 0: 112000 / 114355: Premsel loss 0.3422, acc 0.8394 (0.8620 / 0.8168) Evaluation 0: Premsel loss 0.4231, acc 0.7980 (0.7854 / 0.8106) Training 1: 0 / 114355: Premsel loss 0.4085, acc 0.8105 (0.8476 / 0.7733) Training 1: 7000 / 114355: Premsel loss 0.3472, acc 0.8353 (0.8565 / 0.8142) Training 1: 14000 / 114355: Premsel loss 0.3449, acc 0.8373 (0.8496 / 0.8250) Training 1: 21000 / 114355: Premsel loss 0.3377, acc 0.8420 (0.8541 / 0.8298) Training 1: 28000 / 114355: Premsel loss 0.3619, acc 0.8270 (0.8539 / 0.8001) Training 1: 35000 / 114355: Premsel loss 0.3476, acc 0.8367 (0.8577 / 0.8157) Training 1: 42000 / 114355: Premsel loss 0.3359, acc 0.8434 (0.8665 / 0.8202) Training 1: 49000 / 114355: Premsel loss 0.3260, acc 0.8490 (0.8667 / 0.8313) Training 1: 56000 / 114355: Premsel loss 0.3373, acc 0.8441 (0.8593 / 0.8289) Training 1: 63000 / 114355: Premsel loss 0.5933, acc 0.7692 (0.7872 / 0.7511) Training 1: 70000 / 114355: Premsel loss 0.4472, acc 0.7796 (0.8119 / 0.7473) Training 1: 77000 / 114355: Premsel loss 0.3831, acc 0.8180 (0.8299 / 0.8061) Training 1: 84000 / 114355: Premsel loss 0.3658, acc 0.8281 (0.8491 / 0.8071) Training 1: 91000 / 114355: Premsel loss 0.3511, acc 0.8347 (0.8559 / 0.8135) Training 1: 98000 / 114355: Premsel loss 0.3378, acc 0.8420 (0.8600 / 0.8241) Training 1: 105000 / 114355: Premsel loss 0.3341, acc 0.8432 (0.8575 / 0.8289) Training 1: 112000 / 114355: Premsel loss 0.3223, acc 0.8499 (0.8723 / 0.8275) Evaluation 1: Premsel loss 0.3250, acc 0.8475 (0.8160 / 0.8791) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_03.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz Training 2: 0 / 96191: Premsel loss 0.3192, acc 0.8532 (0.8804 / 0.8259) Training 2: 7000 / 96191: Premsel loss 0.4419, acc 0.7908 (0.8086 / 0.7731) Training 2: 14000 / 96191: Premsel loss 0.4666, acc 0.7718 (0.7829 / 0.7607) Training 2: 21000 / 96191: Premsel loss 0.3717, acc 0.8246 (0.8311 / 0.8181) Training 2: 28000 / 96191: Premsel loss 0.3559, acc 0.8336 (0.8442 / 0.8231) Training 2: 35000 / 96191: Premsel loss 0.3905, acc 0.8179 (0.8403 / 0.7955) Training 2: 42000 / 96191: Premsel loss 0.3600, acc 0.8301 (0.8415 / 0.8186) Training 2: 49000 / 96191: Premsel loss 0.3407, acc 0.8407 (0.8587 / 0.8228) Training 2: 56000 / 96191: Premsel loss 0.3448, acc 0.8405 (0.8591 / 0.8219) Training 2: 63000 / 96191: Premsel loss 0.3324, acc 0.8453 (0.8628 / 0.8277) Training 2: 70000 / 96191: Premsel loss 0.3224, acc 0.8520 (0.8674 / 0.8367) Training 2: 77000 / 96191: Premsel loss 0.3278, acc 0.8470 (0.8726 / 0.8214) Training 2: 84000 / 96191: Premsel loss 0.3102, acc 0.8583 (0.8763 / 0.8403) Training 2: 91000 / 96191: Premsel loss 0.3060, acc 0.8603 (0.8876 / 0.8330) Evaluation 2: Premsel loss 0.3063, acc 0.8603 (0.8639 / 0.8566) Training 3: 0 / 96191: Premsel loss 0.3195, acc 0.8536 (0.8767 / 0.8305) Training 3: 7000 / 96191: Premsel loss 0.3112, acc 0.8582 (0.8821 / 0.8343) Training 3: 14000 / 96191: Premsel loss 0.3039, acc 0.8615 (0.8832 / 0.8397) Training 3: 21000 / 96191: Premsel loss 0.3034, acc 0.8624 (0.8866 / 0.8382) Training 3: 28000 / 96191: Premsel loss 0.3080, acc 0.8606 (0.8811 / 0.8400) Training 3: 35000 / 96191: Premsel loss 0.2961, acc 0.8663 (0.8897 / 0.8430) Training 3: 42000 / 96191: Premsel loss 0.3018, acc 0.8627 (0.8930 / 0.8324) Training 3: 49000 / 96191: Premsel loss 0.2967, acc 0.8656 (0.8830 / 0.8481) Training 3: 56000 / 96191: Premsel loss 0.2988, acc 0.8638 (0.8881 / 0.8394) Training 3: 63000 / 96191: Premsel loss 0.2987, acc 0.8647 (0.8903 / 0.8391) Training 3: 70000 / 96191: Premsel loss 0.3028, acc 0.8623 (0.8870 / 0.8377) Training 3: 77000 / 96191: Premsel loss 0.2819, acc 0.8741 (0.8972 / 0.8511) Training 3: 84000 / 96191: Premsel loss 0.2839, acc 0.8730 (0.8958 / 0.8503) Training 3: 91000 / 96191: Premsel loss 0.2966, acc 0.8666 (0.8844 / 0.8487) Evaluation 3: Premsel loss 0.2875, acc 0.8709 (0.8620 / 0.8797) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_07.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l4800-e0.15+coop-mzr02___out1.pkl.gz Training 4: 0 / 104621: Premsel loss 0.3034, acc 0.8614 (0.8877 / 0.8352) Training 4: 7000 / 104621: Premsel loss 0.2889, acc 0.8700 (0.9022 / 0.8377) Training 4: 14000 / 104621: Premsel loss 0.2887, acc 0.8699 (0.8985 / 0.8412) Training 4: 21000 / 104621: Premsel loss 0.2946, acc 0.8669 (0.9016 / 0.8323) Training 4: 28000 / 104621: Premsel loss 0.2901, acc 0.8696 (0.8923 / 0.8469) Training 4: 35000 / 104621: Premsel loss 0.2956, acc 0.8667 (0.9020 / 0.8314) Training 4: 42000 / 104621: Premsel loss 0.2892, acc 0.8686 (0.8934 / 0.8437) Training 4: 49000 / 104621: Premsel loss 0.2848, acc 0.8731 (0.8993 / 0.8468) Training 4: 56000 / 104621: Premsel loss 0.2867, acc 0.8716 (0.8972 / 0.8459) Training 4: 63000 / 104621: Premsel loss 0.2864, acc 0.8729 (0.9013 / 0.8445) Training 4: 70000 / 104621: Premsel loss 0.2794, acc 0.8751 (0.9018 / 0.8484) Training 4: 77000 / 104621: Premsel loss 0.2818, acc 0.8733 (0.8979 / 0.8487) Training 4: 84000 / 104621: Premsel loss 0.2771, acc 0.8773 (0.9051 / 0.8495) Training 4: 91000 / 104621: Premsel loss 0.2726, acc 0.8786 (0.8985 / 0.8588) Training 4: 98000 / 104621: Premsel loss 0.2802, acc 0.8746 (0.9015 / 0.8477) Evaluation 4: Premsel loss 0.2719, acc 0.8796 (0.9235 / 0.8357) Training 5: 0 / 104621: Premsel loss 0.2778, acc 0.8757 (0.9010 / 0.8503) Training 5: 7000 / 104621: Premsel loss 0.2686, acc 0.8814 (0.9124 / 0.8504) Training 5: 14000 / 104621: Premsel loss 0.2779, acc 0.8771 (0.9024 / 0.8517) Training 5: 21000 / 104621: Premsel loss 0.2717, acc 0.8787 (0.9012 / 0.8561) Training 5: 28000 / 104621: Premsel loss 0.2688, acc 0.8804 (0.9089 / 0.8520) Training 5: 35000 / 104621: Premsel loss 0.2630, acc 0.8840 (0.9058 / 0.8623) Training 5: 42000 / 104621: Premsel loss 0.2637, acc 0.8843 (0.9053 / 0.8633) Training 5: 49000 / 104621: Premsel loss 0.2717, acc 0.8801 (0.9072 / 0.8531) Training 5: 56000 / 104621: Premsel loss 0.2729, acc 0.8780 (0.9091 / 0.8470) Training 5: 63000 / 104621: Premsel loss 0.2731, acc 0.8793 (0.9023 / 0.8563) Training 5: 70000 / 104621: Premsel loss 0.2649, acc 0.8832 (0.9079 / 0.8585) Training 5: 77000 / 104621: Premsel loss 0.2587, acc 0.8868 (0.9118 / 0.8617) Training 5: 84000 / 104621: Premsel loss 0.2678, acc 0.8811 (0.9096 / 0.8526) Training 5: 91000 / 104621: Premsel loss 0.2634, acc 0.8843 (0.9089 / 0.8598) Training 5: 98000 / 104621: Premsel loss 0.3590, acc 0.8362 (0.8577 / 0.8147) Evaluation 5: Premsel loss 0.4440, acc 0.8604 (0.8919 / 0.8288) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_08.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_19.pkl.gz Training 6: 0 / 121616: Premsel loss 0.3131, acc 0.8592 (0.8823 / 0.8361) Training 6: 7000 / 121616: Premsel loss 0.2894, acc 0.8710 (0.8966 / 0.8454) Training 6: 14000 / 121616: Premsel loss 0.3879, acc 0.8156 (0.8413 / 0.7900) Training 6: 21000 / 121616: Premsel loss 0.6516, acc 0.7501 (0.7653 / 0.7350) Training 6: 28000 / 121616: Premsel loss 0.3909, acc 0.8110 (0.8342 / 0.7878) Training 6: 35000 / 121616: Premsel loss 0.4355, acc 0.7911 (0.8153 / 0.7668) Training 6: 42000 / 121616: Premsel loss 0.3753, acc 0.8229 (0.8381 / 0.8076) Training 6: 49000 / 121616: Premsel loss 0.3494, acc 0.8369 (0.8564 / 0.8174) Training 6: 56000 / 121616: Premsel loss 0.3363, acc 0.8447 (0.8717 / 0.8177) Training 6: 63000 / 121616: Premsel loss 0.3341, acc 0.8459 (0.8707 / 0.8211) Training 6: 70000 / 121616: Premsel loss 0.3310, acc 0.8471 (0.8743 / 0.8199) Training 6: 77000 / 121616: Premsel loss 0.3177, acc 0.8538 (0.8771 / 0.8305) Training 6: 84000 / 121616: Premsel loss 0.3196, acc 0.8536 (0.8760 / 0.8312) Training 6: 91000 / 121616: Premsel loss 0.3126, acc 0.8578 (0.8813 / 0.8343) Training 6: 98000 / 121616: Premsel loss 0.3089, acc 0.8603 (0.8928 / 0.8279) Training 6: 105000 / 121616: Premsel loss 0.3058, acc 0.8622 (0.8865 / 0.8380) Training 6: 112000 / 121616: Premsel loss 0.3011, acc 0.8647 (0.8913 / 0.8380) Training 6: 119000 / 121616: Premsel loss 0.2990, acc 0.8653 (0.8901 / 0.8405) Evaluation 6: Premsel loss 0.2934, acc 0.8685 (0.9095 / 0.8274) Training 7: 0 / 121616: Premsel loss 0.2941, acc 0.8682 (0.8934 / 0.8429) Training 7: 7000 / 121616: Premsel loss 0.2953, acc 0.8659 (0.8955 / 0.8363) Training 7: 14000 / 121616: Premsel loss 0.2943, acc 0.8680 (0.8947 / 0.8412) Training 7: 21000 / 121616: Premsel loss 0.3033, acc 0.8625 (0.8827 / 0.8422) Training 7: 28000 / 121616: Premsel loss 0.2953, acc 0.8666 (0.8881 / 0.8451) Training 7: 35000 / 121616: Premsel loss 0.2890, acc 0.8706 (0.9025 / 0.8386) Training 7: 42000 / 121616: Premsel loss 0.2801, acc 0.8756 (0.9068 / 0.8444) Training 7: 49000 / 121616: Premsel loss 0.2774, acc 0.8775 (0.9078 / 0.8471) Training 7: 56000 / 121616: Premsel loss 0.2800, acc 0.8750 (0.9008 / 0.8492) Training 7: 63000 / 121616: Premsel loss 0.2866, acc 0.8722 (0.8989 / 0.8455) Training 7: 70000 / 121616: Premsel loss 0.2832, acc 0.8750 (0.9095 / 0.8404) Training 7: 77000 / 121616: Premsel loss 0.2752, acc 0.8788 (0.9128 / 0.8448) Training 7: 84000 / 121616: Premsel loss 0.2680, acc 0.8824 (0.9110 / 0.8539) Training 7: 91000 / 121616: Premsel loss 0.2766, acc 0.8781 (0.9099 / 0.8463) Training 7: 98000 / 121616: Premsel loss 0.2746, acc 0.8789 (0.9055 / 0.8524) Training 7: 105000 / 121616: Premsel loss 0.2762, acc 0.8773 (0.9079 / 0.8466) Training 7: 112000 / 121616: Premsel loss 0.2761, acc 0.8780 (0.9142 / 0.8417) Training 7: 119000 / 121616: Premsel loss 0.2649, acc 0.8847 (0.9129 / 0.8564) Evaluation 7: Premsel loss 0.2631, acc 0.8854 (0.9245 / 0.8463) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_27.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-coop___out1.pkl.gz Training 8: 0 / 108509: Premsel loss 0.2651, acc 0.8845 (0.9171 / 0.8519) Training 8: 7000 / 108509: Premsel loss 0.2573, acc 0.8868 (0.9172 / 0.8563) Training 8: 14000 / 108509: Premsel loss 0.2510, acc 0.8909 (0.9184 / 0.8633) Training 8: 21000 / 108509: Premsel loss 0.2550, acc 0.8885 (0.9140 / 0.8630) Training 8: 28000 / 108509: Premsel loss 0.2512, acc 0.8907 (0.9205 / 0.8608) Training 8: 35000 / 108509: Premsel loss 0.2489, acc 0.8916 (0.9224 / 0.8607) Training 8: 42000 / 108509: Premsel loss 0.2566, acc 0.8870 (0.9196 / 0.8545) Training 8: 49000 / 108509: Premsel loss 0.2425, acc 0.8945 (0.9257 / 0.8634) Training 8: 56000 / 108509: Premsel loss 0.2500, acc 0.8899 (0.9178 / 0.8621) Training 8: 63000 / 108509: Premsel loss 0.2505, acc 0.8910 (0.9190 / 0.8629) Training 8: 70000 / 108509: Premsel loss 0.2581, acc 0.8858 (0.9193 / 0.8523) Training 8: 77000 / 108509: Premsel loss 0.2540, acc 0.8907 (0.9227 / 0.8588) Training 8: 84000 / 108509: Premsel loss 0.2427, acc 0.8946 (0.9236 / 0.8656) Training 8: 91000 / 108509: Premsel loss 0.2499, acc 0.8902 (0.9191 / 0.8614) Training 8: 98000 / 108509: Premsel loss 0.2531, acc 0.8890 (0.9197 / 0.8583) Training 8: 105000 / 108509: Premsel loss 0.2419, acc 0.8962 (0.9243 / 0.8680) Evaluation 8: Premsel loss 0.2402, acc 0.8965 (0.9418 / 0.8512) Training 9: 0 / 108509: Premsel loss 0.2455, acc 0.8940 (0.9265 / 0.8616) Training 9: 7000 / 108509: Premsel loss 0.2384, acc 0.8965 (0.9224 / 0.8707) Training 9: 14000 / 108509: Premsel loss 0.2425, acc 0.8950 (0.9221 / 0.8679) Training 9: 21000 / 108509: Premsel loss 0.2392, acc 0.8959 (0.9275 / 0.8643) Training 9: 28000 / 108509: Premsel loss 0.2456, acc 0.8928 (0.9204 / 0.8652) Training 9: 35000 / 108509: Premsel loss 0.2363, acc 0.8981 (0.9258 / 0.8705) Training 9: 42000 / 108509: Premsel loss 0.2414, acc 0.8965 (0.9217 / 0.8713) Training 9: 49000 / 108509: Premsel loss 0.2370, acc 0.8978 (0.9242 / 0.8714) Training 9: 56000 / 108509: Premsel loss 0.2427, acc 0.8945 (0.9255 / 0.8634) Training 9: 63000 / 108509: Premsel loss 0.2351, acc 0.8991 (0.9289 / 0.8694) Training 9: 70000 / 108509: Premsel loss 0.2374, acc 0.8975 (0.9316 / 0.8633) Training 9: 77000 / 108509: Premsel loss 0.2429, acc 0.8951 (0.9210 / 0.8692) Training 9: 84000 / 108509: Premsel loss 0.2363, acc 0.8978 (0.9249 / 0.8708) Training 9: 91000 / 108509: Premsel loss 0.2270, acc 0.9022 (0.9300 / 0.8744) Training 9: 98000 / 108509: Premsel loss 0.2352, acc 0.8985 (0.9259 / 0.8712) Training 9: 105000 / 108509: Premsel loss 0.2308, acc 0.9006 (0.9282 / 0.8730) Evaluation 9: Premsel loss 0.2333, acc 0.8994 (0.8972 / 0.9016) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_06.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz Training 10: 0 / 129451: Premsel loss 0.2386, acc 0.8965 (0.9274 / 0.8655) Training 10: 7000 / 129451: Premsel loss 0.2574, acc 0.8873 (0.9139 / 0.8606) Training 10: 14000 / 129451: Premsel loss 0.2603, acc 0.8852 (0.9117 / 0.8587) Training 10: 21000 / 129451: Premsel loss 0.2621, acc 0.8837 (0.9118 / 0.8556) Training 10: 28000 / 129451: Premsel loss 0.2675, acc 0.8846 (0.9114 / 0.8578) Training 10: 35000 / 129451: Premsel loss 0.2587, acc 0.8871 (0.9088 / 0.8654) Training 10: 42000 / 129451: Premsel loss 0.2515, acc 0.8911 (0.9158 / 0.8665) Training 10: 49000 / 129451: Premsel loss 0.2573, acc 0.8886 (0.9091 / 0.8681) Training 10: 56000 / 129451: Premsel loss 0.2656, acc 0.8831 (0.9092 / 0.8570) Training 10: 63000 / 129451: Premsel loss 0.2507, acc 0.8900 (0.9147 / 0.8652) Training 10: 70000 / 129451: Premsel loss 0.2529, acc 0.8894 (0.9145 / 0.8643) Training 10: 77000 / 129451: Premsel loss 0.2477, acc 0.8918 (0.9187 / 0.8650) Training 10: 84000 / 129451: Premsel loss 0.2483, acc 0.8919 (0.9184 / 0.8655) Training 10: 91000 / 129451: Premsel loss 0.2516, acc 0.8907 (0.9135 / 0.8680) Training 10: 98000 / 129451: Premsel loss 0.2546, acc 0.8883 (0.9082 / 0.8684) Training 10: 105000 / 129451: Premsel loss 0.2508, acc 0.8908 (0.9151 / 0.8665) Training 10: 112000 / 129451: Premsel loss 0.2670, acc 0.8819 (0.9124 / 0.8514) Training 10: 119000 / 129451: Premsel loss 0.2536, acc 0.8896 (0.9124 / 0.8669) Training 10: 126000 / 129451: Premsel loss 0.2597, acc 0.8867 (0.9098 / 0.8636) Evaluation 10: Premsel loss 0.2456, acc 0.8930 (0.8974 / 0.8887) Training 11: 0 / 129451: Premsel loss 0.2479, acc 0.8923 (0.9208 / 0.8638) Training 11: 7000 / 129451: Premsel loss 0.2526, acc 0.8884 (0.9117 / 0.8652) Training 11: 14000 / 129451: Premsel loss 0.2484, acc 0.8917 (0.9188 / 0.8646) Training 11: 21000 / 129451: Premsel loss 0.2576, acc 0.8865 (0.9131 / 0.8598) Training 11: 28000 / 129451: Premsel loss 0.2501, acc 0.8908 (0.9213 / 0.8603) Training 11: 35000 / 129451: Premsel loss 0.2475, acc 0.8928 (0.9166 / 0.8690) Training 11: 42000 / 129451: Premsel loss 0.2460, acc 0.8924 (0.9198 / 0.8651) Training 11: 49000 / 129451: Premsel loss 0.2393, acc 0.8966 (0.9215 / 0.8717) Training 11: 56000 / 129451: Premsel loss 0.2405, acc 0.8965 (0.9195 / 0.8735) Training 11: 63000 / 129451: Premsel loss 0.2473, acc 0.8927 (0.9161 / 0.8693) Training 11: 70000 / 129451: Premsel loss 0.2454, acc 0.8935 (0.9208 / 0.8661) Training 11: 77000 / 129451: Premsel loss 0.2403, acc 0.8965 (0.9228 / 0.8701) Training 11: 84000 / 129451: Premsel loss 0.2402, acc 0.8960 (0.9212 / 0.8707) Training 11: 91000 / 129451: Premsel loss 0.2456, acc 0.8919 (0.9151 / 0.8688) Training 11: 98000 / 129451: Premsel loss 0.2409, acc 0.8965 (0.9196 / 0.8735) Training 11: 105000 / 129451: Premsel loss 0.2444, acc 0.8946 (0.9215 / 0.8676) Training 11: 112000 / 129451: Premsel loss 0.2417, acc 0.8956 (0.9242 / 0.8671) Training 11: 119000 / 129451: Premsel loss 0.2391, acc 0.8966 (0.9227 / 0.8705) Training 11: 126000 / 129451: Premsel loss 0.2449, acc 0.8930 (0.9266 / 0.8594) Evaluation 11: Premsel loss 0.2305, acc 0.9016 (0.9290 / 0.8742) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_74avg-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_00.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 12: 0 / 123545: Premsel loss 0.2479, acc 0.8917 (0.9137 / 0.8696) Training 12: 7000 / 123545: Premsel loss 0.2381, acc 0.8980 (0.9245 / 0.8715) Training 12: 14000 / 123545: Premsel loss 0.2332, acc 0.8997 (0.9263 / 0.8731) Training 12: 21000 / 123545: Premsel loss 0.2355, acc 0.8985 (0.9173 / 0.8798) Training 12: 28000 / 123545: Premsel loss 0.2384, acc 0.8970 (0.9243 / 0.8696) Training 12: 35000 / 123545: Premsel loss 0.2371, acc 0.8972 (0.9162 / 0.8783) Training 12: 42000 / 123545: Premsel loss 0.2383, acc 0.8978 (0.9258 / 0.8699) Training 12: 49000 / 123545: Premsel loss 0.2341, acc 0.8988 (0.9153 / 0.8824) Training 12: 56000 / 123545: Premsel loss 0.2451, acc 0.8939 (0.9227 / 0.8651) Training 12: 63000 / 123545: Premsel loss 0.2376, acc 0.8981 (0.9172 / 0.8790) Training 12: 70000 / 123545: Premsel loss 0.2399, acc 0.8960 (0.9212 / 0.8709) Training 12: 77000 / 123545: Premsel loss 0.2312, acc 0.9003 (0.9257 / 0.8749) Training 12: 84000 / 123545: Premsel loss 0.2395, acc 0.8959 (0.9167 / 0.8752) Training 12: 91000 / 123545: Premsel loss 0.2396, acc 0.8970 (0.9166 / 0.8775) Training 12: 98000 / 123545: Premsel loss 0.2331, acc 0.8994 (0.9206 / 0.8783) Training 12: 105000 / 123545: Premsel loss 0.2329, acc 0.8996 (0.9276 / 0.8716) Training 12: 112000 / 123545: Premsel loss 0.2290, acc 0.9023 (0.9264 / 0.8781) Training 12: 119000 / 123545: Premsel loss 0.2297, acc 0.9020 (0.9307 / 0.8733) Evaluation 12: Premsel loss 0.2330, acc 0.9003 (0.9494 / 0.8512) Training 13: 0 / 123545: Premsel loss 0.2344, acc 0.8997 (0.9226 / 0.8768) Training 13: 7000 / 123545: Premsel loss 0.2227, acc 0.9053 (0.9319 / 0.8786) Training 13: 14000 / 123545: Premsel loss 0.2260, acc 0.9049 (0.9335 / 0.8762) Training 13: 21000 / 123545: Premsel loss 0.2321, acc 0.9007 (0.9293 / 0.8721) Training 13: 28000 / 123545: Premsel loss 0.2214, acc 0.9051 (0.9327 / 0.8776) Training 13: 35000 / 123545: Premsel loss 0.2254, acc 0.9033 (0.9298 / 0.8768) Training 13: 42000 / 123545: Premsel loss 0.2232, acc 0.9049 (0.9313 / 0.8785) Training 13: 49000 / 123545: Premsel loss 0.2222, acc 0.9062 (0.9319 / 0.8806) Training 13: 56000 / 123545: Premsel loss 0.2422, acc 0.8956 (0.9235 / 0.8677) Training 13: 63000 / 123545: Premsel loss 0.2213, acc 0.9054 (0.9295 / 0.8813) Training 13: 70000 / 123545: Premsel loss 0.2316, acc 0.9011 (0.9240 / 0.8782) Training 13: 77000 / 123545: Premsel loss 0.2277, acc 0.9020 (0.9278 / 0.8761) Training 13: 84000 / 123545: Premsel loss 0.2319, acc 0.9002 (0.9287 / 0.8718) Training 13: 91000 / 123545: Premsel loss 0.2256, acc 0.9027 (0.9283 / 0.8772) Training 13: 98000 / 123545: Premsel loss 0.2308, acc 0.9011 (0.9237 / 0.8786) Training 13: 105000 / 123545: Premsel loss 0.2248, acc 0.9052 (0.9291 / 0.8814) Training 13: 112000 / 123545: Premsel loss 0.2315, acc 0.9008 (0.9206 / 0.8810) Training 13: 119000 / 123545: Premsel loss 0.2246, acc 0.9040 (0.9248 / 0.8832) Evaluation 13: Premsel loss 0.2201, acc 0.9068 (0.9290 / 0.8845) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_02.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_09.pkl.gz Training 14: 0 / 117851: Premsel loss 0.2241, acc 0.9048 (0.9266 / 0.8830) Training 14: 7000 / 117851: Premsel loss 0.2248, acc 0.9049 (0.9297 / 0.8802) Training 14: 14000 / 117851: Premsel loss 0.2237, acc 0.9050 (0.9287 / 0.8814) Training 14: 21000 / 117851: Premsel loss 0.2251, acc 0.9055 (0.9333 / 0.8778) Training 14: 28000 / 117851: Premsel loss 0.2201, acc 0.9073 (0.9325 / 0.8821) Training 14: 35000 / 117851: Premsel loss 0.2069, acc 0.9143 (0.9399 / 0.8888) Training 14: 42000 / 117851: Premsel loss 0.2183, acc 0.9091 (0.9378 / 0.8805) Training 14: 49000 / 117851: Premsel loss 0.2188, acc 0.9082 (0.9327 / 0.8837) Training 14: 56000 / 117851: Premsel loss 0.2210, acc 0.9079 (0.9323 / 0.8835) Training 14: 63000 / 117851: Premsel loss 0.2145, acc 0.9100 (0.9373 / 0.8828) Training 14: 70000 / 117851: Premsel loss 0.2175, acc 0.9093 (0.9378 / 0.8808) Training 14: 77000 / 117851: Premsel loss 0.2151, acc 0.9104 (0.9343 / 0.8866) Training 14: 84000 / 117851: Premsel loss 0.2186, acc 0.9076 (0.9346 / 0.8807) Training 14: 91000 / 117851: Premsel loss 0.2133, acc 0.9107 (0.9372 / 0.8842) Training 14: 98000 / 117851: Premsel loss 0.2086, acc 0.9133 (0.9386 / 0.8879) Training 14: 105000 / 117851: Premsel loss 0.2230, acc 0.9060 (0.9335 / 0.8785) Training 14: 112000 / 117851: Premsel loss 0.2142, acc 0.9105 (0.9369 / 0.8842) Evaluation 14: Premsel loss 0.2110, acc 0.9116 (0.9161 / 0.9071) Training 15: 0 / 117851: Premsel loss 0.2183, acc 0.9083 (0.9366 / 0.8799) Training 15: 7000 / 117851: Premsel loss 0.2144, acc 0.9108 (0.9392 / 0.8824) Training 15: 14000 / 117851: Premsel loss 0.2105, acc 0.9127 (0.9390 / 0.8864) Training 15: 21000 / 117851: Premsel loss 0.2111, acc 0.9129 (0.9368 / 0.8891) Training 15: 28000 / 117851: Premsel loss 0.2093, acc 0.9132 (0.9415 / 0.8850) Training 15: 35000 / 117851: Premsel loss 0.2011, acc 0.9172 (0.9444 / 0.8900) Training 15: 42000 / 117851: Premsel loss 0.2080, acc 0.9139 (0.9395 / 0.8884) Training 15: 49000 / 117851: Premsel loss 0.2034, acc 0.9158 (0.9409 / 0.8907) Training 15: 56000 / 117851: Premsel loss 0.2128, acc 0.9121 (0.9385 / 0.8857) Training 15: 63000 / 117851: Premsel loss 0.2049, acc 0.9153 (0.9409 / 0.8898) Training 15: 70000 / 117851: Premsel loss 0.2114, acc 0.9125 (0.9384 / 0.8867) Training 15: 77000 / 117851: Premsel loss 0.2112, acc 0.9126 (0.9402 / 0.8849) Training 15: 84000 / 117851: Premsel loss 0.1965, acc 0.9201 (0.9452 / 0.8949) Training 15: 91000 / 117851: Premsel loss 0.2043, acc 0.9163 (0.9442 / 0.8884) Training 15: 98000 / 117851: Premsel loss 0.2043, acc 0.9161 (0.9411 / 0.8911) Training 15: 105000 / 117851: Premsel loss 0.2108, acc 0.9125 (0.9372 / 0.8878) Training 15: 112000 / 117851: Premsel loss 0.2080, acc 0.9139 (0.9371 / 0.8907) Evaluation 15: Premsel loss 0.2039, acc 0.9153 (0.9192 / 0.9115) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_01.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz Training 16: 0 / 130743: Premsel loss 0.2079, acc 0.9137 (0.9405 / 0.8869) Training 16: 7000 / 130743: Premsel loss 0.2202, acc 0.9072 (0.9332 / 0.8812) Training 16: 14000 / 130743: Premsel loss 0.2227, acc 0.9059 (0.9303 / 0.8815) Training 16: 21000 / 130743: Premsel loss 0.2306, acc 0.9012 (0.9271 / 0.8754) Training 16: 28000 / 130743: Premsel loss 0.2270, acc 0.9033 (0.9292 / 0.8774) Training 16: 35000 / 130743: Premsel loss 0.2261, acc 0.9045 (0.9278 / 0.8812) Training 16: 42000 / 130743: Premsel loss 0.2279, acc 0.9041 (0.9265 / 0.8816) Training 16: 49000 / 130743: Premsel loss 0.2327, acc 0.9007 (0.9243 / 0.8771) Training 16: 56000 / 130743: Premsel loss 0.2238, acc 0.9047 (0.9290 / 0.8805) Training 16: 63000 / 130743: Premsel loss 0.2360, acc 0.8992 (0.9248 / 0.8735) Training 16: 70000 / 130743: Premsel loss 0.2320, acc 0.9019 (0.9283 / 0.8755) Training 16: 77000 / 130743: Premsel loss 0.2316, acc 0.9017 (0.9268 / 0.8766) Training 16: 84000 / 130743: Premsel loss 0.2295, acc 0.9038 (0.9294 / 0.8781) Training 16: 91000 / 130743: Premsel loss 0.2269, acc 0.9031 (0.9299 / 0.8763) Training 16: 98000 / 130743: Premsel loss 0.2248, acc 0.9050 (0.9300 / 0.8799) Training 16: 105000 / 130743: Premsel loss 0.2310, acc 0.9020 (0.9254 / 0.8786) Training 16: 112000 / 130743: Premsel loss 0.2345, acc 0.8995 (0.9219 / 0.8770) Training 16: 119000 / 130743: Premsel loss 0.2291, acc 0.9028 (0.9275 / 0.8781) Training 16: 126000 / 130743: Premsel loss 0.2291, acc 0.9021 (0.9275 / 0.8766) Evaluation 16: Premsel loss 0.2206, acc 0.9070 (0.9426 / 0.8713) Training 17: 0 / 130743: Premsel loss 0.2229, acc 0.9050 (0.9282 / 0.8818) Training 17: 7000 / 130743: Premsel loss 0.2153, acc 0.9096 (0.9334 / 0.8857) Training 17: 14000 / 130743: Premsel loss 0.2220, acc 0.9059 (0.9311 / 0.8807) Training 17: 21000 / 130743: Premsel loss 0.2211, acc 0.9059 (0.9292 / 0.8825) Training 17: 28000 / 130743: Premsel loss 0.2273, acc 0.9024 (0.9330 / 0.8719) Training 17: 35000 / 130743: Premsel loss 0.2228, acc 0.9047 (0.9263 / 0.8832) Training 17: 42000 / 130743: Premsel loss 0.2266, acc 0.9034 (0.9343 / 0.8724) Training 17: 49000 / 130743: Premsel loss 0.2126, acc 0.9113 (0.9365 / 0.8861) Training 17: 56000 / 130743: Premsel loss 0.2239, acc 0.9060 (0.9300 / 0.8820) Training 17: 63000 / 130743: Premsel loss 0.2256, acc 0.9036 (0.9266 / 0.8806) Training 17: 70000 / 130743: Premsel loss 0.2217, acc 0.9062 (0.9307 / 0.8817) Training 17: 77000 / 130743: Premsel loss 0.2204, acc 0.9069 (0.9333 / 0.8806) Training 17: 84000 / 130743: Premsel loss 0.2283, acc 0.9026 (0.9333 / 0.8719) Training 17: 91000 / 130743: Premsel loss 0.2189, acc 0.9080 (0.9339 / 0.8820) Training 17: 98000 / 130743: Premsel loss 0.2143, acc 0.9096 (0.9329 / 0.8864) Training 17: 105000 / 130743: Premsel loss 0.2309, acc 0.9014 (0.9306 / 0.8723) Training 17: 112000 / 130743: Premsel loss 0.2169, acc 0.9087 (0.9380 / 0.8794) Training 17: 119000 / 130743: Premsel loss 0.2186, acc 0.9082 (0.9341 / 0.8823) Training 17: 126000 / 130743: Premsel loss 0.2223, acc 0.9061 (0.9296 / 0.8827) Evaluation 17: Premsel loss 0.2147, acc 0.9096 (0.9384 / 0.8807) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_16.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz Training 18: 0 / 123254: Premsel loss 0.2239, acc 0.9048 (0.9300 / 0.8795) Training 18: 7000 / 123254: Premsel loss 0.2239, acc 0.9046 (0.9302 / 0.8789) Training 18: 14000 / 123254: Premsel loss 0.2228, acc 0.9064 (0.9241 / 0.8887) Training 18: 21000 / 123254: Premsel loss 0.2270, acc 0.9049 (0.9229 / 0.8869) Training 18: 28000 / 123254: Premsel loss 0.2232, acc 0.9060 (0.9309 / 0.8811) Training 18: 35000 / 123254: Premsel loss 0.2216, acc 0.9056 (0.9269 / 0.8843) Training 18: 42000 / 123254: Premsel loss 0.2171, acc 0.9091 (0.9341 / 0.8841) Training 18: 49000 / 123254: Premsel loss 0.2131, acc 0.9106 (0.9374 / 0.8839) Training 18: 56000 / 123254: Premsel loss 0.2344, acc 0.8988 (0.9181 / 0.8794) Training 18: 63000 / 123254: Premsel loss 0.2199, acc 0.9070 (0.9260 / 0.8880) Training 18: 70000 / 123254: Premsel loss 0.2181, acc 0.9093 (0.9312 / 0.8874) Training 18: 77000 / 123254: Premsel loss 0.2169, acc 0.9079 (0.9290 / 0.8868) Training 18: 84000 / 123254: Premsel loss 0.2233, acc 0.9049 (0.9268 / 0.8831) Training 18: 91000 / 123254: Premsel loss 0.2277, acc 0.9034 (0.9280 / 0.8787) Training 18: 98000 / 123254: Premsel loss 0.2187, acc 0.9073 (0.9337 / 0.8810) Training 18: 105000 / 123254: Premsel loss 0.2191, acc 0.9069 (0.9300 / 0.8838) Training 18: 112000 / 123254: Premsel loss 0.2125, acc 0.9110 (0.9339 / 0.8880) Training 18: 119000 / 123254: Premsel loss 0.2165, acc 0.9080 (0.9306 / 0.8854) Evaluation 18: Premsel loss 0.2156, acc 0.9087 (0.9246 / 0.8928) Training 19: 0 / 123254: Premsel loss 0.2240, acc 0.9053 (0.9275 / 0.8831) Training 19: 7000 / 123254: Premsel loss 0.2200, acc 0.9083 (0.9338 / 0.8828) Training 19: 14000 / 123254: Premsel loss 0.2153, acc 0.9092 (0.9332 / 0.8852) Training 19: 21000 / 123254: Premsel loss 0.2215, acc 0.9077 (0.9357 / 0.8796) Training 19: 28000 / 123254: Premsel loss 0.2133, acc 0.9104 (0.9311 / 0.8896) Training 19: 35000 / 123254: Premsel loss 0.2201, acc 0.9065 (0.9324 / 0.8806) Training 19: 42000 / 123254: Premsel loss 0.2202, acc 0.9072 (0.9267 / 0.8876) Training 19: 49000 / 123254: Premsel loss 0.2201, acc 0.9072 (0.9313 / 0.8831) Training 19: 56000 / 123254: Premsel loss 0.2119, acc 0.9107 (0.9319 / 0.8896) Training 19: 63000 / 123254: Premsel loss 0.2177, acc 0.9082 (0.9252 / 0.8912) Training 19: 70000 / 123254: Premsel loss 0.2135, acc 0.9096 (0.9333 / 0.8859) Training 19: 77000 / 123254: Premsel loss 0.2168, acc 0.9096 (0.9334 / 0.8859) Training 19: 84000 / 123254: Premsel loss 0.2153, acc 0.9084 (0.9314 / 0.8853) Training 19: 91000 / 123254: Premsel loss 0.2137, acc 0.9105 (0.9341 / 0.8869) Training 19: 98000 / 123254: Premsel loss 0.2107, acc 0.9118 (0.9365 / 0.8870) Training 19: 105000 / 123254: Premsel loss 0.2170, acc 0.9085 (0.9303 / 0.8868) Training 19: 112000 / 123254: Premsel loss 0.2183, acc 0.9090 (0.9289 / 0.8892) Training 19: 119000 / 123254: Premsel loss 0.2126, acc 0.9108 (0.9341 / 0.8875) Evaluation 19: Premsel loss 0.2103, acc 0.9114 (0.9560 / 0.8669) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l700-e0.20+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_04.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d50-l900-e0.15+coop-mzr02___out1.pkl.gz Training 20: 0 / 127944: Premsel loss 0.2143, acc 0.9102 (0.9335 / 0.8869) Training 20: 7000 / 127944: Premsel loss 0.2154, acc 0.9095 (0.9311 / 0.8880) Training 20: 14000 / 127944: Premsel loss 0.2054, acc 0.9142 (0.9404 / 0.8881) Training 20: 21000 / 127944: Premsel loss 0.2197, acc 0.9077 (0.9299 / 0.8855) Training 20: 28000 / 127944: Premsel loss 0.2139, acc 0.9101 (0.9353 / 0.8849) Training 20: 35000 / 127944: Premsel loss 0.2183, acc 0.9077 (0.9351 / 0.8803) Training 20: 42000 / 127944: Premsel loss 0.2163, acc 0.9091 (0.9316 / 0.8865) Training 20: 49000 / 127944: Premsel loss 0.2071, acc 0.9132 (0.9350 / 0.8913) Training 20: 56000 / 127944: Premsel loss 0.2093, acc 0.9121 (0.9366 / 0.8877) Training 20: 63000 / 127944: Premsel loss 0.2066, acc 0.9137 (0.9387 / 0.8886) Training 20: 70000 / 127944: Premsel loss 0.2107, acc 0.9114 (0.9328 / 0.8899) Training 20: 77000 / 127944: Premsel loss 0.2118, acc 0.9120 (0.9361 / 0.8878) Training 20: 84000 / 127944: Premsel loss 0.2120, acc 0.9111 (0.9342 / 0.8879) Training 20: 91000 / 127944: Premsel loss 0.2135, acc 0.9104 (0.9330 / 0.8879) Training 20: 98000 / 127944: Premsel loss 0.2104, acc 0.9123 (0.9353 / 0.8893) Training 20: 105000 / 127944: Premsel loss 0.2111, acc 0.9109 (0.9364 / 0.8855) Training 20: 112000 / 127944: Premsel loss 0.2128, acc 0.9102 (0.9339 / 0.8866) Training 20: 119000 / 127944: Premsel loss 0.2153, acc 0.9098 (0.9344 / 0.8852) Training 20: 126000 / 127944: Premsel loss 0.2123, acc 0.9114 (0.9390 / 0.8838) Evaluation 20: Premsel loss 0.2086, acc 0.9125 (0.9407 / 0.8844) Training 21: 0 / 127944: Premsel loss 0.2160, acc 0.9092 (0.9329 / 0.8856) Training 21: 7000 / 127944: Premsel loss 0.2117, acc 0.9123 (0.9351 / 0.8894) Training 21: 14000 / 127944: Premsel loss 0.2090, acc 0.9120 (0.9353 / 0.8888) Training 21: 21000 / 127944: Premsel loss 0.2063, acc 0.9141 (0.9410 / 0.8872) Training 21: 28000 / 127944: Premsel loss 0.2050, acc 0.9137 (0.9347 / 0.8927) Training 21: 35000 / 127944: Premsel loss 0.2060, acc 0.9128 (0.9379 / 0.8878) Training 21: 42000 / 127944: Premsel loss 0.2087, acc 0.9128 (0.9363 / 0.8894) Training 21: 49000 / 127944: Premsel loss 0.2027, acc 0.9159 (0.9403 / 0.8914) Training 21: 56000 / 127944: Premsel loss 0.2013, acc 0.9163 (0.9389 / 0.8938) Training 21: 63000 / 127944: Premsel loss 0.2120, acc 0.9116 (0.9353 / 0.8879) Training 21: 70000 / 127944: Premsel loss 0.2138, acc 0.9115 (0.9376 / 0.8854) Training 21: 77000 / 127944: Premsel loss 0.2062, acc 0.9138 (0.9391 / 0.8885) Training 21: 84000 / 127944: Premsel loss 0.2103, acc 0.9119 (0.9367 / 0.8871) Training 21: 91000 / 127944: Premsel loss 0.2129, acc 0.9106 (0.9358 / 0.8855) Training 21: 98000 / 127944: Premsel loss 0.2082, acc 0.9128 (0.9369 / 0.8888) Training 21: 105000 / 127944: Premsel loss 0.2061, acc 0.9135 (0.9370 / 0.8899) Training 21: 112000 / 127944: Premsel loss 0.2038, acc 0.9147 (0.9401 / 0.8894) Training 21: 119000 / 127944: Premsel loss 0.2100, acc 0.9122 (0.9348 / 0.8895) Training 21: 126000 / 127944: Premsel loss 0.2082, acc 0.9130 (0.9369 / 0.8891) Evaluation 21: Premsel loss 0.2010, acc 0.9163 (0.9374 / 0.8951) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_25.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_05.pkl.gz Training 22: 0 / 107372: Premsel loss 0.2083, acc 0.9116 (0.9354 / 0.8878) Training 22: 7000 / 107372: Premsel loss 0.1935, acc 0.9210 (0.9452 / 0.8968) Training 22: 14000 / 107372: Premsel loss 0.1941, acc 0.9213 (0.9452 / 0.8974) Training 22: 21000 / 107372: Premsel loss 0.1845, acc 0.9256 (0.9494 / 0.9018) Training 22: 28000 / 107372: Premsel loss 0.1887, acc 0.9235 (0.9474 / 0.8996) Training 22: 35000 / 107372: Premsel loss 0.1864, acc 0.9252 (0.9502 / 0.9001) Training 22: 42000 / 107372: Premsel loss 0.1820, acc 0.9263 (0.9492 / 0.9034) Training 22: 49000 / 107372: Premsel loss 0.1832, acc 0.9268 (0.9514 / 0.9021) Training 22: 56000 / 107372: Premsel loss 0.1880, acc 0.9243 (0.9460 / 0.9025) Training 22: 63000 / 107372: Premsel loss 0.1848, acc 0.9259 (0.9489 / 0.9029) Training 22: 70000 / 107372: Premsel loss 0.1912, acc 0.9228 (0.9468 / 0.8988) Training 22: 77000 / 107372: Premsel loss 0.1845, acc 0.9260 (0.9505 / 0.9016) Training 22: 84000 / 107372: Premsel loss 0.1904, acc 0.9231 (0.9470 / 0.8992) Training 22: 91000 / 107372: Premsel loss 0.1879, acc 0.9238 (0.9462 / 0.9014) Training 22: 98000 / 107372: Premsel loss 0.1807, acc 0.9281 (0.9521 / 0.9042) Training 22: 105000 / 107372: Premsel loss 0.1813, acc 0.9273 (0.9486 / 0.9061) Evaluation 22: Premsel loss 0.1768, acc 0.9296 (0.9580 / 0.9012) Training 23: 0 / 107372: Premsel loss 0.1801, acc 0.9278 (0.9520 / 0.9036) Training 23: 7000 / 107372: Premsel loss 0.1843, acc 0.9265 (0.9510 / 0.9020) Training 23: 14000 / 107372: Premsel loss 0.1786, acc 0.9290 (0.9514 / 0.9066) Training 23: 21000 / 107372: Premsel loss 0.1839, acc 0.9264 (0.9497 / 0.9032) Training 23: 28000 / 107372: Premsel loss 0.1808, acc 0.9279 (0.9520 / 0.9039) Training 23: 35000 / 107372: Premsel loss 0.1834, acc 0.9262 (0.9507 / 0.9018) Training 23: 42000 / 107372: Premsel loss 0.1906, acc 0.9225 (0.9533 / 0.8918) Training 23: 49000 / 107372: Premsel loss 0.1796, acc 0.9282 (0.9489 / 0.9075) Training 23: 56000 / 107372: Premsel loss 0.1800, acc 0.9278 (0.9526 / 0.9029) Training 23: 63000 / 107372: Premsel loss 0.1835, acc 0.9267 (0.9498 / 0.9036) Training 23: 70000 / 107372: Premsel loss 0.1746, acc 0.9299 (0.9529 / 0.9069) Training 23: 77000 / 107372: Premsel loss 0.1808, acc 0.9279 (0.9510 / 0.9049) Training 23: 84000 / 107372: Premsel loss 0.1786, acc 0.9289 (0.9504 / 0.9075) Training 23: 91000 / 107372: Premsel loss 0.1820, acc 0.9266 (0.9518 / 0.9015) Training 23: 98000 / 107372: Premsel loss 0.1780, acc 0.9292 (0.9522 / 0.9061) Training 23: 105000 / 107372: Premsel loss 0.1833, acc 0.9276 (0.9555 / 0.8997) Evaluation 23: Premsel loss 0.1733, acc 0.9311 (0.9520 / 0.9103) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t300-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_22.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz Training 24: 0 / 134014: Premsel loss 0.1761, acc 0.9299 (0.9531 / 0.9068) Training 24: 7000 / 134014: Premsel loss 0.2137, acc 0.9095 (0.9352 / 0.8838) Training 24: 14000 / 134014: Premsel loss 0.2138, acc 0.9100 (0.9347 / 0.8853) Training 24: 21000 / 134014: Premsel loss 0.2106, acc 0.9119 (0.9389 / 0.8850) Training 24: 28000 / 134014: Premsel loss 0.2124, acc 0.9108 (0.9384 / 0.8831) Training 24: 35000 / 134014: Premsel loss 0.2125, acc 0.9108 (0.9387 / 0.8830) Training 24: 42000 / 134014: Premsel loss 0.2124, acc 0.9107 (0.9370 / 0.8845) Training 24: 49000 / 134014: Premsel loss 0.2164, acc 0.9089 (0.9325 / 0.8853) Training 24: 56000 / 134014: Premsel loss 0.2175, acc 0.9086 (0.9350 / 0.8822) Training 24: 63000 / 134014: Premsel loss 0.2178, acc 0.9078 (0.9339 / 0.8817) Training 24: 70000 / 134014: Premsel loss 0.2135, acc 0.9104 (0.9347 / 0.8862) Training 24: 77000 / 134014: Premsel loss 0.2152, acc 0.9088 (0.9348 / 0.8828) Training 24: 84000 / 134014: Premsel loss 0.2099, acc 0.9116 (0.9363 / 0.8869) Training 24: 91000 / 134014: Premsel loss 0.2154, acc 0.9082 (0.9359 / 0.8806) Training 24: 98000 / 134014: Premsel loss 0.2118, acc 0.9110 (0.9359 / 0.8861) Training 24: 105000 / 134014: Premsel loss 0.2111, acc 0.9116 (0.9379 / 0.8853) Training 24: 112000 / 134014: Premsel loss 0.2053, acc 0.9143 (0.9411 / 0.8876) Training 24: 119000 / 134014: Premsel loss 0.2133, acc 0.9098 (0.9354 / 0.8842) Training 24: 126000 / 134014: Premsel loss 0.2200, acc 0.9062 (0.9315 / 0.8809) Training 24: 133000 / 134014: Premsel loss 0.2090, acc 0.9119 (0.9372 / 0.8867) Evaluation 24: Premsel loss 0.2147, acc 0.9094 (0.9069 / 0.9119) Training 25: 0 / 134014: Premsel loss 0.2163, acc 0.9079 (0.9379 / 0.8779) Training 25: 7000 / 134014: Premsel loss 0.2036, acc 0.9153 (0.9417 / 0.8889) Training 25: 14000 / 134014: Premsel loss 0.2072, acc 0.9128 (0.9405 / 0.8852) Training 25: 21000 / 134014: Premsel loss 0.2043, acc 0.9142 (0.9424 / 0.8861) Training 25: 28000 / 134014: Premsel loss 0.2013, acc 0.9163 (0.9417 / 0.8909) Training 25: 35000 / 134014: Premsel loss 0.2097, acc 0.9116 (0.9412 / 0.8819) Training 25: 42000 / 134014: Premsel loss 0.2058, acc 0.9135 (0.9437 / 0.8834) Training 25: 49000 / 134014: Premsel loss 0.2081, acc 0.9117 (0.9352 / 0.8883) Training 25: 56000 / 134014: Premsel loss 0.2102, acc 0.9110 (0.9384 / 0.8837) Training 25: 63000 / 134014: Premsel loss 0.2090, acc 0.9117 (0.9386 / 0.8848) Training 25: 70000 / 134014: Premsel loss 0.2117, acc 0.9109 (0.9350 / 0.8868) Training 25: 77000 / 134014: Premsel loss 0.2163, acc 0.9075 (0.9362 / 0.8788) Training 25: 84000 / 134014: Premsel loss 0.2058, acc 0.9145 (0.9394 / 0.8895) Training 25: 91000 / 134014: Premsel loss 0.2093, acc 0.9123 (0.9379 / 0.8867) Training 25: 98000 / 134014: Premsel loss 0.2091, acc 0.9121 (0.9408 / 0.8835) Training 25: 105000 / 134014: Premsel loss 0.2125, acc 0.9099 (0.9374 / 0.8823) Training 25: 112000 / 134014: Premsel loss 0.2066, acc 0.9128 (0.9406 / 0.8850) Training 25: 119000 / 134014: Premsel loss 0.2155, acc 0.9090 (0.9378 / 0.8802) Training 25: 126000 / 134014: Premsel loss 0.2013, acc 0.9163 (0.9413 / 0.8912) Training 25: 133000 / 134014: Premsel loss 0.2078, acc 0.9130 (0.9419 / 0.8842) Evaluation 25: Premsel loss 0.2071, acc 0.9132 (0.9545 / 0.8720) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_13.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.005___out1.pkl.gz Training 26: 0 / 49868: Premsel loss 0.2108, acc 0.9109 (0.9365 / 0.8854) Training 26: 7000 / 49868: Premsel loss 0.1929, acc 0.9212 (0.9461 / 0.8963) Training 26: 14000 / 49868: Premsel loss 0.1875, acc 0.9240 (0.9482 / 0.8999) Training 26: 21000 / 49868: Premsel loss 0.1856, acc 0.9249 (0.9499 / 0.8999) Training 26: 28000 / 49868: Premsel loss 0.1893, acc 0.9232 (0.9454 / 0.9010) Training 26: 35000 / 49868: Premsel loss 0.1909, acc 0.9227 (0.9450 / 0.9004) Training 26: 42000 / 49868: Premsel loss 0.1821, acc 0.9265 (0.9487 / 0.9043) Training 26: 49000 / 49868: Premsel loss 0.1821, acc 0.9263 (0.9489 / 0.9037) Evaluation 26: Premsel loss 0.1777, acc 0.9286 (0.9492 / 0.9080) Training 27: 0 / 49868: Premsel loss 0.1843, acc 0.9258 (0.9524 / 0.8991) Training 27: 7000 / 49868: Premsel loss 0.1785, acc 0.9287 (0.9526 / 0.9048) Training 27: 14000 / 49868: Premsel loss 0.1826, acc 0.9258 (0.9494 / 0.9023) Training 27: 21000 / 49868: Premsel loss 0.1815, acc 0.9269 (0.9496 / 0.9041) Training 27: 28000 / 49868: Premsel loss 0.1780, acc 0.9284 (0.9506 / 0.9061) Training 27: 35000 / 49868: Premsel loss 0.1809, acc 0.9272 (0.9487 / 0.9058) Training 27: 42000 / 49868: Premsel loss 0.1850, acc 0.9261 (0.9487 / 0.9035) Training 27: 49000 / 49868: Premsel loss 0.1738, acc 0.9312 (0.9550 / 0.9074) Evaluation 27: Premsel loss 0.1730, acc 0.9310 (0.9581 / 0.9040) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_14.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz Training 28: 0 / 95832: Premsel loss 0.1754, acc 0.9296 (0.9514 / 0.9078) Training 28: 7000 / 95832: Premsel loss 0.1944, acc 0.9191 (0.9410 / 0.8972) Training 28: 14000 / 95832: Premsel loss 0.1883, acc 0.9227 (0.9475 / 0.8980) Training 28: 21000 / 95832: Premsel loss 0.1875, acc 0.9239 (0.9487 / 0.8990) Training 28: 28000 / 95832: Premsel loss 0.1894, acc 0.9220 (0.9480 / 0.8960) Training 28: 35000 / 95832: Premsel loss 0.1935, acc 0.9206 (0.9468 / 0.8944) Training 28: 42000 / 95832: Premsel loss 0.1889, acc 0.9229 (0.9458 / 0.8999) Training 28: 49000 / 95832: Premsel loss 0.1879, acc 0.9236 (0.9517 / 0.8954) Training 28: 56000 / 95832: Premsel loss 0.1904, acc 0.9218 (0.9446 / 0.8990) Training 28: 63000 / 95832: Premsel loss 0.1861, acc 0.9237 (0.9489 / 0.8984) Training 28: 70000 / 95832: Premsel loss 0.1885, acc 0.9227 (0.9449 / 0.9004) Training 28: 77000 / 95832: Premsel loss 0.1912, acc 0.9219 (0.9444 / 0.8994) Training 28: 84000 / 95832: Premsel loss 0.1930, acc 0.9206 (0.9442 / 0.8970) Training 28: 91000 / 95832: Premsel loss 0.1872, acc 0.9231 (0.9461 / 0.9000) Evaluation 28: Premsel loss 0.2038, acc 0.9145 (0.9039 / 0.9251) Training 29: 0 / 95832: Premsel loss 0.1897, acc 0.9229 (0.9507 / 0.8950) Training 29: 7000 / 95832: Premsel loss 0.1856, acc 0.9239 (0.9471 / 0.9008) Training 29: 14000 / 95832: Premsel loss 0.1886, acc 0.9227 (0.9473 / 0.8981) Training 29: 21000 / 95832: Premsel loss 0.1895, acc 0.9219 (0.9415 / 0.9022) Training 29: 28000 / 95832: Premsel loss 0.1848, acc 0.9240 (0.9467 / 0.9014) Training 29: 35000 / 95832: Premsel loss 0.1877, acc 0.9230 (0.9483 / 0.8976) Training 29: 42000 / 95832: Premsel loss 0.1890, acc 0.9221 (0.9459 / 0.8983) Training 29: 49000 / 95832: Premsel loss 0.1789, acc 0.9269 (0.9499 / 0.9040) Training 29: 56000 / 95832: Premsel loss 0.1870, acc 0.9233 (0.9481 / 0.8984) Training 29: 63000 / 95832: Premsel loss 0.1832, acc 0.9257 (0.9496 / 0.9018) Training 29: 70000 / 95832: Premsel loss 0.1824, acc 0.9256 (0.9486 / 0.9027) Training 29: 77000 / 95832: Premsel loss 0.1827, acc 0.9251 (0.9477 / 0.9024) Training 29: 84000 / 95832: Premsel loss 0.1923, acc 0.9210 (0.9463 / 0.8956) Training 29: 91000 / 95832: Premsel loss 0.1879, acc 0.9228 (0.9455 / 0.9001) Evaluation 29: Premsel loss 0.1850, acc 0.9243 (0.9359 / 0.9128) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_11.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_17.pkl.gz Training 30: 0 / 116440: Premsel loss 0.1802, acc 0.9266 (0.9539 / 0.8994) Training 30: 7000 / 116440: Premsel loss 0.1809, acc 0.9268 (0.9486 / 0.9050) Training 30: 14000 / 116440: Premsel loss 0.1867, acc 0.9245 (0.9463 / 0.9026) Training 30: 21000 / 116440: Premsel loss 0.2572, acc 0.8880 (0.9168 / 0.8593) Training 30: 28000 / 116440: Premsel loss 0.2135, acc 0.9096 (0.9355 / 0.8838) Training 30: 35000 / 116440: Premsel loss 0.1947, acc 0.9195 (0.9449 / 0.8941) Training 30: 42000 / 116440: Premsel loss 0.1981, acc 0.9189 (0.9454 / 0.8924) Training 30: 49000 / 116440: Premsel loss 0.1938, acc 0.9209 (0.9437 / 0.8981) Training 30: 56000 / 116440: Premsel loss 0.1937, acc 0.9213 (0.9439 / 0.8986) Training 30: 63000 / 116440: Premsel loss 0.1908, acc 0.9228 (0.9477 / 0.8980) Training 30: 70000 / 116440: Premsel loss 0.1879, acc 0.9239 (0.9472 / 0.9006) Training 30: 77000 / 116440: Premsel loss 0.1876, acc 0.9236 (0.9464 / 0.9007) Training 30: 84000 / 116440: Premsel loss 0.1871, acc 0.9246 (0.9497 / 0.8994) Training 30: 91000 / 116440: Premsel loss 0.1860, acc 0.9254 (0.9485 / 0.9022) Training 30: 98000 / 116440: Premsel loss 0.1833, acc 0.9261 (0.9472 / 0.9050) Training 30: 105000 / 116440: Premsel loss 0.1817, acc 0.9269 (0.9496 / 0.9041) Training 30: 112000 / 116440: Premsel loss 0.1809, acc 0.9274 (0.9501 / 0.9046) Evaluation 30: Premsel loss 0.1763, acc 0.9296 (0.9474 / 0.9118) Training 31: 0 / 116440: Premsel loss 0.1751, acc 0.9293 (0.9498 / 0.9089) Training 31: 7000 / 116440: Premsel loss 0.1782, acc 0.9283 (0.9501 / 0.9066) Training 31: 14000 / 116440: Premsel loss 0.1774, acc 0.9287 (0.9517 / 0.9056) Training 31: 21000 / 116440: Premsel loss 0.1824, acc 0.9268 (0.9489 / 0.9046) Training 31: 28000 / 116440: Premsel loss 0.1758, acc 0.9290 (0.9504 / 0.9077) Training 31: 35000 / 116440: Premsel loss 0.1792, acc 0.9281 (0.9517 / 0.9046) Training 31: 42000 / 116440: Premsel loss 0.1780, acc 0.9291 (0.9490 / 0.9092) Training 31: 49000 / 116440: Premsel loss 0.1764, acc 0.9297 (0.9529 / 0.9066) Training 31: 56000 / 116440: Premsel loss 0.1815, acc 0.9269 (0.9479 / 0.9058) Training 31: 63000 / 116440: Premsel loss 0.1776, acc 0.9293 (0.9527 / 0.9060) Training 31: 70000 / 116440: Premsel loss 0.1767, acc 0.9295 (0.9519 / 0.9070) Training 31: 77000 / 116440: Premsel loss 0.1714, acc 0.9321 (0.9533 / 0.9108) Training 31: 84000 / 116440: Premsel loss 0.1808, acc 0.9268 (0.9497 / 0.9039) Training 31: 91000 / 116440: Premsel loss 0.1831, acc 0.9261 (0.9490 / 0.9033) Training 31: 98000 / 116440: Premsel loss 0.1770, acc 0.9289 (0.9530 / 0.9049) Training 31: 105000 / 116440: Premsel loss 0.1781, acc 0.9282 (0.9502 / 0.9061) Training 31: 112000 / 116440: Premsel loss 0.1803, acc 0.9281 (0.9497 / 0.9064) Evaluation 31: Premsel loss 0.1747, acc 0.9300 (0.9379 / 0.9220) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_24.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo2___bb_preds__160___out1.pkl.gz Training 32: 0 / 116631: Premsel loss 0.1778, acc 0.9289 (0.9524 / 0.9054) Training 32: 7000 / 116631: Premsel loss 0.1990, acc 0.9180 (0.9404 / 0.8956) Training 32: 14000 / 116631: Premsel loss 0.1970, acc 0.9183 (0.9401 / 0.8964) Training 32: 21000 / 116631: Premsel loss 0.1981, acc 0.9178 (0.9390 / 0.8965) Training 32: 28000 / 116631: Premsel loss 0.2033, acc 0.9157 (0.9412 / 0.8901) Training 32: 35000 / 116631: Premsel loss 0.2018, acc 0.9164 (0.9406 / 0.8922) Training 32: 42000 / 116631: Premsel loss 0.1941, acc 0.9197 (0.9460 / 0.8934) Training 32: 49000 / 116631: Premsel loss 0.1989, acc 0.9178 (0.9423 / 0.8933) Training 32: 56000 / 116631: Premsel loss 0.2028, acc 0.9155 (0.9428 / 0.8882) Training 32: 63000 / 116631: Premsel loss 0.2018, acc 0.9158 (0.9371 / 0.8944) Training 32: 70000 / 116631: Premsel loss 0.2011, acc 0.9164 (0.9406 / 0.8922) Training 32: 77000 / 116631: Premsel loss 0.1957, acc 0.9195 (0.9443 / 0.8946) Training 32: 84000 / 116631: Premsel loss 0.2037, acc 0.9150 (0.9366 / 0.8933) Training 32: 91000 / 116631: Premsel loss 0.1982, acc 0.9180 (0.9379 / 0.8981) Training 32: 98000 / 116631: Premsel loss 0.2056, acc 0.9148 (0.9401 / 0.8895) Training 32: 105000 / 116631: Premsel loss 0.1962, acc 0.9192 (0.9390 / 0.8994) Training 32: 112000 / 116631: Premsel loss 0.2054, acc 0.9147 (0.9392 / 0.8901) Evaluation 32: Premsel loss 0.1937, acc 0.9206 (0.9412 / 0.9000) Training 33: 0 / 116631: Premsel loss 0.2031, acc 0.9154 (0.9399 / 0.8909) Training 33: 7000 / 116631: Premsel loss 0.1956, acc 0.9195 (0.9471 / 0.8919) Training 33: 14000 / 116631: Premsel loss 0.1998, acc 0.9170 (0.9414 / 0.8926) Training 33: 21000 / 116631: Premsel loss 0.1917, acc 0.9207 (0.9439 / 0.8975) Training 33: 28000 / 116631: Premsel loss 0.1955, acc 0.9195 (0.9476 / 0.8913) Training 33: 35000 / 116631: Premsel loss 0.1970, acc 0.9184 (0.9420 / 0.8947) Training 33: 42000 / 116631: Premsel loss 0.1969, acc 0.9185 (0.9413 / 0.8957) Training 33: 49000 / 116631: Premsel loss 0.1961, acc 0.9181 (0.9434 / 0.8927) Training 33: 56000 / 116631: Premsel loss 0.1962, acc 0.9191 (0.9436 / 0.8947) Training 33: 63000 / 116631: Premsel loss 0.1926, acc 0.9207 (0.9433 / 0.8980) Training 33: 70000 / 116631: Premsel loss 0.1961, acc 0.9187 (0.9445 / 0.8930) Training 33: 77000 / 116631: Premsel loss 0.1931, acc 0.9206 (0.9436 / 0.8975) Training 33: 84000 / 116631: Premsel loss 0.2009, acc 0.9164 (0.9393 / 0.8934) Training 33: 91000 / 116631: Premsel loss 0.2029, acc 0.9155 (0.9398 / 0.8911) Training 33: 98000 / 116631: Premsel loss 0.2011, acc 0.9160 (0.9404 / 0.8916) Training 33: 105000 / 116631: Premsel loss 0.1953, acc 0.9193 (0.9426 / 0.8961) Training 33: 112000 / 116631: Premsel loss 0.1930, acc 0.9202 (0.9445 / 0.8958) Evaluation 33: Premsel loss 0.1933, acc 0.9204 (0.9528 / 0.8881) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_21.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz Training 34: 0 / 119848: Premsel loss 0.1970, acc 0.9177 (0.9381 / 0.8973) Training 34: 7000 / 119848: Premsel loss 0.1904, acc 0.9217 (0.9398 / 0.9035) Training 34: 14000 / 119848: Premsel loss 0.1950, acc 0.9201 (0.9435 / 0.8968) Training 34: 21000 / 119848: Premsel loss 0.1871, acc 0.9234 (0.9465 / 0.9003) Training 34: 28000 / 119848: Premsel loss 0.1850, acc 0.9245 (0.9468 / 0.9022) Training 34: 35000 / 119848: Premsel loss 0.1855, acc 0.9234 (0.9468 / 0.8999) Training 34: 42000 / 119848: Premsel loss 0.1872, acc 0.9228 (0.9451 / 0.9004) Training 34: 49000 / 119848: Premsel loss 0.1823, acc 0.9252 (0.9486 / 0.9018) Training 34: 56000 / 119848: Premsel loss 0.1849, acc 0.9240 (0.9459 / 0.9020) Training 34: 63000 / 119848: Premsel loss 0.1817, acc 0.9255 (0.9477 / 0.9033) Training 34: 70000 / 119848: Premsel loss 0.1839, acc 0.9258 (0.9491 / 0.9024) Training 34: 77000 / 119848: Premsel loss 0.1864, acc 0.9239 (0.9467 / 0.9011) Training 34: 84000 / 119848: Premsel loss 0.1851, acc 0.9239 (0.9468 / 0.9011) Training 34: 91000 / 119848: Premsel loss 0.1866, acc 0.9229 (0.9517 / 0.8941) Training 34: 98000 / 119848: Premsel loss 0.1901, acc 0.9221 (0.9420 / 0.9022) Training 34: 105000 / 119848: Premsel loss 0.1829, acc 0.9252 (0.9502 / 0.9002) Training 34: 112000 / 119848: Premsel loss 0.1901, acc 0.9213 (0.9421 / 0.9004) Training 34: 119000 / 119848: Premsel loss 0.1851, acc 0.9240 (0.9478 / 0.9001) Evaluation 34: Premsel loss 0.1796, acc 0.9264 (0.9548 / 0.8980) Training 35: 0 / 119848: Premsel loss 0.1865, acc 0.9231 (0.9439 / 0.9023) Training 35: 7000 / 119848: Premsel loss 0.1827, acc 0.9247 (0.9454 / 0.9039) Training 35: 14000 / 119848: Premsel loss 0.1831, acc 0.9246 (0.9477 / 0.9014) Training 35: 21000 / 119848: Premsel loss 0.1844, acc 0.9240 (0.9475 / 0.9006) Training 35: 28000 / 119848: Premsel loss 0.1797, acc 0.9270 (0.9513 / 0.9027) Training 35: 35000 / 119848: Premsel loss 0.1818, acc 0.9251 (0.9484 / 0.9017) Training 35: 42000 / 119848: Premsel loss 0.1814, acc 0.9256 (0.9490 / 0.9022) Training 35: 49000 / 119848: Premsel loss 0.1792, acc 0.9264 (0.9482 / 0.9046) Training 35: 56000 / 119848: Premsel loss 0.1774, acc 0.9277 (0.9520 / 0.9035) Training 35: 63000 / 119848: Premsel loss 0.1832, acc 0.9252 (0.9468 / 0.9037) Training 35: 70000 / 119848: Premsel loss 0.1859, acc 0.9236 (0.9461 / 0.9011) Training 35: 77000 / 119848: Premsel loss 0.1873, acc 0.9232 (0.9447 / 0.9017) Training 35: 84000 / 119848: Premsel loss 0.1849, acc 0.9242 (0.9447 / 0.9037) Training 35: 91000 / 119848: Premsel loss 0.1829, acc 0.9248 (0.9461 / 0.9035) Training 35: 98000 / 119848: Premsel loss 0.1822, acc 0.9250 (0.9447 / 0.9053) Training 35: 105000 / 119848: Premsel loss 0.1819, acc 0.9256 (0.9485 / 0.9027) Training 35: 112000 / 119848: Premsel loss 0.1830, acc 0.9253 (0.9497 / 0.9009) Training 35: 119000 / 119848: Premsel loss 0.1793, acc 0.9270 (0.9490 / 0.9050) Evaluation 35: Premsel loss 0.1773, acc 0.9275 (0.9597 / 0.8954) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_23.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l1800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz Training 36: 0 / 123621: Premsel loss 0.1772, acc 0.9278 (0.9485 / 0.9072) Training 36: 7000 / 123621: Premsel loss 0.1968, acc 0.9193 (0.9491 / 0.8894) Training 36: 14000 / 123621: Premsel loss 0.1912, acc 0.9214 (0.9453 / 0.8974) Training 36: 21000 / 123621: Premsel loss 0.1892, acc 0.9219 (0.9483 / 0.8954) Training 36: 28000 / 123621: Premsel loss 0.1912, acc 0.9212 (0.9419 / 0.9006) Training 36: 35000 / 123621: Premsel loss 0.1887, acc 0.9220 (0.9467 / 0.8972) Training 36: 42000 / 123621: Premsel loss 0.1902, acc 0.9228 (0.9473 / 0.8983) Training 36: 49000 / 123621: Premsel loss 0.1903, acc 0.9208 (0.9452 / 0.8964) Training 36: 56000 / 123621: Premsel loss 0.1966, acc 0.9183 (0.9405 / 0.8961) Training 36: 63000 / 123621: Premsel loss 0.1930, acc 0.9210 (0.9432 / 0.8988) Training 36: 70000 / 123621: Premsel loss 0.1883, acc 0.9224 (0.9439 / 0.9008) Training 36: 77000 / 123621: Premsel loss 0.1948, acc 0.9197 (0.9446 / 0.8948) Training 36: 84000 / 123621: Premsel loss 0.1905, acc 0.9212 (0.9464 / 0.8960) Training 36: 91000 / 123621: Premsel loss 0.1894, acc 0.9213 (0.9451 / 0.8975) Training 36: 98000 / 123621: Premsel loss 0.1932, acc 0.9202 (0.9452 / 0.8951) Training 36: 105000 / 123621: Premsel loss 0.1908, acc 0.9209 (0.9451 / 0.8968) Training 36: 112000 / 123621: Premsel loss 0.1890, acc 0.9219 (0.9465 / 0.8974) Training 36: 119000 / 123621: Premsel loss 0.1882, acc 0.9223 (0.9495 / 0.8950) Evaluation 36: Premsel loss 0.1989, acc 0.9167 (0.9122 / 0.9212) Training 37: 0 / 123621: Premsel loss 0.2062, acc 0.9118 (0.9466 / 0.8770) Training 37: 7000 / 123621: Premsel loss 0.1842, acc 0.9247 (0.9487 / 0.9007) Training 37: 14000 / 123621: Premsel loss 0.1833, acc 0.9257 (0.9526 / 0.8988) Training 37: 21000 / 123621: Premsel loss 0.1915, acc 0.9211 (0.9466 / 0.8956) Training 37: 28000 / 123621: Premsel loss 0.1862, acc 0.9240 (0.9484 / 0.8995) Training 37: 35000 / 123621: Premsel loss 0.1852, acc 0.9238 (0.9492 / 0.8984) Training 37: 42000 / 123621: Premsel loss 0.1869, acc 0.9225 (0.9436 / 0.9015) Training 37: 49000 / 123621: Premsel loss 0.2004, acc 0.9169 (0.9382 / 0.8956) Training 37: 56000 / 123621: Premsel loss 0.1926, acc 0.9205 (0.9438 / 0.8972) Training 37: 63000 / 123621: Premsel loss 0.2000, acc 0.9174 (0.9509 / 0.8840) Training 37: 70000 / 123621: Premsel loss 0.1946, acc 0.9196 (0.9456 / 0.8936) Training 37: 77000 / 123621: Premsel loss 0.1955, acc 0.9191 (0.9485 / 0.8896) Training 37: 84000 / 123621: Premsel loss 0.1822, acc 0.9255 (0.9498 / 0.9012) Training 37: 91000 / 123621: Premsel loss 0.1887, acc 0.9225 (0.9459 / 0.8991) Training 37: 98000 / 123621: Premsel loss 0.1933, acc 0.9203 (0.9438 / 0.8969) Training 37: 105000 / 123621: Premsel loss 0.1861, acc 0.9239 (0.9495 / 0.8982) Training 37: 112000 / 123621: Premsel loss 0.1884, acc 0.9219 (0.9494 / 0.8943) Training 37: 119000 / 123621: Premsel loss 0.1902, acc 0.9224 (0.9486 / 0.8962) Evaluation 37: Premsel loss 0.1843, acc 0.9248 (0.9608 / 0.8889) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_26.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_10.pkl.gz Training 38: 0 / 122252: Premsel loss 0.1870, acc 0.9242 (0.9473 / 0.9012) Training 38: 7000 / 122252: Premsel loss 0.1974, acc 0.9185 (0.9425 / 0.8944) Training 38: 14000 / 122252: Premsel loss 0.1832, acc 0.9251 (0.9460 / 0.9042) Training 38: 21000 / 122252: Premsel loss 0.1974, acc 0.9194 (0.9437 / 0.8952) Training 38: 28000 / 122252: Premsel loss 0.1895, acc 0.9220 (0.9441 / 0.8998) Training 38: 35000 / 122252: Premsel loss 0.1919, acc 0.9205 (0.9414 / 0.8995) Training 38: 42000 / 122252: Premsel loss 0.1887, acc 0.9234 (0.9485 / 0.8983) Training 38: 49000 / 122252: Premsel loss 0.1869, acc 0.9236 (0.9475 / 0.8998) Training 38: 56000 / 122252: Premsel loss 0.1894, acc 0.9233 (0.9443 / 0.9022) Training 38: 63000 / 122252: Premsel loss 0.1895, acc 0.9229 (0.9410 / 0.9048) Training 38: 70000 / 122252: Premsel loss 0.1861, acc 0.9247 (0.9467 / 0.9027) Training 38: 77000 / 122252: Premsel loss 0.1892, acc 0.9224 (0.9446 / 0.9002) Training 38: 84000 / 122252: Premsel loss 0.1887, acc 0.9236 (0.9505 / 0.8968) Training 38: 91000 / 122252: Premsel loss 0.1848, acc 0.9257 (0.9489 / 0.9024) Training 38: 98000 / 122252: Premsel loss 0.1888, acc 0.9227 (0.9491 / 0.8963) Training 38: 105000 / 122252: Premsel loss 0.1929, acc 0.9211 (0.9406 / 0.9016) Training 38: 112000 / 122252: Premsel loss 0.1880, acc 0.9235 (0.9439 / 0.9031) Training 38: 119000 / 122252: Premsel loss 0.1929, acc 0.9209 (0.9413 / 0.9006) Evaluation 38: Premsel loss 0.1851, acc 0.9248 (0.9620 / 0.8876) Training 39: 0 / 122252: Premsel loss 0.1864, acc 0.9242 (0.9440 / 0.9044) Training 39: 7000 / 122252: Premsel loss 0.1877, acc 0.9235 (0.9466 / 0.9005) Training 39: 14000 / 122252: Premsel loss 0.1837, acc 0.9252 (0.9501 / 0.9004) Training 39: 21000 / 122252: Premsel loss 0.1792, acc 0.9273 (0.9482 / 0.9063) Training 39: 28000 / 122252: Premsel loss 0.1844, acc 0.9250 (0.9468 / 0.9032) Training 39: 35000 / 122252: Premsel loss 0.1862, acc 0.9242 (0.9455 / 0.9029) Training 39: 42000 / 122252: Premsel loss 0.1818, acc 0.9266 (0.9496 / 0.9036) Training 39: 49000 / 122252: Premsel loss 0.1881, acc 0.9226 (0.9425 / 0.9027) Training 39: 56000 / 122252: Premsel loss 0.1814, acc 0.9271 (0.9502 / 0.9040) Training 39: 63000 / 122252: Premsel loss 0.1868, acc 0.9236 (0.9467 / 0.9006) Training 39: 70000 / 122252: Premsel loss 0.1829, acc 0.9259 (0.9471 / 0.9046) Training 39: 77000 / 122252: Premsel loss 0.1861, acc 0.9245 (0.9477 / 0.9012) Training 39: 84000 / 122252: Premsel loss 0.1846, acc 0.9255 (0.9485 / 0.9024) Training 39: 91000 / 122252: Premsel loss 0.1779, acc 0.9280 (0.9514 / 0.9046) Training 39: 98000 / 122252: Premsel loss 0.1822, acc 0.9262 (0.9510 / 0.9014) Training 39: 105000 / 122252: Premsel loss 0.1798, acc 0.9282 (0.9527 / 0.9036) Training 39: 112000 / 122252: Premsel loss 0.1831, acc 0.9263 (0.9475 / 0.9051) Training 39: 119000 / 122252: Premsel loss 0.1939, acc 0.9212 (0.9412 / 0.9013) Evaluation 39: Premsel loss 0.1822, acc 0.9264 (0.9616 / 0.8911) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_20.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz Training 40: 0 / 96215: Premsel loss 0.1866, acc 0.9236 (0.9459 / 0.9013) Training 40: 7000 / 96215: Premsel loss 0.1936, acc 0.9205 (0.9436 / 0.8974) Training 40: 14000 / 96215: Premsel loss 0.1971, acc 0.9186 (0.9453 / 0.8918) Training 40: 21000 / 96215: Premsel loss 0.1945, acc 0.9192 (0.9426 / 0.8958) Training 40: 28000 / 96215: Premsel loss 0.2033, acc 0.9157 (0.9412 / 0.8902) Training 40: 35000 / 96215: Premsel loss 0.2025, acc 0.9157 (0.9383 / 0.8930) Training 40: 42000 / 96215: Premsel loss 0.1911, acc 0.9222 (0.9448 / 0.8996) Training 40: 49000 / 96215: Premsel loss 0.1991, acc 0.9188 (0.9404 / 0.8973) Training 40: 56000 / 96215: Premsel loss 0.1914, acc 0.9214 (0.9423 / 0.9006) Training 40: 63000 / 96215: Premsel loss 0.1888, acc 0.9236 (0.9461 / 0.9011) Training 40: 70000 / 96215: Premsel loss 0.2000, acc 0.9172 (0.9380 / 0.8965) Training 40: 77000 / 96215: Premsel loss 0.1989, acc 0.9175 (0.9409 / 0.8940) Training 40: 84000 / 96215: Premsel loss 0.1931, acc 0.9205 (0.9421 / 0.8988) Training 40: 91000 / 96215: Premsel loss 0.1957, acc 0.9188 (0.9440 / 0.8936) Evaluation 40: Premsel loss 0.1907, acc 0.9214 (0.9358 / 0.9069) Training 41: 0 / 96215: Premsel loss 0.1972, acc 0.9181 (0.9414 / 0.8947) Training 41: 7000 / 96215: Premsel loss 0.1957, acc 0.9187 (0.9423 / 0.8950) Training 41: 14000 / 96215: Premsel loss 0.1947, acc 0.9194 (0.9413 / 0.8975) Training 41: 21000 / 96215: Premsel loss 0.1892, acc 0.9220 (0.9463 / 0.8977) Training 41: 28000 / 96215: Premsel loss 0.1912, acc 0.9218 (0.9461 / 0.8975) Training 41: 35000 / 96215: Premsel loss 0.1928, acc 0.9206 (0.9429 / 0.8982) Training 41: 42000 / 96215: Premsel loss 0.1938, acc 0.9196 (0.9447 / 0.8946) Training 41: 49000 / 96215: Premsel loss 0.1967, acc 0.9193 (0.9457 / 0.8929) Training 41: 56000 / 96215: Premsel loss 0.1996, acc 0.9169 (0.9376 / 0.8962) Training 41: 63000 / 96215: Premsel loss 0.1976, acc 0.9179 (0.9406 / 0.8951) Training 41: 70000 / 96215: Premsel loss 0.1948, acc 0.9199 (0.9429 / 0.8970) Training 41: 77000 / 96215: Premsel loss 0.1947, acc 0.9194 (0.9398 / 0.8991) Training 41: 84000 / 96215: Premsel loss 0.1968, acc 0.9188 (0.9407 / 0.8970) Training 41: 91000 / 96215: Premsel loss 0.1942, acc 0.9192 (0.9417 / 0.8967) Evaluation 41: Premsel loss 0.1897, acc 0.9220 (0.9369 / 0.9071) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_18.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz Training 42: 0 / 130835: Premsel loss 0.1959, acc 0.9180 (0.9429 / 0.8930) Training 42: 7000 / 130835: Premsel loss 0.1913, acc 0.9211 (0.9419 / 0.9004) Training 42: 14000 / 130835: Premsel loss 0.1830, acc 0.9250 (0.9506 / 0.8993) Training 42: 21000 / 130835: Premsel loss 0.1893, acc 0.9217 (0.9480 / 0.8955) Training 42: 28000 / 130835: Premsel loss 0.1905, acc 0.9214 (0.9464 / 0.8965) Training 42: 35000 / 130835: Premsel loss 0.1842, acc 0.9250 (0.9480 / 0.9020) Training 42: 42000 / 130835: Premsel loss 0.1856, acc 0.9233 (0.9455 / 0.9011) Training 42: 49000 / 130835: Premsel loss 0.1853, acc 0.9234 (0.9487 / 0.8981) Training 42: 56000 / 130835: Premsel loss 0.1889, acc 0.9226 (0.9508 / 0.8943) Training 42: 63000 / 130835: Premsel loss 0.1881, acc 0.9225 (0.9482 / 0.8967) Training 42: 70000 / 130835: Premsel loss 0.1942, acc 0.9189 (0.9478 / 0.8901) Training 42: 77000 / 130835: Premsel loss 0.1833, acc 0.9255 (0.9502 / 0.9008) Training 42: 84000 / 130835: Premsel loss 0.1839, acc 0.9239 (0.9525 / 0.8954) Training 42: 91000 / 130835: Premsel loss 0.1826, acc 0.9257 (0.9471 / 0.9042) Training 42: 98000 / 130835: Premsel loss 0.1900, acc 0.9217 (0.9440 / 0.8993) Training 42: 105000 / 130835: Premsel loss 0.1858, acc 0.9232 (0.9495 / 0.8970) Training 42: 112000 / 130835: Premsel loss 0.1844, acc 0.9235 (0.9435 / 0.9034) Training 42: 119000 / 130835: Premsel loss 0.1857, acc 0.9237 (0.9476 / 0.8997) Training 42: 126000 / 130835: Premsel loss 0.1880, acc 0.9231 (0.9493 / 0.8969) Evaluation 42: Premsel loss 0.1873, acc 0.9226 (0.9298 / 0.9155) Training 43: 0 / 130835: Premsel loss 0.1852, acc 0.9239 (0.9524 / 0.8954) Training 43: 7000 / 130835: Premsel loss 0.1822, acc 0.9253 (0.9509 / 0.8996) Training 43: 14000 / 130835: Premsel loss 0.1749, acc 0.9286 (0.9545 / 0.9027) Training 43: 21000 / 130835: Premsel loss 0.1796, acc 0.9266 (0.9520 / 0.9012) Training 43: 28000 / 130835: Premsel loss 0.1845, acc 0.9244 (0.9492 / 0.8996) Training 43: 35000 / 130835: Premsel loss 0.1934, acc 0.9196 (0.9443 / 0.8949) Training 43: 42000 / 130835: Premsel loss 0.1847, acc 0.9236 (0.9457 / 0.9016) Training 43: 49000 / 130835: Premsel loss 0.1884, acc 0.9221 (0.9476 / 0.8966) Training 43: 56000 / 130835: Premsel loss 0.1823, acc 0.9257 (0.9508 / 0.9007) Training 43: 63000 / 130835: Premsel loss 0.1815, acc 0.9258 (0.9520 / 0.8996) Training 43: 70000 / 130835: Premsel loss 0.1876, acc 0.9229 (0.9473 / 0.8985) Training 43: 77000 / 130835: Premsel loss 0.1846, acc 0.9234 (0.9441 / 0.9027) Training 43: 84000 / 130835: Premsel loss 0.1813, acc 0.9265 (0.9526 / 0.9005) Training 43: 91000 / 130835: Premsel loss 0.1875, acc 0.9221 (0.9492 / 0.8950) Training 43: 98000 / 130835: Premsel loss 0.1781, acc 0.9277 (0.9528 / 0.9025) Training 43: 105000 / 130835: Premsel loss 0.1896, acc 0.9216 (0.9442 / 0.8990) Training 43: 112000 / 130835: Premsel loss 0.1881, acc 0.9218 (0.9460 / 0.8975) Training 43: 119000 / 130835: Premsel loss 0.1931, acc 0.9194 (0.9434 / 0.8954) Training 43: 126000 / 130835: Premsel loss 0.1860, acc 0.9223 (0.9481 / 0.8965) Evaluation 43: Premsel loss 0.1843, acc 0.9245 (0.9612 / 0.8879) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_12.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop01-VHSLCAXPh+lgb-d50-l900-e0.15loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz Training 44: 0 / 98290: Premsel loss 0.1807, acc 0.9263 (0.9527 / 0.8999) Training 44: 7000 / 98290: Premsel loss 0.1913, acc 0.9209 (0.9438 / 0.8980) Training 44: 14000 / 98290: Premsel loss 0.1894, acc 0.9222 (0.9460 / 0.8983) Training 44: 21000 / 98290: Premsel loss 0.1855, acc 0.9236 (0.9462 / 0.9011) Training 44: 28000 / 98290: Premsel loss 0.1827, acc 0.9250 (0.9486 / 0.9014) Training 44: 35000 / 98290: Premsel loss 0.1854, acc 0.9239 (0.9499 / 0.8980) Training 44: 42000 / 98290: Premsel loss 0.1851, acc 0.9243 (0.9473 / 0.9013) Training 44: 49000 / 98290: Premsel loss 0.1839, acc 0.9255 (0.9486 / 0.9023) Training 44: 56000 / 98290: Premsel loss 0.1868, acc 0.9240 (0.9481 / 0.8998) Training 44: 63000 / 98290: Premsel loss 0.1889, acc 0.9222 (0.9413 / 0.9030) Training 44: 70000 / 98290: Premsel loss 0.1888, acc 0.9223 (0.9435 / 0.9011) Training 44: 77000 / 98290: Premsel loss 0.1851, acc 0.9251 (0.9449 / 0.9052) Training 44: 84000 / 98290: Premsel loss 0.1836, acc 0.9252 (0.9470 / 0.9035) Training 44: 91000 / 98290: Premsel loss 0.1878, acc 0.9232 (0.9481 / 0.8984) Training 44: 98000 / 98290: Premsel loss 0.1890, acc 0.9219 (0.9425 / 0.9013) Evaluation 44: Premsel loss 0.1801, acc 0.9266 (0.9430 / 0.9103) Training 45: 0 / 98290: Premsel loss 0.1818, acc 0.9257 (0.9481 / 0.9033) Training 45: 7000 / 98290: Premsel loss 0.1783, acc 0.9275 (0.9494 / 0.9055) ^[Training 45: 14000 / 98290: Premsel loss 0.1862, acc 0.9242 (0.9473 / 0.9010) Training 45: 21000 / 98290: Premsel loss 0.1850, acc 0.9245 (0.9500 / 0.8990) Training 45: 28000 / 98290: Premsel loss 0.1816, acc 0.9261 (0.9462 / 0.9059) Training 45: 35000 / 98290: Premsel loss 0.1810, acc 0.9265 (0.9475 / 0.9055) Training 45: 42000 / 98290: Premsel loss 0.1870, acc 0.9241 (0.9455 / 0.9026) Training 45: 49000 / 98290: Premsel loss 0.1879, acc 0.9226 (0.9450 / 0.9002) Training 45: 56000 / 98290: Premsel loss 0.1844, acc 0.9252 (0.9479 / 0.9025) Training 45: 63000 / 98290: Premsel loss 0.1809, acc 0.9261 (0.9491 / 0.9031) Training 45: 70000 / 98290: Premsel loss 0.1787, acc 0.9280 (0.9508 / 0.9053) Training 45: 77000 / 98290: Premsel loss 0.1760, acc 0.9283 (0.9488 / 0.9078) Training 45: 84000 / 98290: Premsel loss 0.1785, acc 0.9278 (0.9505 / 0.9052) Training 45: 91000 / 98290: Premsel loss 0.1819, acc 0.9262 (0.9487 / 0.9037) Training 45: 98000 / 98290: Premsel loss 0.1785, acc 0.9274 (0.9502 / 0.9046) Evaluation 45: Premsel loss 0.1780, acc 0.9277 (0.9552 / 0.9002) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz Training 46: 0 / 106255: Premsel loss 0.1838, acc 0.9247 (0.9471 / 0.9023) Training 46: 7000 / 106255: Premsel loss 0.1915, acc 0.9202 (0.9415 / 0.8989) Training 46: 14000 / 106255: Premsel loss 0.1828, acc 0.9230 (0.9422 / 0.9038) Training 46: 21000 / 106255: Premsel loss 0.1884, acc 0.9205 (0.9371 / 0.9038) Training 46: 28000 / 106255: Premsel loss 0.1838, acc 0.9228 (0.9437 / 0.9019) Training 46: 35000 / 106255: Premsel loss 0.1848, acc 0.9229 (0.9442 / 0.9016) Training 46: 42000 / 106255: Premsel loss 0.1873, acc 0.9215 (0.9438 / 0.8992) Training 46: 49000 / 106255: Premsel loss 0.1891, acc 0.9201 (0.9405 / 0.8997) Training 46: 56000 / 106255: Premsel loss 0.1859, acc 0.9219 (0.9415 / 0.9023) Training 46: 63000 / 106255: Premsel loss 0.1887, acc 0.9203 (0.9414 / 0.8991) Training 46: 70000 / 106255: Premsel loss 0.1851, acc 0.9217 (0.9400 / 0.9033) Training 46: 77000 / 106255: Premsel loss 0.1821, acc 0.9238 (0.9459 / 0.9018) Training 46: 84000 / 106255: Premsel loss 0.1879, acc 0.9210 (0.9417 / 0.9002) Training 46: 91000 / 106255: Premsel loss 0.1868, acc 0.9218 (0.9418 / 0.9018) Training 46: 98000 / 106255: Premsel loss 0.1807, acc 0.9246 (0.9481 / 0.9011) Training 46: 105000 / 106255: Premsel loss 0.1885, acc 0.9205 (0.9403 / 0.9006) Evaluation 46: Premsel loss 0.1808, acc 0.9243 (0.9356 / 0.9130) Training 47: 0 / 106255: Premsel loss 0.1837, acc 0.9232 (0.9425 / 0.9038) Training 47: 7000 / 106255: Premsel loss 0.1756, acc 0.9266 (0.9454 / 0.9079) Training 47: 14000 / 106255: Premsel loss 0.1800, acc 0.9243 (0.9440 / 0.9046) Training 47: 21000 / 106255: Premsel loss 0.1720, acc 0.9289 (0.9475 / 0.9103) Training 47: 28000 / 106255: Premsel loss 0.1866, acc 0.9210 (0.9382 / 0.9037) Training 47: 35000 / 106255: Premsel loss 0.1770, acc 0.9261 (0.9477 / 0.9045) Training 47: 42000 / 106255: Premsel loss 0.1829, acc 0.9235 (0.9440 / 0.9029) Training 47: 49000 / 106255: Premsel loss 0.1811, acc 0.9235 (0.9436 / 0.9033) Training 47: 56000 / 106255: Premsel loss 0.1703, acc 0.9291 (0.9494 / 0.9088) Training 47: 63000 / 106255: Premsel loss 0.1822, acc 0.9234 (0.9426 / 0.9042) Training 47: 70000 / 106255: Premsel loss 0.1837, acc 0.9229 (0.9450 / 0.9008) Training 47: 77000 / 106255: Premsel loss 0.1750, acc 0.9269 (0.9478 / 0.9060) Training 47: 84000 / 106255: Premsel loss 0.1804, acc 0.9248 (0.9462 / 0.9034) Training 47: 91000 / 106255: Premsel loss 0.1845, acc 0.9220 (0.9456 / 0.8984) Training 47: 98000 / 106255: Premsel loss 0.1838, acc 0.9226 (0.9450 / 0.9002) Training 47: 105000 / 106255: Premsel loss 0.1832, acc 0.9229 (0.9434 / 0.9025) Evaluation 47: Premsel loss 0.1777, acc 0.9261 (0.9394 / 0.9128) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz Training 48: 0 / 81168: Premsel loss 0.1787, acc 0.9248 (0.9457 / 0.9039) Training 48: 7000 / 81168: Premsel loss 0.2101, acc 0.9105 (0.9347 / 0.8864) Training 48: 14000 / 81168: Premsel loss 0.2093, acc 0.9101 (0.9314 / 0.8888) Training 48: 21000 / 81168: Premsel loss 0.2012, acc 0.9147 (0.9312 / 0.8981) Training 48: 28000 / 81168: Premsel loss 0.2028, acc 0.9135 (0.9352 / 0.8917) Training 48: 35000 / 81168: Premsel loss 0.2049, acc 0.9133 (0.9363 / 0.8903) Training 48: 42000 / 81168: Premsel loss 0.2064, acc 0.9120 (0.9350 / 0.8890) Training 48: 49000 / 81168: Premsel loss 0.2035, acc 0.9126 (0.9361 / 0.8891) Training 48: 56000 / 81168: Premsel loss 0.2023, acc 0.9142 (0.9382 / 0.8903) Training 48: 63000 / 81168: Premsel loss 0.2013, acc 0.9144 (0.9358 / 0.8931) Training 48: 70000 / 81168: Premsel loss 0.2051, acc 0.9112 (0.9278 / 0.8947) Training 48: 77000 / 81168: Premsel loss 0.2029, acc 0.9140 (0.9331 / 0.8948) Evaluation 48: Premsel loss 0.1967, acc 0.9170 (0.9411 / 0.8929) Training 49: 0 / 81168: Premsel loss 0.1980, acc 0.9149 (0.9384 / 0.8915) Training 49: 7000 / 81168: Premsel loss 0.2035, acc 0.9122 (0.9338 / 0.8907) Training 49: 14000 / 81168: Premsel loss 0.2022, acc 0.9130 (0.9352 / 0.8909) Training 49: 21000 / 81168: Premsel loss 0.2013, acc 0.9143 (0.9382 / 0.8905) Training 49: 28000 / 81168: Premsel loss 0.1976, acc 0.9155 (0.9352 / 0.8957) Training 49: 35000 / 81168: Premsel loss 0.1942, acc 0.9182 (0.9410 / 0.8955) Training 49: 42000 / 81168: Premsel loss 0.1958, acc 0.9165 (0.9399 / 0.8931) Training 49: 49000 / 81168: Premsel loss 0.1995, acc 0.9139 (0.9361 / 0.8917) Training 49: 56000 / 81168: Premsel loss 0.1983, acc 0.9163 (0.9403 / 0.8923) Training 49: 63000 / 81168: Premsel loss 0.2000, acc 0.9151 (0.9377 / 0.8926) Training 49: 70000 / 81168: Premsel loss 0.1991, acc 0.9150 (0.9407 / 0.8894) Training 49: 77000 / 81168: Premsel loss 0.1984, acc 0.9161 (0.9377 / 0.8945) Evaluation 49: Premsel loss 0.1950, acc 0.9177 (0.9522 / 0.8832) Loading data... Full data reached, reshuffling... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_04.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz Training 50: 0 / 125384: Premsel loss 0.2041, acc 0.9123 (0.9359 / 0.8887) Training 50: 7000 / 125384: Premsel loss 0.2147, acc 0.9081 (0.9317 / 0.8845) Training 50: 14000 / 125384: Premsel loss 0.2057, acc 0.9133 (0.9355 / 0.8910) Training 50: 21000 / 125384: Premsel loss 0.2095, acc 0.9105 (0.9327 / 0.8884) Training 50: 28000 / 125384: Premsel loss 0.2045, acc 0.9134 (0.9377 / 0.8891) Training 50: 35000 / 125384: Premsel loss 0.2135, acc 0.9093 (0.9363 / 0.8822) Training 50: 42000 / 125384: Premsel loss 0.2063, acc 0.9127 (0.9368 / 0.8886) Training 50: 49000 / 125384: Premsel loss 0.1976, acc 0.9175 (0.9433 / 0.8918) Training 50: 56000 / 125384: Premsel loss 0.2071, acc 0.9127 (0.9392 / 0.8862) Training 50: 63000 / 125384: Premsel loss 0.1966, acc 0.9177 (0.9404 / 0.8950) Training 50: 70000 / 125384: Premsel loss 0.1989, acc 0.9169 (0.9426 / 0.8912) Training 50: 77000 / 125384: Premsel loss 0.2022, acc 0.9151 (0.9383 / 0.8918) Training 50: 84000 / 125384: Premsel loss 0.1965, acc 0.9185 (0.9421 / 0.8950) Training 50: 91000 / 125384: Premsel loss 0.2088, acc 0.9110 (0.9408 / 0.8812) Training 50: 98000 / 125384: Premsel loss 0.2062, acc 0.9135 (0.9379 / 0.8890) Training 50: 105000 / 125384: Premsel loss 0.2079, acc 0.9129 (0.9411 / 0.8847) Training 50: 112000 / 125384: Premsel loss 0.2055, acc 0.9127 (0.9348 / 0.8907) Training 50: 119000 / 125384: Premsel loss 0.1979, acc 0.9172 (0.9446 / 0.8898) Evaluation 50: Premsel loss 0.1983, acc 0.9172 (0.9307 / 0.9038) Training 51: 0 / 125384: Premsel loss 0.1970, acc 0.9182 (0.9436 / 0.8929) Training 51: 7000 / 125384: Premsel loss 0.1919, acc 0.9195 (0.9426 / 0.8964) Training 51: 14000 / 125384: Premsel loss 0.1979, acc 0.9169 (0.9410 / 0.8928) Training 51: 21000 / 125384: Premsel loss 0.1912, acc 0.9210 (0.9491 / 0.8928) Training 51: 28000 / 125384: Premsel loss 0.1980, acc 0.9177 (0.9469 / 0.8886) Training 51: 35000 / 125384: Premsel loss 0.1984, acc 0.9171 (0.9421 / 0.8920) Training 51: 42000 / 125384: Premsel loss 0.1997, acc 0.9160 (0.9420 / 0.8900) Training 51: 49000 / 125384: Premsel loss 0.1937, acc 0.9193 (0.9441 / 0.8945) Training 51: 56000 / 125384: Premsel loss 0.1918, acc 0.9202 (0.9480 / 0.8924) Training 51: 63000 / 125384: Premsel loss 0.1999, acc 0.9163 (0.9421 / 0.8906) Training 51: 70000 / 125384: Premsel loss 0.1944, acc 0.9199 (0.9479 / 0.8920) Training 51: 77000 / 125384: Premsel loss 0.1954, acc 0.9188 (0.9449 / 0.8928) Training 51: 84000 / 125384: Premsel loss 0.1928, acc 0.9206 (0.9459 / 0.8952) Training 51: 91000 / 125384: Premsel loss 0.1990, acc 0.9162 (0.9428 / 0.8895) Training 51: 98000 / 125384: Premsel loss 0.1980, acc 0.9179 (0.9414 / 0.8944) Training 51: 105000 / 125384: Premsel loss 0.1988, acc 0.9163 (0.9395 / 0.8930) Training 51: 112000 / 125384: Premsel loss 0.1969, acc 0.9185 (0.9460 / 0.8909) Training 51: 119000 / 125384: Premsel loss 0.1935, acc 0.9196 (0.9418 / 0.8974) Evaluation 51: Premsel loss 0.1913, acc 0.9209 (0.9488 / 0.8929) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_10.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-coop___out1.pkl.gz Training 52: 0 / 111887: Premsel loss 0.2011, acc 0.9157 (0.9395 / 0.8920) Training 52: 7000 / 111887: Premsel loss 0.1808, acc 0.9255 (0.9473 / 0.9037) Training 52: 14000 / 111887: Premsel loss 0.1794, acc 0.9279 (0.9459 / 0.9098) Training 52: 21000 / 111887: Premsel loss 0.1875, acc 0.9225 (0.9436 / 0.9013) Training 52: 28000 / 111887: Premsel loss 0.1742, acc 0.9293 (0.9489 / 0.9097) Training 52: 35000 / 111887: Premsel loss 0.1914, acc 0.9216 (0.9438 / 0.8994) Training 52: 42000 / 111887: Premsel loss 0.1765, acc 0.9278 (0.9532 / 0.9023) Training 52: 49000 / 111887: Premsel loss 0.1842, acc 0.9240 (0.9509 / 0.8972) Training 52: 56000 / 111887: Premsel loss 0.1767, acc 0.9276 (0.9530 / 0.9022) Training 52: 63000 / 111887: Premsel loss 0.1820, acc 0.9250 (0.9484 / 0.9017) Training 52: 70000 / 111887: Premsel loss 0.1838, acc 0.9247 (0.9470 / 0.9024) Training 52: 77000 / 111887: Premsel loss 0.1785, acc 0.9268 (0.9530 / 0.9005) Training 52: 84000 / 111887: Premsel loss 0.1778, acc 0.9270 (0.9491 / 0.9049) Training 52: 91000 / 111887: Premsel loss 0.1826, acc 0.9254 (0.9500 / 0.9009) Training 52: 98000 / 111887: Premsel loss 0.1715, acc 0.9306 (0.9536 / 0.9076) Training 52: 105000 / 111887: Premsel loss 0.1723, acc 0.9300 (0.9541 / 0.9059) Evaluation 52: Premsel loss 0.1737, acc 0.9291 (0.9585 / 0.8996) Training 53: 0 / 111887: Premsel loss 0.1811, acc 0.9252 (0.9446 / 0.9057) Training 53: 7000 / 111887: Premsel loss 0.1750, acc 0.9289 (0.9493 / 0.9085) Training 53: 14000 / 111887: Premsel loss 0.1708, acc 0.9307 (0.9525 / 0.9088) Training 53: 21000 / 111887: Premsel loss 0.1726, acc 0.9304 (0.9528 / 0.9079) Training 53: 28000 / 111887: Premsel loss 0.1750, acc 0.9286 (0.9509 / 0.9063) Training 53: 35000 / 111887: Premsel loss 0.1776, acc 0.9266 (0.9488 / 0.9043) Training 53: 42000 / 111887: Premsel loss 0.1715, acc 0.9297 (0.9542 / 0.9052) Training 53: 49000 / 111887: Premsel loss 0.1780, acc 0.9274 (0.9473 / 0.9075) Training 53: 56000 / 111887: Premsel loss 0.1761, acc 0.9281 (0.9510 / 0.9052) Training 53: 63000 / 111887: Premsel loss 0.1736, acc 0.9290 (0.9550 / 0.9029) Training 53: 70000 / 111887: Premsel loss 0.1775, acc 0.9277 (0.9505 / 0.9050) Training 53: 77000 / 111887: Premsel loss 0.1764, acc 0.9269 (0.9476 / 0.9063) Training 53: 84000 / 111887: Premsel loss 0.1737, acc 0.9297 (0.9526 / 0.9067) Training 53: 91000 / 111887: Premsel loss 0.1806, acc 0.9260 (0.9451 / 0.9068) Training 53: 98000 / 111887: Premsel loss 0.1747, acc 0.9286 (0.9507 / 0.9066) Training 53: 105000 / 111887: Premsel loss 0.1738, acc 0.9292 (0.9527 / 0.9057) Evaluation 53: Premsel loss 0.1717, acc 0.9300 (0.9453 / 0.9147) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_00.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz Training 54: 0 / 119216: Premsel loss 0.1793, acc 0.9270 (0.9520 / 0.9020) Training 54: 7000 / 119216: Premsel loss 0.2094, acc 0.9113 (0.9342 / 0.8884) Training 54: 14000 / 119216: Premsel loss 0.2084, acc 0.9111 (0.9352 / 0.8869) Training 54: 21000 / 119216: Premsel loss 0.2080, acc 0.9119 (0.9355 / 0.8884) Training 54: 28000 / 119216: Premsel loss 0.2128, acc 0.9097 (0.9296 / 0.8899) Training 54: 35000 / 119216: Premsel loss 0.2005, acc 0.9149 (0.9379 / 0.8918) Training 54: 42000 / 119216: Premsel loss 0.2105, acc 0.9114 (0.9281 / 0.8947) Training 54: 49000 / 119216: Premsel loss 0.2039, acc 0.9141 (0.9379 / 0.8903) Training 54: 56000 / 119216: Premsel loss 0.2098, acc 0.9107 (0.9328 / 0.8885) Training 54: 63000 / 119216: Premsel loss 0.2154, acc 0.9083 (0.9287 / 0.8879) Training 54: 70000 / 119216: Premsel loss 0.2084, acc 0.9120 (0.9349 / 0.8891) Training 54: 77000 / 119216: Premsel loss 0.2045, acc 0.9136 (0.9345 / 0.8928) Training 54: 84000 / 119216: Premsel loss 0.2036, acc 0.9149 (0.9350 / 0.8947) Training 54: 91000 / 119216: Premsel loss 0.2051, acc 0.9136 (0.9339 / 0.8933) Training 54: 98000 / 119216: Premsel loss 0.2065, acc 0.9125 (0.9323 / 0.8927) Training 54: 105000 / 119216: Premsel loss 0.1983, acc 0.9178 (0.9404 / 0.8952) Training 54: 112000 / 119216: Premsel loss 0.2070, acc 0.9124 (0.9326 / 0.8921) Training 54: 119000 / 119216: Premsel loss 0.2078, acc 0.9114 (0.9346 / 0.8882) Evaluation 54: Premsel loss 0.2048, acc 0.9140 (0.9513 / 0.8767) Training 55: 0 / 119216: Premsel loss 0.2028, acc 0.9150 (0.9379 / 0.8922) Training 55: 7000 / 119216: Premsel loss 0.1997, acc 0.9162 (0.9384 / 0.8941) Training 55: 14000 / 119216: Premsel loss 0.1983, acc 0.9165 (0.9361 / 0.8968) Training 55: 21000 / 119216: Premsel loss 0.2047, acc 0.9138 (0.9342 / 0.8934) Training 55: 28000 / 119216: Premsel loss 0.2066, acc 0.9136 (0.9325 / 0.8946) Training 55: 35000 / 119216: Premsel loss 0.2036, acc 0.9143 (0.9406 / 0.8880) Training 55: 42000 / 119216: Premsel loss 0.2069, acc 0.9124 (0.9354 / 0.8894) Training 55: 49000 / 119216: Premsel loss 0.2023, acc 0.9148 (0.9379 / 0.8917) Training 55: 56000 / 119216: Premsel loss 0.2084, acc 0.9117 (0.9329 / 0.8904) Training 55: 63000 / 119216: Premsel loss 0.2082, acc 0.9127 (0.9308 / 0.8947) Training 55: 70000 / 119216: Premsel loss 0.2096, acc 0.9106 (0.9359 / 0.8853) Training 55: 77000 / 119216: Premsel loss 0.2044, acc 0.9138 (0.9341 / 0.8935) Training 55: 84000 / 119216: Premsel loss 0.2050, acc 0.9135 (0.9332 / 0.8939) Training 55: 91000 / 119216: Premsel loss 0.2049, acc 0.9143 (0.9373 / 0.8914) Training 55: 98000 / 119216: Premsel loss 0.2029, acc 0.9155 (0.9369 / 0.8942) Training 55: 105000 / 119216: Premsel loss 0.2074, acc 0.9114 (0.9326 / 0.8903) Training 55: 112000 / 119216: Premsel loss 0.2029, acc 0.9152 (0.9374 / 0.8930) Training 55: 119000 / 119216: Premsel loss 0.2016, acc 0.9153 (0.9345 / 0.8960) Evaluation 55: Premsel loss 0.2055, acc 0.9140 (0.9520 / 0.8760) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_13.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_08.pkl.gz Training 56: 0 / 123903: Premsel loss 0.1966, acc 0.9180 (0.9422 / 0.8938) Training 56: 7000 / 123903: Premsel loss 0.1937, acc 0.9197 (0.9440 / 0.8954) Training 56: 14000 / 123903: Premsel loss 0.1966, acc 0.9187 (0.9424 / 0.8950) Training 56: 21000 / 123903: Premsel loss 0.1960, acc 0.9197 (0.9398 / 0.8996) Training 56: 28000 / 123903: Premsel loss 0.1928, acc 0.9206 (0.9432 / 0.8981) Training 56: 35000 / 123903: Premsel loss 0.1916, acc 0.9214 (0.9436 / 0.8991) Training 56: 42000 / 123903: Premsel loss 0.1891, acc 0.9229 (0.9468 / 0.8990) Training 56: 49000 / 123903: Premsel loss 0.1887, acc 0.9226 (0.9432 / 0.9019) Training 56: 56000 / 123903: Premsel loss 0.1881, acc 0.9241 (0.9473 / 0.9009) ^[Training 56: 63000 / 123903: Premsel loss 0.1875, acc 0.9230 (0.9446 / 0.9015) Training 56: 70000 / 123903: Premsel loss 0.1873, acc 0.9242 (0.9474 / 0.9011) Training 56: 77000 / 123903: Premsel loss 0.1867, acc 0.9238 (0.9465 / 0.9011) Training 56: 84000 / 123903: Premsel loss 0.1897, acc 0.9215 (0.9439 / 0.8990) Training 56: 91000 / 123903: Premsel loss 0.1862, acc 0.9240 (0.9465 / 0.9015) Training 56: 98000 / 123903: Premsel loss 0.1914, acc 0.9216 (0.9467 / 0.8966) Training 56: 105000 / 123903: Premsel loss 0.1963, acc 0.9190 (0.9404 / 0.8976) Training 56: 112000 / 123903: Premsel loss 0.1859, acc 0.9235 (0.9460 / 0.9011) Training 56: 119000 / 123903: Premsel loss 0.1894, acc 0.9230 (0.9484 / 0.8976) Evaluation 56: Premsel loss 0.1864, acc 0.9235 (0.9439 / 0.9030) Training 57: 0 / 123903: Premsel loss 0.1821, acc 0.9255 (0.9497 / 0.9012) Training 57: 7000 / 123903: Premsel loss 0.1858, acc 0.9243 (0.9487 / 0.8999) Training 57: 14000 / 123903: Premsel loss 0.1893, acc 0.9230 (0.9465 / 0.8994) Training 57: 21000 / 123903: Premsel loss 0.1891, acc 0.9222 (0.9458 / 0.8986) Training 57: 28000 / 123903: Premsel loss 0.1834, acc 0.9251 (0.9443 / 0.9059) Training 57: 35000 / 123903: Premsel loss 0.1873, acc 0.9237 (0.9471 / 0.9004) Training 57: 42000 / 123903: Premsel loss 0.1846, acc 0.9253 (0.9499 / 0.9007) Training 57: 49000 / 123903: Premsel loss 0.1870, acc 0.9238 (0.9494 / 0.8983) Training 57: 56000 / 123903: Premsel loss 0.1888, acc 0.9234 (0.9474 / 0.8993) Training 57: 63000 / 123903: Premsel loss 0.1823, acc 0.9264 (0.9500 / 0.9028) Training 57: 70000 / 123903: Premsel loss 0.1873, acc 0.9236 (0.9485 / 0.8987) Training 57: 77000 / 123903: Premsel loss 0.1862, acc 0.9248 (0.9487 / 0.9010) Training 57: 84000 / 123903: Premsel loss 0.1825, acc 0.9253 (0.9491 / 0.9015) Training 57: 91000 / 123903: Premsel loss 0.1833, acc 0.9254 (0.9483 / 0.9025) Training 57: 98000 / 123903: Premsel loss 0.1875, acc 0.9235 (0.9469 / 0.9001) Training 57: 105000 / 123903: Premsel loss 0.1924, acc 0.9214 (0.9439 / 0.8989) Training 57: 112000 / 123903: Premsel loss 0.1825, acc 0.9261 (0.9473 / 0.9048) Training 57: 119000 / 123903: Premsel loss 0.1858, acc 0.9239 (0.9459 / 0.9019) Evaluation 57: Premsel loss 0.1861, acc 0.9246 (0.9682 / 0.8810) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_15.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz Training 58: 0 / 123673: Premsel loss 0.1852, acc 0.9240 (0.9399 / 0.9081) Training 58: 7000 / 123673: Premsel loss 0.1989, acc 0.9174 (0.9410 / 0.8938) Training 58: 14000 / 123673: Premsel loss 0.1959, acc 0.9186 (0.9404 / 0.8968) Training 58: 21000 / 123673: Premsel loss 0.1946, acc 0.9194 (0.9416 / 0.8973) Training 58: 28000 / 123673: Premsel loss 0.1953, acc 0.9193 (0.9396 / 0.8990) Training 58: 35000 / 123673: Premsel loss 0.1965, acc 0.9187 (0.9409 / 0.8965) Training 58: 42000 / 123673: Premsel loss 0.1966, acc 0.9185 (0.9422 / 0.8949) Training 58: 49000 / 123673: Premsel loss 0.1947, acc 0.9194 (0.9435 / 0.8953) Training 58: 56000 / 123673: Premsel loss 0.1928, acc 0.9195 (0.9425 / 0.8964) Training 58: 63000 / 123673: Premsel loss 0.1974, acc 0.9188 (0.9438 / 0.8939) Training 58: 70000 / 123673: Premsel loss 0.1925, acc 0.9202 (0.9394 / 0.9010) Training 58: 77000 / 123673: Premsel loss 0.1943, acc 0.9193 (0.9419 / 0.8967) Training 58: 84000 / 123673: Premsel loss 0.1962, acc 0.9182 (0.9412 / 0.8951) Training 58: 91000 / 123673: Premsel loss 0.1952, acc 0.9194 (0.9385 / 0.9002) Training 58: 98000 / 123673: Premsel loss 0.1940, acc 0.9194 (0.9447 / 0.8941) Training 58: 105000 / 123673: Premsel loss 0.1918, acc 0.9200 (0.9415 / 0.8985) Training 58: 112000 / 123673: Premsel loss 0.2006, acc 0.9166 (0.9451 / 0.8881) Training 58: 119000 / 123673: Premsel loss 0.1892, acc 0.9217 (0.9417 / 0.9017) Evaluation 58: Premsel loss 0.1915, acc 0.9205 (0.9507 / 0.8903) Training 59: 0 / 123673: Premsel loss 0.1978, acc 0.9181 (0.9382 / 0.8980) Training 59: 7000 / 123673: Premsel loss 0.1915, acc 0.9205 (0.9432 / 0.8978) Training 59: 14000 / 123673: Premsel loss 0.1906, acc 0.9208 (0.9443 / 0.8973) Training 59: 21000 / 123673: Premsel loss 0.1983, acc 0.9171 (0.9368 / 0.8974) Training 59: 28000 / 123673: Premsel loss 0.1934, acc 0.9192 (0.9438 / 0.8947) Training 59: 35000 / 123673: Premsel loss 0.1954, acc 0.9184 (0.9416 / 0.8952) Training 59: 42000 / 123673: Premsel loss 0.1939, acc 0.9195 (0.9459 / 0.8931) Training 59: 49000 / 123673: Premsel loss 0.1911, acc 0.9210 (0.9448 / 0.8972) Training 59: 56000 / 123673: Premsel loss 0.1944, acc 0.9183 (0.9405 / 0.8961) Training 59: 63000 / 123673: Premsel loss 0.1936, acc 0.9199 (0.9456 / 0.8942) Training 59: 70000 / 123673: Premsel loss 0.1946, acc 0.9183 (0.9440 / 0.8927) Training 59: 77000 / 123673: Premsel loss 0.1880, acc 0.9224 (0.9463 / 0.8985) Training 59: 84000 / 123673: Premsel loss 0.1913, acc 0.9207 (0.9449 / 0.8965) Training 59: 91000 / 123673: Premsel loss 0.1964, acc 0.9185 (0.9416 / 0.8954) Training 59: 98000 / 123673: Premsel loss 0.1955, acc 0.9177 (0.9420 / 0.8933) Training 59: 105000 / 123673: Premsel loss 0.1954, acc 0.9185 (0.9418 / 0.8952) Training 59: 112000 / 123673: Premsel loss 0.1910, acc 0.9213 (0.9455 / 0.8971) Training 59: 119000 / 123673: Premsel loss 0.1912, acc 0.9214 (0.9427 / 0.9002) Evaluation 59: Premsel loss 0.1945, acc 0.9190 (0.9620 / 0.8760) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_74avg-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop01-VHSLCAXPh+lgb-d50-l900-e0.15loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_03.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo2___bb_preds__160___out1.pkl.gz Training 60: 0 / 107819: Premsel loss 0.1980, acc 0.9175 (0.9447 / 0.8904) Training 60: 7000 / 107819: Premsel loss 0.1868, acc 0.9233 (0.9465 / 0.9001) Training 60: 14000 / 107819: Premsel loss 0.1847, acc 0.9241 (0.9480 / 0.9003) Training 60: 21000 / 107819: Premsel loss 0.1880, acc 0.9225 (0.9445 / 0.9005) Training 60: 28000 / 107819: Premsel loss 0.1918, acc 0.9208 (0.9443 / 0.8973) Training 60: 35000 / 107819: Premsel loss 0.1847, acc 0.9244 (0.9452 / 0.9036) Training 60: 42000 / 107819: Premsel loss 0.1870, acc 0.9228 (0.9473 / 0.8983) Training 60: 49000 / 107819: Premsel loss 0.1821, acc 0.9252 (0.9468 / 0.9036) Training 60: 56000 / 107819: Premsel loss 0.1915, acc 0.9208 (0.9441 / 0.8975) Training 60: 63000 / 107819: Premsel loss 0.1842, acc 0.9246 (0.9464 / 0.9028) Training 60: 70000 / 107819: Premsel loss 0.1842, acc 0.9240 (0.9458 / 0.9022) Training 60: 77000 / 107819: Premsel loss 0.1852, acc 0.9243 (0.9482 / 0.9004) Training 60: 84000 / 107819: Premsel loss 0.1898, acc 0.9217 (0.9458 / 0.8975) Training 60: 91000 / 107819: Premsel loss 0.1894, acc 0.9222 (0.9438 / 0.9005) Training 60: 98000 / 107819: Premsel loss 0.1875, acc 0.9235 (0.9467 / 0.9003) Training 60: 105000 / 107819: Premsel loss 0.1816, acc 0.9258 (0.9476 / 0.9039) Evaluation 60: Premsel loss 0.1852, acc 0.9243 (0.9588 / 0.8898) Training 61: 0 / 107819: Premsel loss 0.1874, acc 0.9228 (0.9435 / 0.9022) Training 61: 7000 / 107819: Premsel loss 0.1806, acc 0.9264 (0.9486 / 0.9042) Training 61: 14000 / 107819: Premsel loss 0.1813, acc 0.9261 (0.9500 / 0.9023) Training 61: 21000 / 107819: Premsel loss 0.1837, acc 0.9250 (0.9491 / 0.9010) Training 61: 28000 / 107819: Premsel loss 0.1807, acc 0.9261 (0.9521 / 0.9001) Training 61: 35000 / 107819: Premsel loss 0.1817, acc 0.9260 (0.9492 / 0.9028) Training 61: 42000 / 107819: Premsel loss 0.1818, acc 0.9255 (0.9475 / 0.9036) Training 61: 49000 / 107819: Premsel loss 0.1860, acc 0.9241 (0.9470 / 0.9012) Training 61: 56000 / 107819: Premsel loss 0.1804, acc 0.9267 (0.9509 / 0.9025) Training 61: 63000 / 107819: Premsel loss 0.1811, acc 0.9260 (0.9481 / 0.9040) Training 61: 70000 / 107819: Premsel loss 0.1796, acc 0.9266 (0.9507 / 0.9025) Training 61: 77000 / 107819: Premsel loss 0.1872, acc 0.9223 (0.9436 / 0.9010) Training 61: 84000 / 107819: Premsel loss 0.1893, acc 0.9225 (0.9425 / 0.9025) Training 61: 91000 / 107819: Premsel loss 0.1800, acc 0.9265 (0.9481 / 0.9049) Training 61: 98000 / 107819: Premsel loss 0.1854, acc 0.9234 (0.9449 / 0.9019) Training 61: 105000 / 107819: Premsel loss 0.1846, acc 0.9235 (0.9421 / 0.9048) Evaluation 61: Premsel loss 0.1810, acc 0.9259 (0.9516 / 0.9002) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_21.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+solo___out1.pkl.gz Training 62: 0 / 117222: Premsel loss 0.1841, acc 0.9245 (0.9492 / 0.8998) Training 62: 7000 / 117222: Premsel loss 0.1959, acc 0.9187 (0.9424 / 0.8949) Training 62: 14000 / 117222: Premsel loss 0.1962, acc 0.9184 (0.9415 / 0.8953) Training 62: 21000 / 117222: Premsel loss 0.1932, acc 0.9203 (0.9465 / 0.8941) Training 62: 28000 / 117222: Premsel loss 0.1909, acc 0.9214 (0.9455 / 0.8974) Training 62: 35000 / 117222: Premsel loss 0.1933, acc 0.9198 (0.9438 / 0.8959) Training 62: 42000 / 117222: Premsel loss 0.1894, acc 0.9221 (0.9465 / 0.8978) Training 62: 49000 / 117222: Premsel loss 0.1912, acc 0.9204 (0.9430 / 0.8978) Training 62: 56000 / 117222: Premsel loss 0.1866, acc 0.9229 (0.9460 / 0.8999) Training 62: 63000 / 117222: Premsel loss 0.1911, acc 0.9213 (0.9461 / 0.8965) Training 62: 70000 / 117222: Premsel loss 0.1896, acc 0.9200 (0.9449 / 0.8951) Training 62: 77000 / 117222: Premsel loss 0.1946, acc 0.9191 (0.9462 / 0.8920) Training 62: 84000 / 117222: Premsel loss 0.1860, acc 0.9231 (0.9463 / 0.8999) Training 62: 91000 / 117222: Premsel loss 0.1915, acc 0.9205 (0.9419 / 0.8990) Training 62: 98000 / 117222: Premsel loss 0.1881, acc 0.9223 (0.9460 / 0.8987) Training 62: 105000 / 117222: Premsel loss 0.1918, acc 0.9210 (0.9447 / 0.8973) Training 62: 112000 / 117222: Premsel loss 0.1932, acc 0.9194 (0.9418 / 0.8970) Evaluation 62: Premsel loss 0.1886, acc 0.9221 (0.9453 / 0.8990) Training 63: 0 / 117222: Premsel loss 0.1914, acc 0.9200 (0.9420 / 0.8980) Training 63: 7000 / 117222: Premsel loss 0.1870, acc 0.9221 (0.9465 / 0.8978) Training 63: 14000 / 117222: Premsel loss 0.1844, acc 0.9243 (0.9457 / 0.9029) Training 63: 21000 / 117222: Premsel loss 0.1935, acc 0.9195 (0.9429 / 0.8962) Training 63: 28000 / 117222: Premsel loss 0.1854, acc 0.9237 (0.9457 / 0.9018) Training 63: 35000 / 117222: Premsel loss 0.1885, acc 0.9219 (0.9415 / 0.9024) Training 63: 42000 / 117222: Premsel loss 0.1892, acc 0.9213 (0.9440 / 0.8985) Training 63: 49000 / 117222: Premsel loss 0.1924, acc 0.9202 (0.9465 / 0.8938) Training 63: 56000 / 117222: Premsel loss 0.1877, acc 0.9226 (0.9467 / 0.8986) Training 63: 63000 / 117222: Premsel loss 0.1891, acc 0.9212 (0.9404 / 0.9019) Training 63: 70000 / 117222: Premsel loss 0.1883, acc 0.9218 (0.9436 / 0.9000) Training 63: 77000 / 117222: Premsel loss 0.1903, acc 0.9215 (0.9419 / 0.9011) Training 63: 84000 / 117222: Premsel loss 0.1873, acc 0.9229 (0.9475 / 0.8983) Training 63: 91000 / 117222: Premsel loss 0.1874, acc 0.9233 (0.9449 / 0.9017) Training 63: 98000 / 117222: Premsel loss 0.1896, acc 0.9217 (0.9430 / 0.9005) Training 63: 105000 / 117222: Premsel loss 0.1954, acc 0.9193 (0.9415 / 0.8970) Training 63: 112000 / 117222: Premsel loss 0.1925, acc 0.9210 (0.9430 / 0.8990) Evaluation 63: Premsel loss 0.1854, acc 0.9237 (0.9537 / 0.8936) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_26.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_23.pkl.gz Training 64: 0 / 99777: Premsel loss 0.1870, acc 0.9230 (0.9476 / 0.8984) Training 64: 7000 / 99777: Premsel loss 0.1890, acc 0.9235 (0.9504 / 0.8967) Training 64: 14000 / 99777: Premsel loss 0.1925, acc 0.9216 (0.9456 / 0.8977) Training 64: 21000 / 99777: Premsel loss 0.1865, acc 0.9245 (0.9483 / 0.9006) Training 64: 28000 / 99777: Premsel loss 0.1917, acc 0.9221 (0.9451 / 0.8992) Training 64: 35000 / 99777: Premsel loss 0.1898, acc 0.9223 (0.9485 / 0.8961) Training 64: 42000 / 99777: Premsel loss 0.1877, acc 0.9243 (0.9485 / 0.9002) Training 64: 49000 / 99777: Premsel loss 0.1852, acc 0.9249 (0.9472 / 0.9026) Training 64: 56000 / 99777: Premsel loss 0.1867, acc 0.9244 (0.9492 / 0.8996) Training 64: 63000 / 99777: Premsel loss 0.1910, acc 0.9217 (0.9463 / 0.8971) Training 64: 70000 / 99777: Premsel loss 0.1886, acc 0.9226 (0.9465 / 0.8986) Training 64: 77000 / 99777: Premsel loss 0.1874, acc 0.9245 (0.9506 / 0.8983) Training 64: 84000 / 99777: Premsel loss 0.1818, acc 0.9273 (0.9519 / 0.9026) Training 64: 91000 / 99777: Premsel loss 0.1893, acc 0.9229 (0.9492 / 0.8965) Training 64: 98000 / 99777: Premsel loss 0.1884, acc 0.9236 (0.9489 / 0.8983) Evaluation 64: Premsel loss 0.1856, acc 0.9250 (0.9507 / 0.8994) Training 65: 0 / 99777: Premsel loss 0.1845, acc 0.9261 (0.9484 / 0.9037) Training 65: 7000 / 99777: Premsel loss 0.1799, acc 0.9273 (0.9509 / 0.9037) Training 65: 14000 / 99777: Premsel loss 0.1865, acc 0.9244 (0.9500 / 0.8989) Training 65: 21000 / 99777: Premsel loss 0.1854, acc 0.9253 (0.9508 / 0.8999) Training 65: 28000 / 99777: Premsel loss 0.1821, acc 0.9266 (0.9521 / 0.9010) Training 65: 35000 / 99777: Premsel loss 0.1820, acc 0.9271 (0.9505 / 0.9036) Training 65: 42000 / 99777: Premsel loss 0.1820, acc 0.9270 (0.9518 / 0.9021) Training 65: 49000 / 99777: Premsel loss 0.1863, acc 0.9241 (0.9488 / 0.8994) Training 65: 56000 / 99777: Premsel loss 0.1831, acc 0.9266 (0.9512 / 0.9020) Training 65: 63000 / 99777: Premsel loss 0.1852, acc 0.9251 (0.9484 / 0.9019) Training 65: 70000 / 99777: Premsel loss 0.1859, acc 0.9243 (0.9482 / 0.9004) Training 65: 77000 / 99777: Premsel loss 0.1835, acc 0.9257 (0.9469 / 0.9045) Training 65: 84000 / 99777: Premsel loss 0.1827, acc 0.9259 (0.9489 / 0.9028) Training 65: 91000 / 99777: Premsel loss 0.1826, acc 0.9257 (0.9501 / 0.9014) Training 65: 98000 / 99777: Premsel loss 0.1840, acc 0.9255 (0.9482 / 0.9029) Evaluation 65: Premsel loss 0.1823, acc 0.9262 (0.9363 / 0.9160) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_02.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__192___out1.pkl.gz Training 66: 0 / 74616: Premsel loss 0.1790, acc 0.9290 (0.9529 / 0.9051) Training 66: 7000 / 74616: Premsel loss 0.2018, acc 0.9164 (0.9353 / 0.8975) Training 66: 14000 / 74616: Premsel loss 0.2079, acc 0.9135 (0.9363 / 0.8907) Training 66: 21000 / 74616: Premsel loss 0.2132, acc 0.9100 (0.9335 / 0.8864) Training 66: 28000 / 74616: Premsel loss 0.2051, acc 0.9135 (0.9389 / 0.8880) Training 66: 35000 / 74616: Premsel loss 0.2004, acc 0.9169 (0.9413 / 0.8924) Training 66: 42000 / 74616: Premsel loss 0.1998, acc 0.9169 (0.9388 / 0.8950) Training 66: 49000 / 74616: Premsel loss 0.1954, acc 0.9181 (0.9415 / 0.8948) Training 66: 56000 / 74616: Premsel loss 0.2031, acc 0.9155 (0.9375 / 0.8934) Training 66: 63000 / 74616: Premsel loss 0.2052, acc 0.9153 (0.9409 / 0.8896) Training 66: 70000 / 74616: Premsel loss 0.1924, acc 0.9210 (0.9424 / 0.8996) Evaluation 66: Premsel loss 0.1997, acc 0.9162 (0.9163 / 0.9161) Training 67: 0 / 74616: Premsel loss 0.2062, acc 0.9141 (0.9390 / 0.8893) Training 67: 7000 / 74616: Premsel loss 0.2028, acc 0.9155 (0.9352 / 0.8957) Training 67: 14000 / 74616: Premsel loss 0.1943, acc 0.9201 (0.9406 / 0.8996) Training 67: 21000 / 74616: Premsel loss 0.1987, acc 0.9185 (0.9404 / 0.8965) Training 67: 28000 / 74616: Premsel loss 0.2027, acc 0.9154 (0.9372 / 0.8936) Training 67: 35000 / 74616: Premsel loss 0.1991, acc 0.9170 (0.9408 / 0.8932) Training 67: 42000 / 74616: Premsel loss 0.2029, acc 0.9152 (0.9390 / 0.8913) Training 67: 49000 / 74616: Premsel loss 0.1985, acc 0.9170 (0.9421 / 0.8919) Training 67: 56000 / 74616: Premsel loss 0.1985, acc 0.9180 (0.9404 / 0.8956) Training 67: 63000 / 74616: Premsel loss 0.1974, acc 0.9180 (0.9377 / 0.8984) Training 67: 70000 / 74616: Premsel loss 0.2019, acc 0.9154 (0.9385 / 0.8923) Evaluation 67: Premsel loss 0.1920, acc 0.9212 (0.9553 / 0.8871) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l700-e0.20+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_05.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.05___out1.pkl.gz Training 68: 0 / 133535: Premsel loss 0.1999, acc 0.9167 (0.9360 / 0.8974) Training 68: 7000 / 133535: Premsel loss 0.2161, acc 0.9090 (0.9359 / 0.8820) Training 68: 14000 / 133535: Premsel loss 0.2124, acc 0.9100 (0.9319 / 0.8881) Training 68: 21000 / 133535: Premsel loss 0.2300, acc 0.9019 (0.9348 / 0.8690) Training 68: 28000 / 133535: Premsel loss 0.2198, acc 0.9072 (0.9338 / 0.8807) Training 68: 35000 / 133535: Premsel loss 0.2161, acc 0.9089 (0.9348 / 0.8831) Training 68: 42000 / 133535: Premsel loss 0.2221, acc 0.9051 (0.9339 / 0.8763) Training 68: 49000 / 133535: Premsel loss 0.2208, acc 0.9070 (0.9318 / 0.8822) Training 68: 56000 / 133535: Premsel loss 0.2195, acc 0.9069 (0.9301 / 0.8838) Training 68: 63000 / 133535: Premsel loss 0.2213, acc 0.9061 (0.9334 / 0.8788) Training 68: 70000 / 133535: Premsel loss 0.2185, acc 0.9066 (0.9296 / 0.8835) Training 68: 77000 / 133535: Premsel loss 0.2180, acc 0.9078 (0.9337 / 0.8819) Training 68: 84000 / 133535: Premsel loss 0.2311, acc 0.9002 (0.9302 / 0.8703) Training 68: 91000 / 133535: Premsel loss 0.2247, acc 0.9043 (0.9279 / 0.8807) Training 68: 98000 / 133535: Premsel loss 0.2186, acc 0.9069 (0.9328 / 0.8811) Training 68: 105000 / 133535: Premsel loss 0.2208, acc 0.9061 (0.9308 / 0.8815) Training 68: 112000 / 133535: Premsel loss 0.2190, acc 0.9071 (0.9349 / 0.8793) Training 68: 119000 / 133535: Premsel loss 0.2126, acc 0.9101 (0.9349 / 0.8853) Training 68: 126000 / 133535: Premsel loss 0.2213, acc 0.9054 (0.9302 / 0.8806) Training 68: 133000 / 133535: Premsel loss 0.2177, acc 0.9076 (0.9315 / 0.8838) Evaluation 68: Premsel loss 0.2126, acc 0.9097 (0.9355 / 0.8839) Training 69: 0 / 133535: Premsel loss 0.2128, acc 0.9105 (0.9334 / 0.8876) Training 69: 7000 / 133535: Premsel loss 0.2146, acc 0.9091 (0.9375 / 0.8807) Training 69: 14000 / 133535: Premsel loss 0.2160, acc 0.9091 (0.9367 / 0.8815) Training 69: 21000 / 133535: Premsel loss 0.2143, acc 0.9096 (0.9348 / 0.8843) Training 69: 28000 / 133535: Premsel loss 0.2225, acc 0.9048 (0.9250 / 0.8847) Training 69: 35000 / 133535: Premsel loss 0.2069, acc 0.9132 (0.9379 / 0.8885) Training 69: 42000 / 133535: Premsel loss 0.2139, acc 0.9088 (0.9347 / 0.8829) Training 69: 49000 / 133535: Premsel loss 0.2150, acc 0.9084 (0.9375 / 0.8794) Training 69: 56000 / 133535: Premsel loss 0.2129, acc 0.9096 (0.9345 / 0.8848) Training 69: 63000 / 133535: Premsel loss 0.2186, acc 0.9074 (0.9310 / 0.8838) Training 69: 70000 / 133535: Premsel loss 0.2151, acc 0.9085 (0.9368 / 0.8801) Training 69: 77000 / 133535: Premsel loss 0.2143, acc 0.9096 (0.9376 / 0.8817) Training 69: 84000 / 133535: Premsel loss 0.2165, acc 0.9081 (0.9375 / 0.8788) Training 69: 91000 / 133535: Premsel loss 0.2189, acc 0.9070 (0.9310 / 0.8829) Training 69: 98000 / 133535: Premsel loss 0.2156, acc 0.9079 (0.9331 / 0.8827) Training 69: 105000 / 133535: Premsel loss 0.2194, acc 0.9074 (0.9360 / 0.8789) Training 69: 112000 / 133535: Premsel loss 0.2160, acc 0.9079 (0.9337 / 0.8822) Training 69: 119000 / 133535: Premsel loss 0.2202, acc 0.9067 (0.9333 / 0.8801) Training 69: 126000 / 133535: Premsel loss 0.2173, acc 0.9076 (0.9325 / 0.8827) Training 69: 133000 / 133535: Premsel loss 0.2182, acc 0.9075 (0.9309 / 0.8841) Evaluation 69: Premsel loss 0.2186, acc 0.9070 (0.9073 / 0.9066) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_16.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz Training 70: 0 / 101389: Premsel loss 0.2126, acc 0.9109 (0.9414 / 0.8804) Training 70: 7000 / 101389: Premsel loss 0.2059, acc 0.9137 (0.9364 / 0.8909) Training 70: 14000 / 101389: Premsel loss 0.2111, acc 0.9104 (0.9359 / 0.8850) Training 70: 21000 / 101389: Premsel loss 0.2112, acc 0.9106 (0.9366 / 0.8846) Training 70: 28000 / 101389: Premsel loss 0.2198, acc 0.9065 (0.9315 / 0.8814) Training 70: 35000 / 101389: Premsel loss 0.2080, acc 0.9127 (0.9354 / 0.8901) Training 70: 42000 / 101389: Premsel loss 0.2064, acc 0.9129 (0.9352 / 0.8907) Training 70: 49000 / 101389: Premsel loss 0.2118, acc 0.9109 (0.9318 / 0.8900) Training 70: 56000 / 101389: Premsel loss 0.2119, acc 0.9103 (0.9343 / 0.8863) Training 70: 63000 / 101389: Premsel loss 0.2031, acc 0.9151 (0.9390 / 0.8912) Training 70: 70000 / 101389: Premsel loss 0.2136, acc 0.9093 (0.9342 / 0.8844) Training 70: 77000 / 101389: Premsel loss 0.2061, acc 0.9139 (0.9384 / 0.8895) Training 70: 84000 / 101389: Premsel loss 0.2082, acc 0.9129 (0.9374 / 0.8884) Training 70: 91000 / 101389: Premsel loss 0.2020, acc 0.9150 (0.9412 / 0.8889) Training 70: 98000 / 101389: Premsel loss 0.2130, acc 0.9096 (0.9290 / 0.8902) Evaluation 70: Premsel loss 0.2073, acc 0.9130 (0.9376 / 0.8884) Training 71: 0 / 101389: Premsel loss 0.2034, acc 0.9149 (0.9418 / 0.8879) Training 71: 7000 / 101389: Premsel loss 0.2053, acc 0.9135 (0.9377 / 0.8893) Training 71: 14000 / 101389: Premsel loss 0.2045, acc 0.9146 (0.9359 / 0.8933) Training 71: 21000 / 101389: Premsel loss 0.2040, acc 0.9150 (0.9360 / 0.8940) Training 71: 28000 / 101389: Premsel loss 0.2127, acc 0.9094 (0.9279 / 0.8909) Training 71: 35000 / 101389: Premsel loss 0.2122, acc 0.9113 (0.9336 / 0.8891) Training 71: 42000 / 101389: Premsel loss 0.2132, acc 0.9113 (0.9344 / 0.8881) Training 71: 49000 / 101389: Premsel loss 0.1986, acc 0.9176 (0.9376 / 0.8976) Training 71: 56000 / 101389: Premsel loss 0.2135, acc 0.9095 (0.9348 / 0.8842) Training 71: 63000 / 101389: Premsel loss 0.2110, acc 0.9112 (0.9351 / 0.8873) Training 71: 70000 / 101389: Premsel loss 0.2109, acc 0.9108 (0.9317 / 0.8899) Training 71: 77000 / 101389: Premsel loss 0.2091, acc 0.9120 (0.9352 / 0.8887) Training 71: 84000 / 101389: Premsel loss 0.2099, acc 0.9114 (0.9347 / 0.8881) Training 71: 91000 / 101389: Premsel loss 0.2030, acc 0.9151 (0.9384 / 0.8919) Training 71: 98000 / 101389: Premsel loss 0.2050, acc 0.9150 (0.9373 / 0.8927) Evaluation 71: Premsel loss 0.2104, acc 0.9110 (0.9270 / 0.8949) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_12.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_27.pkl.gz Training 72: 0 / 92456: Premsel loss 0.2132, acc 0.9095 (0.9378 / 0.8812) Training 72: 7000 / 92456: Premsel loss 0.1857, acc 0.9242 (0.9442 / 0.9042) Training 72: 14000 / 92456: Premsel loss 0.1832, acc 0.9252 (0.9486 / 0.9018) Training 72: 21000 / 92456: Premsel loss 0.1821, acc 0.9269 (0.9501 / 0.9038) Training 72: 28000 / 92456: Premsel loss 0.1794, acc 0.9276 (0.9490 / 0.9063) Training 72: 35000 / 92456: Premsel loss 0.1904, acc 0.9229 (0.9451 / 0.9007) Training 72: 42000 / 92456: Premsel loss 0.1851, acc 0.9242 (0.9484 / 0.9000) Training 72: 49000 / 92456: Premsel loss 0.1869, acc 0.9235 (0.9495 / 0.8974) Training 72: 56000 / 92456: Premsel loss 0.1869, acc 0.9235 (0.9449 / 0.9021) Training 72: 63000 / 92456: Premsel loss 0.1833, acc 0.9250 (0.9480 / 0.9020) Training 72: 70000 / 92456: Premsel loss 0.1819, acc 0.9259 (0.9489 / 0.9029) Training 72: 77000 / 92456: Premsel loss 0.1809, acc 0.9261 (0.9499 / 0.9023) Training 72: 84000 / 92456: Premsel loss 0.1790, acc 0.9280 (0.9503 / 0.9056) Training 72: 91000 / 92456: Premsel loss 0.1806, acc 0.9265 (0.9499 / 0.9031) Evaluation 72: Premsel loss 0.1804, acc 0.9267 (0.9493 / 0.9041) Training 73: 0 / 92456: Premsel loss 0.1833, acc 0.9259 (0.9504 / 0.9015) Training 73: 7000 / 92456: Premsel loss 0.1773, acc 0.9289 (0.9516 / 0.9061) Training 73: 14000 / 92456: Premsel loss 0.1825, acc 0.9253 (0.9492 / 0.9015) Training 73: 21000 / 92456: Premsel loss 0.1861, acc 0.9242 (0.9491 / 0.8993) Training 73: 28000 / 92456: Premsel loss 0.1744, acc 0.9302 (0.9538 / 0.9066) Training 73: 35000 / 92456: Premsel loss 0.1836, acc 0.9256 (0.9472 / 0.9040) Training 73: 42000 / 92456: Premsel loss 0.1818, acc 0.9270 (0.9492 / 0.9048) Training 73: 49000 / 92456: Premsel loss 0.1810, acc 0.9267 (0.9527 / 0.9007) Training 73: 56000 / 92456: Premsel loss 0.1786, acc 0.9279 (0.9501 / 0.9056) Training 73: 63000 / 92456: Premsel loss 0.1784, acc 0.9280 (0.9518 / 0.9042) Training 73: 70000 / 92456: Premsel loss 0.1779, acc 0.9281 (0.9495 / 0.9066) Training 73: 77000 / 92456: Premsel loss 0.1846, acc 0.9256 (0.9487 / 0.9025) Training 73: 84000 / 92456: Premsel loss 0.1779, acc 0.9281 (0.9512 / 0.9051) Training 73: 91000 / 92456: Premsel loss 0.1803, acc 0.9269 (0.9505 / 0.9032) Evaluation 73: Premsel loss 0.1774, acc 0.9282 (0.9451 / 0.9114) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_22.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-solo___out1.pkl.gz Training 74: 0 / 93372: Premsel loss 0.1827, acc 0.9259 (0.9493 / 0.9026) Training 74: 7000 / 93372: Premsel loss 0.1890, acc 0.9223 (0.9460 / 0.8987) Training 74: 14000 / 93372: Premsel loss 0.1854, acc 0.9239 (0.9465 / 0.9014) Training 74: 21000 / 93372: Premsel loss 0.1871, acc 0.9233 (0.9475 / 0.8992) Training 74: 28000 / 93372: Premsel loss 0.1834, acc 0.9250 (0.9473 / 0.9028) Training 74: 35000 / 93372: Premsel loss 0.1845, acc 0.9247 (0.9437 / 0.9058) Training 74: 42000 / 93372: Premsel loss 0.1870, acc 0.9235 (0.9516 / 0.8955) Training 74: 49000 / 93372: Premsel loss 0.1859, acc 0.9243 (0.9479 / 0.9006) Training 74: 56000 / 93372: Premsel loss 0.1826, acc 0.9254 (0.9500 / 0.9008) Training 74: 63000 / 93372: Premsel loss 0.1859, acc 0.9229 (0.9446 / 0.9012) Training 74: 70000 / 93372: Premsel loss 0.1842, acc 0.9249 (0.9477 / 0.9021) Training 74: 77000 / 93372: Premsel loss 0.1873, acc 0.9228 (0.9464 / 0.8993) Training 74: 84000 / 93372: Premsel loss 0.1869, acc 0.9237 (0.9461 / 0.9012) Training 74: 91000 / 93372: Premsel loss 0.1923, acc 0.9199 (0.9416 / 0.8983) Evaluation 74: Premsel loss 0.1813, acc 0.9259 (0.9530 / 0.8987) Training 75: 0 / 93372: Premsel loss 0.1815, acc 0.9254 (0.9490 / 0.9019) Training 75: 7000 / 93372: Premsel loss 0.1841, acc 0.9251 (0.9500 / 0.9002) Training 75: 14000 / 93372: Premsel loss 0.1893, acc 0.9217 (0.9452 / 0.8982) Training 75: 21000 / 93372: Premsel loss 0.1814, acc 0.9259 (0.9501 / 0.9018) Training 75: 28000 / 93372: Premsel loss 0.1855, acc 0.9237 (0.9463 / 0.9011) Training 75: 35000 / 93372: Premsel loss 0.1889, acc 0.9227 (0.9433 / 0.9021) Training 75: 42000 / 93372: Premsel loss 0.1886, acc 0.9226 (0.9475 / 0.8977) Training 75: 49000 / 93372: Premsel loss 0.1876, acc 0.9231 (0.9469 / 0.8993) Training 75: 56000 / 93372: Premsel loss 0.1834, acc 0.9252 (0.9483 / 0.9021) Training 75: 63000 / 93372: Premsel loss 0.1861, acc 0.9233 (0.9483 / 0.8984) Training 75: 70000 / 93372: Premsel loss 0.1899, acc 0.9239 (0.9469 / 0.9008) Training 75: 77000 / 93372: Premsel loss 0.1835, acc 0.9256 (0.9490 / 0.9022) Training 75: 84000 / 93372: Premsel loss 0.1860, acc 0.9245 (0.9527 / 0.8963) Training 75: 91000 / 93372: Premsel loss 0.1802, acc 0.9270 (0.9537 / 0.9004) Evaluation 75: Premsel loss 0.1773, acc 0.9281 (0.9544 / 0.9017) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_01.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz Training 76: 0 / 77262: Premsel loss 0.1814, acc 0.9254 (0.9476 / 0.9031) Training 76: 7000 / 77262: Premsel loss 0.2080, acc 0.9121 (0.9340 / 0.8901) Training 76: 14000 / 77262: Premsel loss 0.2108, acc 0.9111 (0.9378 / 0.8845) Training 76: 21000 / 77262: Premsel loss 0.1980, acc 0.9179 (0.9417 / 0.8941) Training 76: 28000 / 77262: Premsel loss 0.2170, acc 0.9088 (0.9312 / 0.8864) Training 76: 35000 / 77262: Premsel loss 0.2061, acc 0.9141 (0.9391 / 0.8892) Training 76: 42000 / 77262: Premsel loss 0.2062, acc 0.9138 (0.9364 / 0.8913) Training 76: 49000 / 77262: Premsel loss 0.2092, acc 0.9124 (0.9408 / 0.8840) Training 76: 56000 / 77262: Premsel loss 0.2145, acc 0.9099 (0.9373 / 0.8825) Training 76: 63000 / 77262: Premsel loss 0.2161, acc 0.9086 (0.9367 / 0.8804) Training 76: 70000 / 77262: Premsel loss 0.2029, acc 0.9160 (0.9405 / 0.8915) Training 76: 77000 / 77262: Premsel loss 0.2167, acc 0.9079 (0.9298 / 0.8861) Evaluation 76: Premsel loss 0.2093, acc 0.9122 (0.9473 / 0.8771) Training 77: 0 / 77262: Premsel loss 0.2230, acc 0.9051 (0.9285 / 0.8817) Training 77: 7000 / 77262: Premsel loss 0.2088, acc 0.9123 (0.9398 / 0.8848) Training 77: 14000 / 77262: Premsel loss 0.2136, acc 0.9104 (0.9363 / 0.8845) Training 77: 21000 / 77262: Premsel loss 0.2065, acc 0.9136 (0.9405 / 0.8866) Training 77: 28000 / 77262: Premsel loss 0.2034, acc 0.9147 (0.9398 / 0.8897) Training 77: 35000 / 77262: Premsel loss 0.2057, acc 0.9142 (0.9413 / 0.8870) Training 77: 42000 / 77262: Premsel loss 0.2052, acc 0.9150 (0.9373 / 0.8927) Training 77: 49000 / 77262: Premsel loss 0.2010, acc 0.9160 (0.9403 / 0.8916) Training 77: 56000 / 77262: Premsel loss 0.2078, acc 0.9132 (0.9366 / 0.8898) Training 77: 63000 / 77262: Premsel loss 0.2080, acc 0.9127 (0.9401 / 0.8854) Training 77: 70000 / 77262: Premsel loss 0.2031, acc 0.9147 (0.9388 / 0.8906) Training 77: 77000 / 77262: Premsel loss 0.2003, acc 0.9172 (0.9440 / 0.8905) Evaluation 77: Premsel loss 0.1961, acc 0.9187 (0.9464 / 0.8910) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_14.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d50-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz Training 78: 0 / 123491: Premsel loss 0.2004, acc 0.9168 (0.9370 / 0.8966) Training 78: 7000 / 123491: Premsel loss 0.1981, acc 0.9177 (0.9414 / 0.8940) Training 78: 14000 / 123491: Premsel loss 0.2005, acc 0.9167 (0.9391 / 0.8942) Training 78: 21000 / 123491: Premsel loss 0.2028, acc 0.9152 (0.9401 / 0.8903) Training 78: 28000 / 123491: Premsel loss 0.1981, acc 0.9169 (0.9437 / 0.8902) Training 78: 35000 / 123491: Premsel loss 0.2091, acc 0.9127 (0.9362 / 0.8892) Training 78: 42000 / 123491: Premsel loss 0.1991, acc 0.9173 (0.9424 / 0.8922) Training 78: 49000 / 123491: Premsel loss 0.2015, acc 0.9165 (0.9387 / 0.8943) Training 78: 56000 / 123491: Premsel loss 0.1952, acc 0.9189 (0.9418 / 0.8961) Training 78: 63000 / 123491: Premsel loss 0.2092, acc 0.9115 (0.9352 / 0.8879) Training 78: 70000 / 123491: Premsel loss 0.2033, acc 0.9142 (0.9398 / 0.8886) Training 78: 77000 / 123491: Premsel loss 0.1995, acc 0.9173 (0.9420 / 0.8925) Training 78: 84000 / 123491: Premsel loss 0.2041, acc 0.9143 (0.9373 / 0.8914) Training 78: 91000 / 123491: Premsel loss 0.2032, acc 0.9149 (0.9377 / 0.8921) Training 78: 98000 / 123491: Premsel loss 0.1982, acc 0.9182 (0.9418 / 0.8947) Training 78: 105000 / 123491: Premsel loss 0.1986, acc 0.9176 (0.9426 / 0.8927) Training 78: 112000 / 123491: Premsel loss 0.2118, acc 0.9110 (0.9389 / 0.8831) Training 78: 119000 / 123491: Premsel loss 0.2043, acc 0.9142 (0.9386 / 0.8898) Evaluation 78: Premsel loss 0.2038, acc 0.9143 (0.9299 / 0.8986) Training 79: 0 / 123491: Premsel loss 0.2089, acc 0.9119 (0.9404 / 0.8833) Training 79: 7000 / 123491: Premsel loss 0.2013, acc 0.9156 (0.9391 / 0.8921) Training 79: 14000 / 123491: Premsel loss 0.1959, acc 0.9181 (0.9416 / 0.8946) Training 79: 21000 / 123491: Premsel loss 0.1975, acc 0.9176 (0.9385 / 0.8967) Training 79: 28000 / 123491: Premsel loss 0.2013, acc 0.9165 (0.9428 / 0.8901) Training 79: 35000 / 123491: Premsel loss 0.2002, acc 0.9163 (0.9432 / 0.8894) Training 79: 42000 / 123491: Premsel loss 0.1934, acc 0.9203 (0.9435 / 0.8970) Training 79: 49000 / 123491: Premsel loss 0.1990, acc 0.9176 (0.9421 / 0.8930) Training 79: 56000 / 123491: Premsel loss 0.2046, acc 0.9143 (0.9391 / 0.8895) Training 79: 63000 / 123491: Premsel loss 0.2020, acc 0.9156 (0.9365 / 0.8947) Training 79: 70000 / 123491: Premsel loss 0.2009, acc 0.9158 (0.9417 / 0.8899) Training 79: 77000 / 123491: Premsel loss 0.1981, acc 0.9171 (0.9399 / 0.8942) Training 79: 84000 / 123491: Premsel loss 0.1909, acc 0.9208 (0.9464 / 0.8952) Training 79: 91000 / 123491: Premsel loss 0.1988, acc 0.9168 (0.9439 / 0.8898) Training 79: 98000 / 123491: Premsel loss 0.1962, acc 0.9190 (0.9408 / 0.8972) Training 79: 105000 / 123491: Premsel loss 0.2046, acc 0.9140 (0.9371 / 0.8909) Training 79: 112000 / 123491: Premsel loss 0.1998, acc 0.9176 (0.9428 / 0.8923) Training 79: 119000 / 123491: Premsel loss 0.1985, acc 0.9168 (0.9386 / 0.8950) Evaluation 79: Premsel loss 0.2049, acc 0.9143 (0.9627 / 0.8658) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_17.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_19.pkl.gz Training 80: 0 / 123988: Premsel loss 0.1999, acc 0.9170 (0.9361 / 0.8980) Training 80: 7000 / 123988: Premsel loss 0.1980, acc 0.9177 (0.9401 / 0.8953) Training 80: 14000 / 123988: Premsel loss 0.1914, acc 0.9205 (0.9441 / 0.8969) Training 80: 21000 / 123988: Premsel loss 0.1943, acc 0.9195 (0.9433 / 0.8957) Training 80: 28000 / 123988: Premsel loss 0.1960, acc 0.9190 (0.9459 / 0.8921) Training 80: 35000 / 123988: Premsel loss 0.1892, acc 0.9220 (0.9455 / 0.8984) Training 80: 42000 / 123988: Premsel loss 0.1927, acc 0.9209 (0.9440 / 0.8979) Training 80: 49000 / 123988: Premsel loss 0.1836, acc 0.9237 (0.9472 / 0.9001) Training 80: 56000 / 123988: Premsel loss 0.1922, acc 0.9213 (0.9445 / 0.8981) Training 80: 63000 / 123988: Premsel loss 0.1944, acc 0.9195 (0.9445 / 0.8946) Training 80: 70000 / 123988: Premsel loss 0.1943, acc 0.9198 (0.9453 / 0.8942) Training 80: 77000 / 123988: Premsel loss 0.1891, acc 0.9225 (0.9446 / 0.9004) Training 80: 84000 / 123988: Premsel loss 0.2046, acc 0.9143 (0.9394 / 0.8892) Training 80: 91000 / 123988: Premsel loss 0.1919, acc 0.9210 (0.9438 / 0.8981) Training 80: 98000 / 123988: Premsel loss 0.2010, acc 0.9169 (0.9433 / 0.8904) Training 80: 105000 / 123988: Premsel loss 0.1949, acc 0.9195 (0.9440 / 0.8949) Training 80: 112000 / 123988: Premsel loss 0.2025, acc 0.9155 (0.9402 / 0.8908) Training 80: 119000 / 123988: Premsel loss 0.1945, acc 0.9198 (0.9441 / 0.8954) Evaluation 80: Premsel loss 0.1892, acc 0.9224 (0.9441 / 0.9006) Training 81: 0 / 123988: Premsel loss 0.2006, acc 0.9168 (0.9457 / 0.8878) Training 81: 7000 / 123988: Premsel loss 0.1925, acc 0.9209 (0.9480 / 0.8937) Training 81: 14000 / 123988: Premsel loss 0.1872, acc 0.9235 (0.9491 / 0.8980) Training 81: 21000 / 123988: Premsel loss 0.1918, acc 0.9211 (0.9459 / 0.8963) Training 81: 28000 / 123988: Premsel loss 0.1905, acc 0.9212 (0.9429 / 0.8994) Training 81: 35000 / 123988: Premsel loss 0.1869, acc 0.9235 (0.9472 / 0.8999) Training 81: 42000 / 123988: Premsel loss 0.1898, acc 0.9225 (0.9457 / 0.8993) Training 81: 49000 / 123988: Premsel loss 0.1885, acc 0.9229 (0.9460 / 0.8999) Training 81: 56000 / 123988: Premsel loss 0.1886, acc 0.9225 (0.9476 / 0.8974) Training 81: 63000 / 123988: Premsel loss 0.1860, acc 0.9240 (0.9494 / 0.8985) Training 81: 70000 / 123988: Premsel loss 0.1847, acc 0.9241 (0.9498 / 0.8985) Training 81: 77000 / 123988: Premsel loss 0.1861, acc 0.9237 (0.9493 / 0.8981) Training 81: 84000 / 123988: Premsel loss 0.1917, acc 0.9214 (0.9447 / 0.8981) Training 81: 91000 / 123988: Premsel loss 0.1828, acc 0.9250 (0.9487 / 0.9012) Training 81: 98000 / 123988: Premsel loss 0.1886, acc 0.9236 (0.9455 / 0.9017) Training 81: 105000 / 123988: Premsel loss 0.1923, acc 0.9211 (0.9442 / 0.8979) Training 81: 112000 / 123988: Premsel loss 0.1909, acc 0.9220 (0.9453 / 0.8986) Training 81: 119000 / 123988: Premsel loss 0.1964, acc 0.9190 (0.9430 / 0.8949) Evaluation 81: Premsel loss 0.1976, acc 0.9176 (0.9218 / 0.9134) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_06.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-solo___out1.pkl.gz Training 82: 0 / 110159: Premsel loss 0.1957, acc 0.9194 (0.9469 / 0.8919) Training 82: 7000 / 110159: Premsel loss 0.2011, acc 0.9162 (0.9357 / 0.8968) Training 82: 14000 / 110159: Premsel loss 0.1952, acc 0.9190 (0.9403 / 0.8976) Training 82: 21000 / 110159: Premsel loss 0.1984, acc 0.9172 (0.9430 / 0.8914) Training 82: 28000 / 110159: Premsel loss 0.1980, acc 0.9166 (0.9405 / 0.8927) Training 82: 35000 / 110159: Premsel loss 0.1998, acc 0.9163 (0.9409 / 0.8916) Training 82: 42000 / 110159: Premsel loss 0.2048, acc 0.9139 (0.9351 / 0.8927) Training 82: 49000 / 110159: Premsel loss 0.2071, acc 0.9134 (0.9379 / 0.8889) Training 82: 56000 / 110159: Premsel loss 0.1979, acc 0.9176 (0.9437 / 0.8915) Training 82: 63000 / 110159: Premsel loss 0.1990, acc 0.9177 (0.9418 / 0.8936) Training 82: 70000 / 110159: Premsel loss 0.2038, acc 0.9149 (0.9466 / 0.8831) Training 82: 77000 / 110159: Premsel loss 0.2020, acc 0.9159 (0.9425 / 0.8893) Training 82: 84000 / 110159: Premsel loss 0.2011, acc 0.9153 (0.9373 / 0.8933) Training 82: 91000 / 110159: Premsel loss 0.2007, acc 0.9154 (0.9376 / 0.8931) Training 82: 98000 / 110159: Premsel loss 0.2018, acc 0.9147 (0.9411 / 0.8883) Training 82: 105000 / 110159: Premsel loss 0.2007, acc 0.9167 (0.9399 / 0.8936) Evaluation 82: Premsel loss 0.1969, acc 0.9182 (0.9529 / 0.8836) Training 83: 0 / 110159: Premsel loss 0.2013, acc 0.9158 (0.9413 / 0.8903) Training 83: 7000 / 110159: Premsel loss 0.1996, acc 0.9171 (0.9393 / 0.8950) Training 83: 14000 / 110159: Premsel loss 0.1986, acc 0.9173 (0.9421 / 0.8925) Training 83: 21000 / 110159: Premsel loss 0.2058, acc 0.9137 (0.9354 / 0.8919) Training 83: 28000 / 110159: Premsel loss 0.2092, acc 0.9118 (0.9327 / 0.8908) Training 83: 35000 / 110159: Premsel loss 0.2069, acc 0.9140 (0.9415 / 0.8865) Training 83: 42000 / 110159: Premsel loss 0.1957, acc 0.9188 (0.9431 / 0.8946) Training 83: 49000 / 110159: Premsel loss 0.2026, acc 0.9156 (0.9352 / 0.8961) Training 83: 56000 / 110159: Premsel loss 0.1942, acc 0.9193 (0.9429 / 0.8957) Training 83: 63000 / 110159: Premsel loss 0.1993, acc 0.9167 (0.9390 / 0.8943) Training 83: 70000 / 110159: Premsel loss 0.2017, acc 0.9166 (0.9412 / 0.8920) Training 83: 77000 / 110159: Premsel loss 0.1994, acc 0.9170 (0.9445 / 0.8894) Training 83: 84000 / 110159: Premsel loss 0.1907, acc 0.9214 (0.9431 / 0.8998) Training 83: 91000 / 110159: Premsel loss 0.2034, acc 0.9138 (0.9345 / 0.8931) Training 83: 98000 / 110159: Premsel loss 0.2004, acc 0.9166 (0.9410 / 0.8922) Training 83: 105000 / 110159: Premsel loss 0.2002, acc 0.9160 (0.9375 / 0.8944) Evaluation 83: Premsel loss 0.1960, acc 0.9190 (0.9422 / 0.8957) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_24.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-solo___out1.pkl.gz Training 84: 0 / 115299: Premsel loss 0.1976, acc 0.9178 (0.9417 / 0.8939) Training 84: 7000 / 115299: Premsel loss 0.1949, acc 0.9189 (0.9482 / 0.8896) Training 84: 14000 / 115299: Premsel loss 0.1971, acc 0.9179 (0.9458 / 0.8899) Training 84: 21000 / 115299: Premsel loss 0.1924, acc 0.9198 (0.9451 / 0.8945) Training 84: 28000 / 115299: Premsel loss 0.1950, acc 0.9185 (0.9435 / 0.8934) Training 84: 35000 / 115299: Premsel loss 0.1984, acc 0.9174 (0.9417 / 0.8931) Training 84: 42000 / 115299: Premsel loss 0.1981, acc 0.9170 (0.9440 / 0.8900) Training 84: 49000 / 115299: Premsel loss 0.2013, acc 0.9155 (0.9425 / 0.8886) Training 84: 56000 / 115299: Premsel loss 0.1967, acc 0.9183 (0.9436 / 0.8931) Training 84: 63000 / 115299: Premsel loss 0.1949, acc 0.9191 (0.9434 / 0.8948) Training 84: 70000 / 115299: Premsel loss 0.1918, acc 0.9198 (0.9464 / 0.8932) Training 84: 77000 / 115299: Premsel loss 0.1981, acc 0.9177 (0.9432 / 0.8923) Training 84: 84000 / 115299: Premsel loss 0.1961, acc 0.9190 (0.9491 / 0.8890) Training 84: 91000 / 115299: Premsel loss 0.1911, acc 0.9214 (0.9449 / 0.8979) Training 84: 98000 / 115299: Premsel loss 0.1991, acc 0.9168 (0.9405 / 0.8931) Training 84: 105000 / 115299: Premsel loss 0.1923, acc 0.9203 (0.9457 / 0.8950) Training 84: 112000 / 115299: Premsel loss 0.1902, acc 0.9218 (0.9470 / 0.8966) Evaluation 84: Premsel loss 0.1925, acc 0.9202 (0.9472 / 0.8932) Training 85: 0 / 115299: Premsel loss 0.1942, acc 0.9192 (0.9434 / 0.8949) Training 85: 7000 / 115299: Premsel loss 0.1870, acc 0.9225 (0.9482 / 0.8969) Training 85: 14000 / 115299: Premsel loss 0.1928, acc 0.9187 (0.9439 / 0.8935) Training 85: 21000 / 115299: Premsel loss 0.1974, acc 0.9173 (0.9426 / 0.8920) Training 85: 28000 / 115299: Premsel loss 0.1930, acc 0.9199 (0.9451 / 0.8947) Training 85: 35000 / 115299: Premsel loss 0.1956, acc 0.9185 (0.9436 / 0.8934) Training 85: 42000 / 115299: Premsel loss 0.1884, acc 0.9221 (0.9483 / 0.8958) Training 85: 49000 / 115299: Premsel loss 0.1963, acc 0.9175 (0.9469 / 0.8881) Training 85: 56000 / 115299: Premsel loss 0.1961, acc 0.9183 (0.9411 / 0.8956) Training 85: 63000 / 115299: Premsel loss 0.2020, acc 0.9154 (0.9422 / 0.8887) Training 85: 70000 / 115299: Premsel loss 0.1984, acc 0.9175 (0.9443 / 0.8907) Training 85: 77000 / 115299: Premsel loss 0.1986, acc 0.9161 (0.9402 / 0.8921) Training 85: 84000 / 115299: Premsel loss 0.1918, acc 0.9198 (0.9428 / 0.8967) Training 85: 91000 / 115299: Premsel loss 0.2033, acc 0.9153 (0.9414 / 0.8892) Training 85: 98000 / 115299: Premsel loss 0.1980, acc 0.9171 (0.9441 / 0.8902) Training 85: 105000 / 115299: Premsel loss 0.1945, acc 0.9186 (0.9432 / 0.8940) Training 85: 112000 / 115299: Premsel loss 0.1940, acc 0.9199 (0.9442 / 0.8955) Evaluation 85: Premsel loss 0.1982, acc 0.9173 (0.9414 / 0.8933) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_25.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 86: 0 / 133420: Premsel loss 0.2057, acc 0.9172 (0.9411 / 0.8933) Training 86: 7000 / 133420: Premsel loss 0.2068, acc 0.9129 (0.9418 / 0.8841) Training 86: 14000 / 133420: Premsel loss 0.2002, acc 0.9162 (0.9395 / 0.8929) Training 86: 21000 / 133420: Premsel loss 0.1957, acc 0.9182 (0.9435 / 0.8930) Training 86: 28000 / 133420: Premsel loss 0.2026, acc 0.9143 (0.9408 / 0.8878) Training 86: 35000 / 133420: Premsel loss 0.2012, acc 0.9160 (0.9401 / 0.8918) Training 86: 42000 / 133420: Premsel loss 0.2018, acc 0.9158 (0.9404 / 0.8912) Training 86: 49000 / 133420: Premsel loss 0.1991, acc 0.9170 (0.9417 / 0.8923) Training 86: 56000 / 133420: Premsel loss 0.2006, acc 0.9162 (0.9410 / 0.8914) Training 86: 63000 / 133420: Premsel loss 0.1950, acc 0.9199 (0.9434 / 0.8963) Training 86: 70000 / 133420: Premsel loss 0.1999, acc 0.9156 (0.9418 / 0.8894) Training 86: 77000 / 133420: Premsel loss 0.1998, acc 0.9165 (0.9412 / 0.8917) Training 86: 84000 / 133420: Premsel loss 0.1986, acc 0.9163 (0.9428 / 0.8899) Training 86: 91000 / 133420: Premsel loss 0.1959, acc 0.9187 (0.9444 / 0.8930) Training 86: 98000 / 133420: Premsel loss 0.1950, acc 0.9185 (0.9407 / 0.8962) Training 86: 105000 / 133420: Premsel loss 0.1975, acc 0.9174 (0.9423 / 0.8925) Training 86: 112000 / 133420: Premsel loss 0.1962, acc 0.9189 (0.9432 / 0.8947) Training 86: 119000 / 133420: Premsel loss 0.2030, acc 0.9141 (0.9425 / 0.8858) Training 86: 126000 / 133420: Premsel loss 0.1962, acc 0.9179 (0.9460 / 0.8898) Training 86: 133000 / 133420: Premsel loss 0.1994, acc 0.9159 (0.9448 / 0.8870) Evaluation 86: Premsel loss 0.1974, acc 0.9176 (0.9433 / 0.8919) Training 87: 0 / 133420: Premsel loss 0.2000, acc 0.9159 (0.9407 / 0.8912) Training 87: 7000 / 133420: Premsel loss 0.1981, acc 0.9165 (0.9407 / 0.8923) Training 87: 14000 / 133420: Premsel loss 0.1967, acc 0.9174 (0.9443 / 0.8905) Training 87: 21000 / 133420: Premsel loss 0.1933, acc 0.9193 (0.9446 / 0.8939) Training 87: 28000 / 133420: Premsel loss 0.1931, acc 0.9194 (0.9442 / 0.8946) Training 87: 35000 / 133420: Premsel loss 0.1929, acc 0.9203 (0.9434 / 0.8972) Training 87: 42000 / 133420: Premsel loss 0.1999, acc 0.9164 (0.9437 / 0.8890) Training 87: 49000 / 133420: Premsel loss 0.1949, acc 0.9179 (0.9418 / 0.8941) Training 87: 56000 / 133420: Premsel loss 0.2016, acc 0.9158 (0.9439 / 0.8878) Training 87: 63000 / 133420: Premsel loss 0.2065, acc 0.9124 (0.9391 / 0.8857) Training 87: 70000 / 133420: Premsel loss 0.2009, acc 0.9158 (0.9404 / 0.8912) Training 87: 77000 / 133420: Premsel loss 0.1979, acc 0.9166 (0.9425 / 0.8907) Training 87: 84000 / 133420: Premsel loss 0.2028, acc 0.9144 (0.9374 / 0.8913) Training 87: 91000 / 133420: Premsel loss 0.2023, acc 0.9147 (0.9421 / 0.8874) Training 87: 98000 / 133420: Premsel loss 0.1998, acc 0.9159 (0.9418 / 0.8900) Training 87: 105000 / 133420: Premsel loss 0.2012, acc 0.9158 (0.9449 / 0.8867) Training 87: 112000 / 133420: Premsel loss 0.2019, acc 0.9154 (0.9435 / 0.8873) Training 87: 119000 / 133420: Premsel loss 0.1976, acc 0.9167 (0.9419 / 0.8915) Training 87: 126000 / 133420: Premsel loss 0.2060, acc 0.9136 (0.9426 / 0.8845) Training 87: 133000 / 133420: Premsel loss 0.2015, acc 0.9155 (0.9387 / 0.8923) Evaluation 87: Premsel loss 0.1979, acc 0.9171 (0.9370 / 0.8972) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_18.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_07.pkl.gz Training 88: 0 / 127665: Premsel loss 0.2004, acc 0.9161 (0.9408 / 0.8914) Training 88: 7000 / 127665: Premsel loss 0.1999, acc 0.9170 (0.9455 / 0.8885) Training 88: 14000 / 127665: Premsel loss 0.1967, acc 0.9185 (0.9444 / 0.8925) Training 88: 21000 / 127665: Premsel loss 0.1941, acc 0.9204 (0.9467 / 0.8941) Training 88: 28000 / 127665: Premsel loss 0.1956, acc 0.9187 (0.9404 / 0.8970) Training 88: 35000 / 127665: Premsel loss 0.1977, acc 0.9178 (0.9425 / 0.8931) Training 88: 42000 / 127665: Premsel loss 0.1895, acc 0.9220 (0.9480 / 0.8961) Training 88: 49000 / 127665: Premsel loss 0.1995, acc 0.9181 (0.9447 / 0.8914) Training 88: 56000 / 127665: Premsel loss 0.1950, acc 0.9192 (0.9432 / 0.8952) Training 88: 63000 / 127665: Premsel loss 0.1964, acc 0.9183 (0.9462 / 0.8903) Training 88: 70000 / 127665: Premsel loss 0.1991, acc 0.9178 (0.9414 / 0.8941) Training 88: 77000 / 127665: Premsel loss 0.2039, acc 0.9150 (0.9411 / 0.8890) Training 88: 84000 / 127665: Premsel loss 0.1966, acc 0.9187 (0.9425 / 0.8949) Training 88: 91000 / 127665: Premsel loss 0.2011, acc 0.9162 (0.9426 / 0.8898) Training 88: 98000 / 127665: Premsel loss 0.1994, acc 0.9170 (0.9393 / 0.8947) Training 88: 105000 / 127665: Premsel loss 0.1981, acc 0.9186 (0.9447 / 0.8924) Training 88: 112000 / 127665: Premsel loss 0.1912, acc 0.9220 (0.9449 / 0.8991) Training 88: 119000 / 127665: Premsel loss 0.2015, acc 0.9155 (0.9383 / 0.8928) Training 88: 126000 / 127665: Premsel loss 0.1994, acc 0.9163 (0.9406 / 0.8921) Evaluation 88: Premsel loss 0.1974, acc 0.9183 (0.9435 / 0.8930) Training 89: 0 / 127665: Premsel loss 0.1974, acc 0.9177 (0.9399 / 0.8955) Training 89: 7000 / 127665: Premsel loss 0.1991, acc 0.9177 (0.9428 / 0.8927) Training 89: 14000 / 127665: Premsel loss 0.1952, acc 0.9191 (0.9463 / 0.8920) Training 89: 21000 / 127665: Premsel loss 0.1931, acc 0.9202 (0.9470 / 0.8935) Training 89: 28000 / 127665: Premsel loss 0.2039, acc 0.9142 (0.9376 / 0.8908) Training 89: 35000 / 127665: Premsel loss 0.1964, acc 0.9194 (0.9459 / 0.8930) Training 89: 42000 / 127665: Premsel loss 0.1989, acc 0.9170 (0.9442 / 0.8898) Training 89: 49000 / 127665: Premsel loss 0.2070, acc 0.9139 (0.9435 / 0.8843) Training 89: 56000 / 127665: Premsel loss 0.2168, acc 0.9083 (0.9363 / 0.8803) Training 89: 63000 / 127665: Premsel loss 0.2083, acc 0.9128 (0.9388 / 0.8868) Training 89: 70000 / 127665: Premsel loss 0.2059, acc 0.9141 (0.9385 / 0.8896) Training 89: 77000 / 127665: Premsel loss 0.1994, acc 0.9170 (0.9414 / 0.8926) Training 89: 84000 / 127665: Premsel loss 0.2003, acc 0.9172 (0.9458 / 0.8886) Training 89: 91000 / 127665: Premsel loss 0.1976, acc 0.9178 (0.9435 / 0.8921) Training 89: 98000 / 127665: Premsel loss 0.1930, acc 0.9202 (0.9485 / 0.8919) Training 89: 105000 / 127665: Premsel loss 0.1979, acc 0.9180 (0.9438 / 0.8922) Training 89: 112000 / 127665: Premsel loss 0.1995, acc 0.9180 (0.9457 / 0.8903) Training 89: 119000 / 127665: Premsel loss 0.1979, acc 0.9172 (0.9434 / 0.8911) Training 89: 126000 / 127665: Premsel loss 0.1954, acc 0.9196 (0.9430 / 0.8963) Evaluation 89: Premsel loss 0.1970, acc 0.9188 (0.9412 / 0.8963) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_11.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz Training 90: 0 / 123992: Premsel loss 0.1971, acc 0.9193 (0.9449 / 0.8936) Training 90: 7000 / 123992: Premsel loss 0.2156, acc 0.9088 (0.9279 / 0.8896) Training 90: 14000 / 123992: Premsel loss 0.2057, acc 0.9131 (0.9367 / 0.8895) Training 90: 21000 / 123992: Premsel loss 0.2284, acc 0.9028 (0.9202 / 0.8853) Training 90: 28000 / 123992: Premsel loss 0.2158, acc 0.9083 (0.9316 / 0.8849) Training 90: 35000 / 123992: Premsel loss 0.2204, acc 0.9060 (0.9325 / 0.8795) Training 90: 42000 / 123992: Premsel loss 0.2194, acc 0.9061 (0.9306 / 0.8815) Training 90: 49000 / 123992: Premsel loss 0.2148, acc 0.9082 (0.9337 / 0.8827) Training 90: 56000 / 123992: Premsel loss 0.2240, acc 0.9052 (0.9292 / 0.8812) Training 90: 63000 / 123992: Premsel loss 0.2189, acc 0.9062 (0.9312 / 0.8812) Training 90: 70000 / 123992: Premsel loss 0.2162, acc 0.9078 (0.9277 / 0.8879) Training 90: 77000 / 123992: Premsel loss 0.2168, acc 0.9082 (0.9312 / 0.8852) Training 90: 84000 / 123992: Premsel loss 0.2125, acc 0.9103 (0.9339 / 0.8867) Training 90: 91000 / 123992: Premsel loss 0.2112, acc 0.9107 (0.9319 / 0.8895) Training 90: 98000 / 123992: Premsel loss 0.2185, acc 0.9073 (0.9291 / 0.8855) Training 90: 105000 / 123992: Premsel loss 0.2190, acc 0.9072 (0.9331 / 0.8813) Training 90: 112000 / 123992: Premsel loss 0.2158, acc 0.9093 (0.9305 / 0.8880) Training 90: 119000 / 123992: Premsel loss 0.2164, acc 0.9073 (0.9297 / 0.8849) Evaluation 90: Premsel loss 0.2103, acc 0.9109 (0.9437 / 0.8781) Training 91: 0 / 123992: Premsel loss 0.2067, acc 0.9124 (0.9348 / 0.8900) Training 91: 7000 / 123992: Premsel loss 0.2113, acc 0.9110 (0.9338 / 0.8881) Training 91: 14000 / 123992: Premsel loss 0.2138, acc 0.9095 (0.9295 / 0.8895) Training 91: 21000 / 123992: Premsel loss 0.2206, acc 0.9054 (0.9282 / 0.8826) Training 91: 28000 / 123992: Premsel loss 0.2156, acc 0.9078 (0.9340 / 0.8816) Training 91: 35000 / 123992: Premsel loss 0.2165, acc 0.9079 (0.9340 / 0.8818) Training 91: 42000 / 123992: Premsel loss 0.2114, acc 0.9101 (0.9292 / 0.8910) Training 91: 49000 / 123992: Premsel loss 0.2133, acc 0.9090 (0.9328 / 0.8852) Training 91: 56000 / 123992: Premsel loss 0.2207, acc 0.9058 (0.9255 / 0.8860) Training 91: 63000 / 123992: Premsel loss 0.2197, acc 0.9058 (0.9293 / 0.8824) Training 91: 70000 / 123992: Premsel loss 0.2144, acc 0.9088 (0.9329 / 0.8847) Training 91: 77000 / 123992: Premsel loss 0.2197, acc 0.9062 (0.9291 / 0.8833) Training 91: 84000 / 123992: Premsel loss 0.2213, acc 0.9055 (0.9248 / 0.8861) Training 91: 91000 / 123992: Premsel loss 0.2175, acc 0.9073 (0.9280 / 0.8866) Training 91: 98000 / 123992: Premsel loss 0.2151, acc 0.9078 (0.9304 / 0.8853) Training 91: 105000 / 123992: Premsel loss 0.2196, acc 0.9060 (0.9301 / 0.8819) Training 91: 112000 / 123992: Premsel loss 0.2158, acc 0.9078 (0.9326 / 0.8829) Training 91: 119000 / 123992: Premsel loss 0.2131, acc 0.9092 (0.9322 / 0.8863) Evaluation 91: Premsel loss 0.2126, acc 0.9098 (0.9243 / 0.8953) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_09.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t300-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 92: 0 / 131053: Premsel loss 0.2157, acc 0.9074 (0.9325 / 0.8822) Training 92: 7000 / 131053: Premsel loss 0.2023, acc 0.9152 (0.9396 / 0.8909) Training 92: 14000 / 131053: Premsel loss 0.2083, acc 0.9119 (0.9363 / 0.8875) Training 92: 21000 / 131053: Premsel loss 0.2073, acc 0.9125 (0.9365 / 0.8885) Training 92: 28000 / 131053: Premsel loss 0.2080, acc 0.9123 (0.9372 / 0.8874) Training 92: 35000 / 131053: Premsel loss 0.2085, acc 0.9109 (0.9356 / 0.8861) Training 92: 42000 / 131053: Premsel loss 0.2060, acc 0.9133 (0.9377 / 0.8889) Training 92: 49000 / 131053: Premsel loss 0.2071, acc 0.9119 (0.9385 / 0.8853) Training 92: 56000 / 131053: Premsel loss 0.2087, acc 0.9115 (0.9369 / 0.8860) Training 92: 63000 / 131053: Premsel loss 0.2062, acc 0.9129 (0.9409 / 0.8849) Training 92: 70000 / 131053: Premsel loss 0.2076, acc 0.9121 (0.9372 / 0.8871) Training 92: 77000 / 131053: Premsel loss 0.2064, acc 0.9132 (0.9373 / 0.8891) Training 92: 84000 / 131053: Premsel loss 0.2078, acc 0.9121 (0.9354 / 0.8887) Training 92: 91000 / 131053: Premsel loss 0.2091, acc 0.9116 (0.9385 / 0.8848) Training 92: 98000 / 131053: Premsel loss 0.2094, acc 0.9113 (0.9372 / 0.8854) Training 92: 105000 / 131053: Premsel loss 0.2049, acc 0.9134 (0.9342 / 0.8925) Training 92: 112000 / 131053: Premsel loss 0.2031, acc 0.9149 (0.9389 / 0.8909) Training 92: 119000 / 131053: Premsel loss 0.2091, acc 0.9118 (0.9384 / 0.8852) Training 92: 126000 / 131053: Premsel loss 0.2114, acc 0.9098 (0.9355 / 0.8841) Evaluation 92: Premsel loss 0.2091, acc 0.9120 (0.9472 / 0.8768) Training 93: 0 / 131053: Premsel loss 0.2108, acc 0.9102 (0.9292 / 0.8912) Training 93: 7000 / 131053: Premsel loss 0.2085, acc 0.9118 (0.9339 / 0.8898) Training 93: 14000 / 131053: Premsel loss 0.2045, acc 0.9137 (0.9357 / 0.8917) Training 93: 21000 / 131053: Premsel loss 0.2070, acc 0.9124 (0.9396 / 0.8853) Training 93: 28000 / 131053: Premsel loss 0.2080, acc 0.9123 (0.9403 / 0.8844) Training 93: 35000 / 131053: Premsel loss 0.2082, acc 0.9122 (0.9370 / 0.8873) Training 93: 42000 / 131053: Premsel loss 0.2075, acc 0.9118 (0.9377 / 0.8859) Training 93: 49000 / 131053: Premsel loss 0.2014, acc 0.9149 (0.9384 / 0.8915) Training 93: 56000 / 131053: Premsel loss 0.2068, acc 0.9131 (0.9394 / 0.8869) Training 93: 63000 / 131053: Premsel loss 0.2131, acc 0.9084 (0.9360 / 0.8808) Training 93: 70000 / 131053: Premsel loss 0.2050, acc 0.9146 (0.9378 / 0.8913) Training 93: 77000 / 131053: Premsel loss 0.2063, acc 0.9131 (0.9391 / 0.8871) Training 93: 84000 / 131053: Premsel loss 0.2082, acc 0.9118 (0.9351 / 0.8886) Training 93: 91000 / 131053: Premsel loss 0.2070, acc 0.9121 (0.9378 / 0.8865) Training 93: 98000 / 131053: Premsel loss 0.2129, acc 0.9090 (0.9357 / 0.8823) Training 93: 105000 / 131053: Premsel loss 0.2112, acc 0.9102 (0.9375 / 0.8829) Training 93: 112000 / 131053: Premsel loss 0.2147, acc 0.9091 (0.9330 / 0.8851) Training 93: 119000 / 131053: Premsel loss 0.2089, acc 0.9116 (0.9341 / 0.8891) Training 93: 126000 / 131053: Premsel loss 0.2065, acc 0.9128 (0.9345 / 0.8911) Evaluation 93: Premsel loss 0.2079, acc 0.9124 (0.9244 / 0.9004) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_20.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz Training 94: 0 / 117201: Premsel loss 0.2096, acc 0.9105 (0.9317 / 0.8894) Training 94: 7000 / 117201: Premsel loss 0.2036, acc 0.9145 (0.9375 / 0.8916) Training 94: 14000 / 117201: Premsel loss 0.1983, acc 0.9170 (0.9444 / 0.8896) Training 94: 21000 / 117201: Premsel loss 0.1995, acc 0.9163 (0.9393 / 0.8932) Training 94: 28000 / 117201: Premsel loss 0.2008, acc 0.9162 (0.9427 / 0.8896) Training 94: 35000 / 117201: Premsel loss 0.2058, acc 0.9136 (0.9384 / 0.8887) Training 94: 42000 / 117201: Premsel loss 0.2064, acc 0.9125 (0.9410 / 0.8840) Training 94: 49000 / 117201: Premsel loss 0.2046, acc 0.9133 (0.9387 / 0.8879) Training 94: 56000 / 117201: Premsel loss 0.2029, acc 0.9140 (0.9401 / 0.8879) Training 94: 63000 / 117201: Premsel loss 0.2054, acc 0.9139 (0.9362 / 0.8916) Training 94: 70000 / 117201: Premsel loss 0.1972, acc 0.9173 (0.9426 / 0.8921) Training 94: 77000 / 117201: Premsel loss 0.1993, acc 0.9163 (0.9400 / 0.8925) Training 94: 84000 / 117201: Premsel loss 0.1963, acc 0.9178 (0.9424 / 0.8932) Training 94: 91000 / 117201: Premsel loss 0.1987, acc 0.9166 (0.9463 / 0.8869) Training 94: 98000 / 117201: Premsel loss 0.1974, acc 0.9174 (0.9441 / 0.8907) Training 94: 105000 / 117201: Premsel loss 0.1955, acc 0.9186 (0.9451 / 0.8921) Training 94: 112000 / 117201: Premsel loss 0.1997, acc 0.9160 (0.9415 / 0.8905) Evaluation 94: Premsel loss 0.1972, acc 0.9175 (0.9498 / 0.8853) Training 95: 0 / 117201: Premsel loss 0.1938, acc 0.9188 (0.9415 / 0.8961) Training 95: 7000 / 117201: Premsel loss 0.2030, acc 0.9147 (0.9399 / 0.8896) Training 95: 14000 / 117201: Premsel loss 0.1973, acc 0.9175 (0.9437 / 0.8912) Training 95: 21000 / 117201: Premsel loss 0.2022, acc 0.9146 (0.9396 / 0.8896) Training 95: 28000 / 117201: Premsel loss 0.2040, acc 0.9142 (0.9426 / 0.8858) Training 95: 35000 / 117201: Premsel loss 0.2011, acc 0.9150 (0.9403 / 0.8897) Training 95: 42000 / 117201: Premsel loss 0.1965, acc 0.9180 (0.9432 / 0.8928) Training 95: 49000 / 117201: Premsel loss 0.2045, acc 0.9143 (0.9394 / 0.8891) Training 95: 56000 / 117201: Premsel loss 0.1994, acc 0.9171 (0.9461 / 0.8880) Training 95: 63000 / 117201: Premsel loss 0.2020, acc 0.9158 (0.9388 / 0.8927) Training 95: 70000 / 117201: Premsel loss 0.2023, acc 0.9152 (0.9405 / 0.8898) Training 95: 77000 / 117201: Premsel loss 0.1950, acc 0.9179 (0.9429 / 0.8929) Training 95: 84000 / 117201: Premsel loss 0.1963, acc 0.9186 (0.9415 / 0.8956) Training 95: 91000 / 117201: Premsel loss 0.1953, acc 0.9180 (0.9435 / 0.8925) Training 95: 98000 / 117201: Premsel loss 0.1970, acc 0.9173 (0.9455 / 0.8892) Training 95: 105000 / 117201: Premsel loss 0.2046, acc 0.9144 (0.9396 / 0.8891) Training 95: 112000 / 117201: Premsel loss 0.2025, acc 0.9157 (0.9433 / 0.8880) Evaluation 95: Premsel loss 0.1992, acc 0.9164 (0.9313 / 0.9014) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02___out1.pkl.gz Training 96: 0 / 105417: Premsel loss 0.1950, acc 0.9185 (0.9404 / 0.8965) Training 96: 7000 / 105417: Premsel loss 0.1806, acc 0.9246 (0.9461 / 0.9031) Training 96: 14000 / 105417: Premsel loss 0.1881, acc 0.9211 (0.9457 / 0.8965) Training 96: 21000 / 105417: Premsel loss 0.1748, acc 0.9279 (0.9504 / 0.9054) Training 96: 28000 / 105417: Premsel loss 0.1842, acc 0.9234 (0.9457 / 0.9011) Training 96: 35000 / 105417: Premsel loss 0.1938, acc 0.9197 (0.9437 / 0.8957) Training 96: 42000 / 105417: Premsel loss 0.1884, acc 0.9226 (0.9423 / 0.9028) Training 96: 49000 / 105417: Premsel loss 0.1809, acc 0.9249 (0.9441 / 0.9057) Training 96: 56000 / 105417: Premsel loss 0.1758, acc 0.9276 (0.9486 / 0.9066) Training 96: 63000 / 105417: Premsel loss 0.1802, acc 0.9259 (0.9505 / 0.9014) Training 96: 70000 / 105417: Premsel loss 0.1894, acc 0.9206 (0.9479 / 0.8933) Training 96: 77000 / 105417: Premsel loss 0.1879, acc 0.9219 (0.9491 / 0.8947) Training 96: 84000 / 105417: Premsel loss 0.1885, acc 0.9210 (0.9432 / 0.8988) Training 96: 91000 / 105417: Premsel loss 0.1948, acc 0.9180 (0.9432 / 0.8927) Training 96: 98000 / 105417: Premsel loss 0.1947, acc 0.9188 (0.9443 / 0.8933) Training 96: 105000 / 105417: Premsel loss 0.1890, acc 0.9210 (0.9431 / 0.8988) Evaluation 96: Premsel loss 0.1833, acc 0.9234 (0.9446 / 0.9022) Training 97: 0 / 105417: Premsel loss 0.1832, acc 0.9230 (0.9460 / 0.9001) Training 97: 7000 / 105417: Premsel loss 0.1942, acc 0.9190 (0.9416 / 0.8965) Training 97: 14000 / 105417: Premsel loss 0.1897, acc 0.9201 (0.9475 / 0.8927) Training 97: 21000 / 105417: Premsel loss 0.1847, acc 0.9235 (0.9478 / 0.8992) Training 97: 28000 / 105417: Premsel loss 0.1988, acc 0.9164 (0.9390 / 0.8938) Training 97: 35000 / 105417: Premsel loss 0.1954, acc 0.9177 (0.9452 / 0.8902) Training 97: 42000 / 105417: Premsel loss 0.1942, acc 0.9186 (0.9465 / 0.8906) Training 97: 49000 / 105417: Premsel loss 0.1998, acc 0.9166 (0.9412 / 0.8920) Training 97: 56000 / 105417: Premsel loss 0.1955, acc 0.9178 (0.9410 / 0.8946) Training 97: 63000 / 105417: Premsel loss 0.2001, acc 0.9158 (0.9392 / 0.8923) Training 97: 70000 / 105417: Premsel loss 0.1911, acc 0.9197 (0.9405 / 0.8988) Training 97: 77000 / 105417: Premsel loss 0.1882, acc 0.9208 (0.9428 / 0.8987) Training 97: 84000 / 105417: Premsel loss 0.1906, acc 0.9207 (0.9431 / 0.8982) Training 97: 91000 / 105417: Premsel loss 0.1998, acc 0.9148 (0.9410 / 0.8885) Training 97: 98000 / 105417: Premsel loss 0.1913, acc 0.9203 (0.9385 / 0.9020) Training 97: 105000 / 105417: Premsel loss 0.1794, acc 0.9263 (0.9465 / 0.9061) Evaluation 97: Premsel loss 0.1850, acc 0.9226 (0.9374 / 0.9078) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l1800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz Training 98: 0 / 104878: Premsel loss 0.1889, acc 0.9207 (0.9454 / 0.8960) Training 98: 7000 / 104878: Premsel loss 0.2350, acc 0.8990 (0.9228 / 0.8753) Training 98: 14000 / 104878: Premsel loss 0.2271, acc 0.9028 (0.9277 / 0.8778) Training 98: 21000 / 104878: Premsel loss 0.2377, acc 0.8972 (0.9210 / 0.8734) Training 98: 28000 / 104878: Premsel loss 0.2862, acc 0.8722 (0.8870 / 0.8574) Training 98: 35000 / 104878: Premsel loss 0.2567, acc 0.8872 (0.9010 / 0.8734) Training 98: 42000 / 104878: Premsel loss 0.2480, acc 0.8918 (0.9171 / 0.8665) Training 98: 49000 / 104878: Premsel loss 0.2535, acc 0.8900 (0.9112 / 0.8689) Training 98: 56000 / 104878: Premsel loss 0.2436, acc 0.8947 (0.9123 / 0.8771) Training 98: 63000 / 104878: Premsel loss 0.2395, acc 0.8950 (0.9121 / 0.8780) Training 98: 70000 / 104878: Premsel loss 0.2351, acc 0.8977 (0.9258 / 0.8697) Training 98: 77000 / 104878: Premsel loss 0.2335, acc 0.8984 (0.9213 / 0.8755) Training 98: 84000 / 104878: Premsel loss 0.2425, acc 0.8953 (0.9190 / 0.8715) Training 98: 91000 / 104878: Premsel loss 0.2352, acc 0.8987 (0.9259 / 0.8715) Training 98: 98000 / 104878: Premsel loss 0.2339, acc 0.8984 (0.9214 / 0.8754) Evaluation 98: Premsel loss 0.2325, acc 0.8989 (0.9066 / 0.8912) Training 99: 0 / 104878: Premsel loss 0.2374, acc 0.8961 (0.9180 / 0.8742) Training 99: 7000 / 104878: Premsel loss 0.2263, acc 0.9031 (0.9181 / 0.8880) Training 99: 14000 / 104878: Premsel loss 0.2256, acc 0.9034 (0.9210 / 0.8859) Training 99: 21000 / 104878: Premsel loss 0.2256, acc 0.9032 (0.9239 / 0.8826) Training 99: 28000 / 104878: Premsel loss 0.2262, acc 0.9026 (0.9221 / 0.8832) Training 99: 35000 / 104878: Premsel loss 0.2275, acc 0.9007 (0.9289 / 0.8726) Training 99: 42000 / 104878: Premsel loss 0.2313, acc 0.8999 (0.9199 / 0.8799) Training 99: 49000 / 104878: Premsel loss 0.2381, acc 0.8964 (0.9116 / 0.8812) Training 99: 56000 / 104878: Premsel loss 0.2350, acc 0.8976 (0.9108 / 0.8844) Training 99: 63000 / 104878: Premsel loss 0.2308, acc 0.9014 (0.9190 / 0.8838) Training 99: 70000 / 104878: Premsel loss 0.2416, acc 0.8953 (0.9238 / 0.8668) Training 99: 77000 / 104878: Premsel loss 0.2378, acc 0.8958 (0.9148 / 0.8768) Training 99: 84000 / 104878: Premsel loss 0.2387, acc 0.8975 (0.9186 / 0.8763) Training 99: 91000 / 104878: Premsel loss 0.2368, acc 0.8962 (0.9185 / 0.8740) Training 99: 98000 / 104878: Premsel loss 0.2265, acc 0.9024 (0.9265 / 0.8783) Evaluation 99: Premsel loss 0.2314, acc 0.8993 (0.9086 / 0.8901) Loading data... Full data reached, reshuffling... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_16.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz Training 100: 0 / 131631: Premsel loss 0.2271, acc 0.9022 (0.9225 / 0.8819) Training 100: 7000 / 131631: Premsel loss 0.2416, acc 0.8933 (0.9205 / 0.8660) Training 100: 14000 / 131631: Premsel loss 0.2300, acc 0.8992 (0.9233 / 0.8751) Training 100: 21000 / 131631: Premsel loss 0.2400, acc 0.8946 (0.9176 / 0.8716) Training 100: 28000 / 131631: Premsel loss 0.2430, acc 0.8933 (0.9165 / 0.8702) Training 100: 35000 / 131631: Premsel loss 0.2333, acc 0.8976 (0.9217 / 0.8736) Training 100: 42000 / 131631: Premsel loss 0.2310, acc 0.8990 (0.9189 / 0.8791) Training 100: 49000 / 131631: Premsel loss 0.2363, acc 0.8965 (0.9211 / 0.8719) Training 100: 56000 / 131631: Premsel loss 0.2288, acc 0.9002 (0.9267 / 0.8737) Training 100: 63000 / 131631: Premsel loss 0.2293, acc 0.9002 (0.9267 / 0.8738) Training 100: 70000 / 131631: Premsel loss 0.2392, acc 0.8958 (0.9280 / 0.8636) Training 100: 77000 / 131631: Premsel loss 0.2735, acc 0.8773 (0.9092 / 0.8454) Training 100: 84000 / 131631: Premsel loss 0.2510, acc 0.8881 (0.9184 / 0.8579) Training 100: 91000 / 131631: Premsel loss 0.2404, acc 0.8951 (0.9249 / 0.8653) Training 100: 98000 / 131631: Premsel loss 0.2377, acc 0.8966 (0.9176 / 0.8757) Training 100: 105000 / 131631: Premsel loss 0.2321, acc 0.8985 (0.9254 / 0.8716) Training 100: 112000 / 131631: Premsel loss 0.2369, acc 0.8968 (0.9226 / 0.8711) Training 100: 119000 / 131631: Premsel loss 0.2335, acc 0.8972 (0.9259 / 0.8686) Training 100: 126000 / 131631: Premsel loss 0.2250, acc 0.9022 (0.9285 / 0.8759) Evaluation 100: Premsel loss 0.2294, acc 0.9003 (0.9305 / 0.8701) Training 101: 0 / 131631: Premsel loss 0.2340, acc 0.8978 (0.9258 / 0.8699) Training 101: 7000 / 131631: Premsel loss 0.2328, acc 0.8989 (0.9294 / 0.8683) Training 101: 14000 / 131631: Premsel loss 0.2297, acc 0.9005 (0.9255 / 0.8755) Training 101: 21000 / 131631: Premsel loss 0.2208, acc 0.9047 (0.9327 / 0.8767) Training 101: 28000 / 131631: Premsel loss 0.2276, acc 0.9009 (0.9260 / 0.8759) Training 101: 35000 / 131631: Premsel loss 0.2272, acc 0.9017 (0.9256 / 0.8777) Training 101: 42000 / 131631: Premsel loss 0.2255, acc 0.9018 (0.9293 / 0.8743) Training 101: 49000 / 131631: Premsel loss 0.2227, acc 0.9035 (0.9290 / 0.8781) Training 101: 56000 / 131631: Premsel loss 0.2227, acc 0.9039 (0.9302 / 0.8775) Training 101: 63000 / 131631: Premsel loss 0.2257, acc 0.9019 (0.9314 / 0.8724) Training 101: 70000 / 131631: Premsel loss 0.2280, acc 0.9009 (0.9316 / 0.8701) Training 101: 77000 / 131631: Premsel loss 0.2267, acc 0.9018 (0.9301 / 0.8734) Training 101: 84000 / 131631: Premsel loss 0.2262, acc 0.9022 (0.9271 / 0.8772) Training 101: 91000 / 131631: Premsel loss 0.2302, acc 0.9002 (0.9240 / 0.8764) Training 101: 98000 / 131631: Premsel loss 0.2330, acc 0.8976 (0.9251 / 0.8700) Training 101: 105000 / 131631: Premsel loss 0.2209, acc 0.9050 (0.9328 / 0.8772) Training 101: 112000 / 131631: Premsel loss 0.2286, acc 0.9008 (0.9279 / 0.8736) Training 101: 119000 / 131631: Premsel loss 0.2225, acc 0.9040 (0.9276 / 0.8804) Training 101: 126000 / 131631: Premsel loss 0.2289, acc 0.9004 (0.9255 / 0.8752) Evaluation 101: Premsel loss 0.2228, acc 0.9039 (0.9274 / 0.8804) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_21.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz Training 102: 0 / 71545: Premsel loss 0.2265, acc 0.9014 (0.9294 / 0.8734) Training 102: 7000 / 71545: Premsel loss 0.2288, acc 0.9013 (0.9229 / 0.8797) Training 102: 14000 / 71545: Premsel loss 0.2290, acc 0.8999 (0.9218 / 0.8780) Training 102: 21000 / 71545: Premsel loss 0.2307, acc 0.8999 (0.9209 / 0.8789) Training 102: 28000 / 71545: Premsel loss 0.2279, acc 0.9015 (0.9232 / 0.8797) Training 102: 35000 / 71545: Premsel loss 0.2224, acc 0.9038 (0.9273 / 0.8803) Training 102: 42000 / 71545: Premsel loss 0.2258, acc 0.9021 (0.9280 / 0.8762) Training 102: 49000 / 71545: Premsel loss 0.2257, acc 0.9029 (0.9270 / 0.8787) Training 102: 56000 / 71545: Premsel loss 0.2242, acc 0.9033 (0.9252 / 0.8814) Training 102: 63000 / 71545: Premsel loss 0.2283, acc 0.9014 (0.9281 / 0.8746) Training 102: 70000 / 71545: Premsel loss 0.2258, acc 0.9034 (0.9261 / 0.8806) Evaluation 102: Premsel loss 0.2290, acc 0.9013 (0.9074 / 0.8952) Training 103: 0 / 71545: Premsel loss 0.2304, acc 0.9009 (0.9286 / 0.8733) Training 103: 7000 / 71545: Premsel loss 0.2262, acc 0.9028 (0.9238 / 0.8817) Training 103: 14000 / 71545: Premsel loss 0.2300, acc 0.9002 (0.9256 / 0.8748) Training 103: 21000 / 71545: Premsel loss 0.2272, acc 0.9025 (0.9289 / 0.8760) Training 103: 28000 / 71545: Premsel loss 0.2383, acc 0.8961 (0.9192 / 0.8729) Training 103: 35000 / 71545: Premsel loss 0.2234, acc 0.9040 (0.9278 / 0.8803) Training 103: 42000 / 71545: Premsel loss 0.2254, acc 0.9032 (0.9250 / 0.8814) Training 103: 49000 / 71545: Premsel loss 0.2266, acc 0.9029 (0.9320 / 0.8737) Training 103: 56000 / 71545: Premsel loss 0.2251, acc 0.9033 (0.9261 / 0.8804) Training 103: 63000 / 71545: Premsel loss 0.2217, acc 0.9053 (0.9290 / 0.8816) Training 103: 70000 / 71545: Premsel loss 0.2159, acc 0.9084 (0.9355 / 0.8814) Evaluation 103: Premsel loss 0.2262, acc 0.9027 (0.9101 / 0.8953) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l1800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_22.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop01-VHSLCAXPh+lgb-d50-l900-e0.15loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz Training 104: 0 / 115523: Premsel loss 0.2247, acc 0.9022 (0.9301 / 0.8742) Training 104: 7000 / 115523: Premsel loss 0.2260, acc 0.9019 (0.9270 / 0.8769) Training 104: 14000 / 115523: Premsel loss 0.2302, acc 0.9000 (0.9205 / 0.8795) Training 104: 21000 / 115523: Premsel loss 0.2317, acc 0.9001 (0.9267 / 0.8735) Training 104: 28000 / 115523: Premsel loss 0.2325, acc 0.8997 (0.9209 / 0.8784) Training 104: 35000 / 115523: Premsel loss 0.2322, acc 0.8992 (0.9230 / 0.8755) Training 104: 42000 / 115523: Premsel loss 0.2352, acc 0.8981 (0.9245 / 0.8718) Training 104: 49000 / 115523: Premsel loss 0.2360, acc 0.8973 (0.9179 / 0.8767) Training 104: 56000 / 115523: Premsel loss 0.2379, acc 0.8958 (0.9194 / 0.8723) Training 104: 63000 / 115523: Premsel loss 0.2322, acc 0.8988 (0.9246 / 0.8731) Training 104: 70000 / 115523: Premsel loss 0.2256, acc 0.9033 (0.9258 / 0.8809) Training 104: 77000 / 115523: Premsel loss 0.2359, acc 0.8976 (0.9265 / 0.8686) Training 104: 84000 / 115523: Premsel loss 0.2332, acc 0.8995 (0.9191 / 0.8800) Training 104: 91000 / 115523: Premsel loss 0.2326, acc 0.9005 (0.9222 / 0.8789) Training 104: 98000 / 115523: Premsel loss 0.2350, acc 0.8968 (0.9198 / 0.8737) Training 104: 105000 / 115523: Premsel loss 0.2258, acc 0.9029 (0.9260 / 0.8797) Training 104: 112000 / 115523: Premsel loss 0.2314, acc 0.8988 (0.9285 / 0.8691) Evaluation 104: Premsel loss 0.2331, acc 0.8995 (0.9379 / 0.8610) Training 105: 0 / 115523: Premsel loss 0.2275, acc 0.9025 (0.9207 / 0.8844) Training 105: 7000 / 115523: Premsel loss 0.2306, acc 0.9011 (0.9278 / 0.8744) Training 105: 14000 / 115523: Premsel loss 0.2270, acc 0.9021 (0.9265 / 0.8778) Training 105: 21000 / 115523: Premsel loss 0.2275, acc 0.9020 (0.9253 / 0.8787) Training 105: 28000 / 115523: Premsel loss 0.2284, acc 0.9007 (0.9212 / 0.8802) Training 105: 35000 / 115523: Premsel loss 0.2223, acc 0.9051 (0.9272 / 0.8830) Training 105: 42000 / 115523: Premsel loss 0.2257, acc 0.9022 (0.9215 / 0.8828) Training 105: 49000 / 115523: Premsel loss 0.2337, acc 0.8991 (0.9262 / 0.8719) Training 105: 56000 / 115523: Premsel loss 0.2306, acc 0.8993 (0.9273 / 0.8713) Training 105: 63000 / 115523: Premsel loss 0.2272, acc 0.9021 (0.9230 / 0.8811) Training 105: 70000 / 115523: Premsel loss 0.2245, acc 0.9036 (0.9261 / 0.8810) Training 105: 77000 / 115523: Premsel loss 0.2333, acc 0.8981 (0.9243 / 0.8718) Training 105: 84000 / 115523: Premsel loss 0.2285, acc 0.9007 (0.9198 / 0.8816) Training 105: 91000 / 115523: Premsel loss 0.2298, acc 0.9003 (0.9220 / 0.8786) Training 105: 98000 / 115523: Premsel loss 0.2255, acc 0.9027 (0.9237 / 0.8816) Training 105: 105000 / 115523: Premsel loss 0.2193, acc 0.9058 (0.9304 / 0.8813) Training 105: 112000 / 115523: Premsel loss 0.2206, acc 0.9052 (0.9298 / 0.8807) Evaluation 105: Premsel loss 0.2265, acc 0.9024 (0.9137 / 0.8911) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_23.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_20.pkl.gz Training 106: 0 / 126623: Premsel loss 0.2297, acc 0.9007 (0.9288 / 0.8726) Training 106: 7000 / 126623: Premsel loss 0.2452, acc 0.8929 (0.9193 / 0.8666) Training 106: 14000 / 126623: Premsel loss 0.2396, acc 0.8958 (0.9221 / 0.8695) Training 106: 21000 / 126623: Premsel loss 0.2460, acc 0.8930 (0.9195 / 0.8665) Training 106: 28000 / 126623: Premsel loss 0.2430, acc 0.8939 (0.9209 / 0.8668) Training 106: 35000 / 126623: Premsel loss 0.2389, acc 0.8968 (0.9203 / 0.8733) Training 106: 42000 / 126623: Premsel loss 0.2457, acc 0.8929 (0.9167 / 0.8690) Training 106: 49000 / 126623: Premsel loss 0.2368, acc 0.8980 (0.9243 / 0.8716) Training 106: 56000 / 126623: Premsel loss 0.2412, acc 0.8950 (0.9181 / 0.8719) Training 106: 63000 / 126623: Premsel loss 0.2358, acc 0.8979 (0.9249 / 0.8709) Training 106: 70000 / 126623: Premsel loss 0.2359, acc 0.8973 (0.9232 / 0.8714) Training 106: 77000 / 126623: Premsel loss 0.2417, acc 0.8951 (0.9234 / 0.8668) Training 106: 84000 / 126623: Premsel loss 0.2319, acc 0.8996 (0.9235 / 0.8757) Training 106: 91000 / 126623: Premsel loss 0.2366, acc 0.8970 (0.9236 / 0.8704) Training 106: 98000 / 126623: Premsel loss 0.2411, acc 0.8949 (0.9230 / 0.8668) Training 106: 105000 / 126623: Premsel loss 0.2374, acc 0.8972 (0.9266 / 0.8677) Training 106: 112000 / 126623: Premsel loss 0.2381, acc 0.8979 (0.9245 / 0.8713) Training 106: 119000 / 126623: Premsel loss 0.2319, acc 0.9003 (0.9249 / 0.8757) Training 106: 126000 / 126623: Premsel loss 0.2456, acc 0.8931 (0.9216 / 0.8645) Evaluation 106: Premsel loss 0.2360, acc 0.8985 (0.9460 / 0.8511) Training 107: 0 / 126623: Premsel loss 0.2342, acc 0.8995 (0.9255 / 0.8736) Training 107: 7000 / 126623: Premsel loss 0.2420, acc 0.8939 (0.9173 / 0.8705) Training 107: 14000 / 126623: Premsel loss 0.2417, acc 0.8951 (0.9223 / 0.8678) Training 107: 21000 / 126623: Premsel loss 0.2274, acc 0.9017 (0.9285 / 0.8750) Training 107: 28000 / 126623: Premsel loss 0.2332, acc 0.8990 (0.9257 / 0.8723) Training 107: 35000 / 126623: Premsel loss 0.2377, acc 0.8976 (0.9257 / 0.8696) Training 107: 42000 / 126623: Premsel loss 0.2425, acc 0.8941 (0.9217 / 0.8665) Training 107: 49000 / 126623: Premsel loss 0.2458, acc 0.8928 (0.9233 / 0.8623) Training 107: 56000 / 126623: Premsel loss 0.2430, acc 0.8952 (0.9251 / 0.8653) Training 107: 63000 / 126623: Premsel loss 0.2362, acc 0.8973 (0.9211 / 0.8735) Training 107: 70000 / 126623: Premsel loss 0.2487, acc 0.8926 (0.9170 / 0.8682) Training 107: 77000 / 126623: Premsel loss 0.2405, acc 0.8960 (0.9225 / 0.8695) Training 107: 84000 / 126623: Premsel loss 0.2432, acc 0.8937 (0.9227 / 0.8648) Training 107: 91000 / 126623: Premsel loss 0.2392, acc 0.8963 (0.9207 / 0.8718) Training 107: 98000 / 126623: Premsel loss 0.2416, acc 0.8948 (0.9138 / 0.8758) Training 107: 105000 / 126623: Premsel loss 0.2397, acc 0.8956 (0.9222 / 0.8690) Training 107: 112000 / 126623: Premsel loss 0.2395, acc 0.8967 (0.9241 / 0.8693) Training 107: 119000 / 126623: Premsel loss 0.2414, acc 0.8952 (0.9217 / 0.8686) Training 107: 126000 / 126623: Premsel loss 0.2441, acc 0.8941 (0.9224 / 0.8658) Evaluation 107: Premsel loss 0.2347, acc 0.8993 (0.9237 / 0.8750) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_04.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz Training 108: 0 / 128628: Premsel loss 0.2394, acc 0.8969 (0.9228 / 0.8710) Training 108: 7000 / 128628: Premsel loss 0.2138, acc 0.9088 (0.9347 / 0.8830) Training 108: 14000 / 128628: Premsel loss 0.2076, acc 0.9114 (0.9365 / 0.8863) Training 108: 21000 / 128628: Premsel loss 0.2083, acc 0.9117 (0.9365 / 0.8870) Training 108: 28000 / 128628: Premsel loss 0.2138, acc 0.9087 (0.9284 / 0.8890) Training 108: 35000 / 128628: Premsel loss 0.2061, acc 0.9129 (0.9344 / 0.8913) Training 108: 42000 / 128628: Premsel loss 0.2160, acc 0.9067 (0.9272 / 0.8861) Training 108: 49000 / 128628: Premsel loss 0.2169, acc 0.9072 (0.9331 / 0.8813) Training 108: 56000 / 128628: Premsel loss 0.2172, acc 0.9064 (0.9312 / 0.8815) Training 108: 63000 / 128628: Premsel loss 0.2150, acc 0.9082 (0.9343 / 0.8822) Training 108: 70000 / 128628: Premsel loss 0.2121, acc 0.9100 (0.9376 / 0.8824) Training 108: 77000 / 128628: Premsel loss 0.2114, acc 0.9103 (0.9309 / 0.8896) Training 108: 84000 / 128628: Premsel loss 0.2086, acc 0.9110 (0.9341 / 0.8879) Training 108: 91000 / 128628: Premsel loss 0.2069, acc 0.9114 (0.9342 / 0.8885) Training 108: 98000 / 128628: Premsel loss 0.2083, acc 0.9104 (0.9334 / 0.8874) Training 108: 105000 / 128628: Premsel loss 0.2135, acc 0.9091 (0.9349 / 0.8833) Training 108: 112000 / 128628: Premsel loss 0.2128, acc 0.9098 (0.9331 / 0.8865) Training 108: 119000 / 128628: Premsel loss 0.2142, acc 0.9083 (0.9344 / 0.8823) Training 108: 126000 / 128628: Premsel loss 0.2055, acc 0.9125 (0.9373 / 0.8876) Evaluation 108: Premsel loss 0.2071, acc 0.9119 (0.9451 / 0.8788) Training 109: 0 / 128628: Premsel loss 0.2071, acc 0.9121 (0.9327 / 0.8914) Training 109: 7000 / 128628: Premsel loss 0.2096, acc 0.9111 (0.9381 / 0.8841) Training 109: 14000 / 128628: Premsel loss 0.2162, acc 0.9070 (0.9292 / 0.8848) Training 109: 21000 / 128628: Premsel loss 0.2123, acc 0.9104 (0.9334 / 0.8873) Training 109: 28000 / 128628: Premsel loss 0.2115, acc 0.9098 (0.9327 / 0.8870) Training 109: 35000 / 128628: Premsel loss 0.2084, acc 0.9110 (0.9348 / 0.8873) Training 109: 42000 / 128628: Premsel loss 0.2144, acc 0.9082 (0.9378 / 0.8786) Training 109: 49000 / 128628: Premsel loss 0.2143, acc 0.9087 (0.9298 / 0.8875) Training 109: 56000 / 128628: Premsel loss 0.2113, acc 0.9093 (0.9307 / 0.8879) Training 109: 63000 / 128628: Premsel loss 0.2101, acc 0.9109 (0.9341 / 0.8876) Training 109: 70000 / 128628: Premsel loss 0.2073, acc 0.9112 (0.9361 / 0.8862) Training 109: 77000 / 128628: Premsel loss 0.2057, acc 0.9132 (0.9399 / 0.8864) Training 109: 84000 / 128628: Premsel loss 0.2086, acc 0.9111 (0.9304 / 0.8918) Training 109: 91000 / 128628: Premsel loss 0.2073, acc 0.9116 (0.9357 / 0.8876) Training 109: 98000 / 128628: Premsel loss 0.2110, acc 0.9114 (0.9323 / 0.8905) Training 109: 105000 / 128628: Premsel loss 0.2128, acc 0.9088 (0.9347 / 0.8829) Training 109: 112000 / 128628: Premsel loss 0.2088, acc 0.9109 (0.9359 / 0.8860) Training 109: 119000 / 128628: Premsel loss 0.2090, acc 0.9106 (0.9337 / 0.8876) Training 109: 126000 / 128628: Premsel loss 0.2126, acc 0.9097 (0.9362 / 0.8832) Evaluation 109: Premsel loss 0.2125, acc 0.9090 (0.9387 / 0.8793) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_07.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz Training 110: 0 / 108677: Premsel loss 0.2124, acc 0.9090 (0.9325 / 0.8856) Training 110: 7000 / 108677: Premsel loss 0.2357, acc 0.8987 (0.9227 / 0.8746) Training 110: 14000 / 108677: Premsel loss 0.2324, acc 0.8990 (0.9260 / 0.8720) Training 110: 21000 / 108677: Premsel loss 0.2365, acc 0.8970 (0.9238 / 0.8702) Training 110: 28000 / 108677: Premsel loss 0.2317, acc 0.8991 (0.9240 / 0.8743) Training 110: 35000 / 108677: Premsel loss 0.2306, acc 0.9004 (0.9253 / 0.8756) Training 110: 42000 / 108677: Premsel loss 0.2242, acc 0.9033 (0.9273 / 0.8793) Training 110: 49000 / 108677: Premsel loss 0.2279, acc 0.9010 (0.9278 / 0.8743) Training 110: 56000 / 108677: Premsel loss 0.2289, acc 0.9000 (0.9232 / 0.8767) Training 110: 63000 / 108677: Premsel loss 0.2368, acc 0.8963 (0.9226 / 0.8699) Training 110: 70000 / 108677: Premsel loss 0.2250, acc 0.9028 (0.9305 / 0.8752) Training 110: 77000 / 108677: Premsel loss 0.2327, acc 0.8989 (0.9253 / 0.8725) Training 110: 84000 / 108677: Premsel loss 0.2292, acc 0.9011 (0.9267 / 0.8756) Training 110: 91000 / 108677: Premsel loss 0.2329, acc 0.8991 (0.9262 / 0.8720) Training 110: 98000 / 108677: Premsel loss 0.2332, acc 0.8991 (0.9203 / 0.8779) Training 110: 105000 / 108677: Premsel loss 0.2327, acc 0.8986 (0.9255 / 0.8717) Evaluation 110: Premsel loss 0.2295, acc 0.9006 (0.9189 / 0.8823) Training 111: 0 / 108677: Premsel loss 0.2281, acc 0.9010 (0.9252 / 0.8769) Training 111: 7000 / 108677: Premsel loss 0.2297, acc 0.9006 (0.9261 / 0.8751) Training 111: 14000 / 108677: Premsel loss 0.2234, acc 0.9039 (0.9276 / 0.8803) Training 111: 21000 / 108677: Premsel loss 0.2288, acc 0.9002 (0.9274 / 0.8730) Training 111: 28000 / 108677: Premsel loss 0.2252, acc 0.9033 (0.9274 / 0.8792) Training 111: 35000 / 108677: Premsel loss 0.2332, acc 0.8987 (0.9278 / 0.8697) Training 111: 42000 / 108677: Premsel loss 0.2255, acc 0.9027 (0.9268 / 0.8785) Training 111: 49000 / 108677: Premsel loss 0.2345, acc 0.8979 (0.9226 / 0.8731) Training 111: 56000 / 108677: Premsel loss 0.2311, acc 0.9003 (0.9269 / 0.8738) Training 111: 63000 / 108677: Premsel loss 0.2306, acc 0.8993 (0.9254 / 0.8733) Training 111: 70000 / 108677: Premsel loss 0.2445, acc 0.8922 (0.9195 / 0.8649) Training 111: 77000 / 108677: Premsel loss 0.2389, acc 0.8958 (0.9160 / 0.8756) Training 111: 84000 / 108677: Premsel loss 0.2389, acc 0.8956 (0.9193 / 0.8720) Training 111: 91000 / 108677: Premsel loss 0.2340, acc 0.8982 (0.9222 / 0.8742) Training 111: 98000 / 108677: Premsel loss 0.2403, acc 0.8951 (0.9174 / 0.8729) Training 111: 105000 / 108677: Premsel loss 0.2409, acc 0.8939 (0.9174 / 0.8704) Evaluation 111: Premsel loss 0.2323, acc 0.8987 (0.9259 / 0.8715) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_18.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-solo___out1.pkl.gz Training 112: 0 / 113424: Premsel loss 0.2391, acc 0.8947 (0.9166 / 0.8729) Training 112: 7000 / 113424: Premsel loss 0.2303, acc 0.9000 (0.9262 / 0.8738) Training 112: 14000 / 113424: Premsel loss 0.2207, acc 0.9055 (0.9266 / 0.8844) Training 112: 21000 / 113424: Premsel loss 0.2246, acc 0.9037 (0.9279 / 0.8795) Training 112: 28000 / 113424: Premsel loss 0.2262, acc 0.9030 (0.9258 / 0.8803) Training 112: 35000 / 113424: Premsel loss 0.2230, acc 0.9044 (0.9289 / 0.8800) Training 112: 42000 / 113424: Premsel loss 0.2185, acc 0.9060 (0.9307 / 0.8813) Training 112: 49000 / 113424: Premsel loss 0.2230, acc 0.9036 (0.9301 / 0.8770) Training 112: 56000 / 113424: Premsel loss 0.2248, acc 0.9026 (0.9317 / 0.8736) Training 112: 63000 / 113424: Premsel loss 0.2274, acc 0.9012 (0.9297 / 0.8727) Training 112: 70000 / 113424: Premsel loss 0.2152, acc 0.9079 (0.9352 / 0.8807) Training 112: 77000 / 113424: Premsel loss 0.2239, acc 0.9039 (0.9306 / 0.8771) Training 112: 84000 / 113424: Premsel loss 0.2198, acc 0.9063 (0.9332 / 0.8794) Training 112: 91000 / 113424: Premsel loss 0.2245, acc 0.9033 (0.9308 / 0.8759) Training 112: 98000 / 113424: Premsel loss 0.2212, acc 0.9056 (0.9288 / 0.8823) Training 112: 105000 / 113424: Premsel loss 0.2349, acc 0.8985 (0.9246 / 0.8723) Training 112: 112000 / 113424: Premsel loss 0.2323, acc 0.8999 (0.9262 / 0.8735) Evaluation 112: Premsel loss 0.2246, acc 0.9033 (0.9240 / 0.8825) Training 113: 0 / 113424: Premsel loss 0.2251, acc 0.9028 (0.9311 / 0.8745) Training 113: 7000 / 113424: Premsel loss 0.2271, acc 0.9018 (0.9283 / 0.8752) Training 113: 14000 / 113424: Premsel loss 0.2275, acc 0.9015 (0.9284 / 0.8747) Training 113: 21000 / 113424: Premsel loss 0.2187, acc 0.9064 (0.9333 / 0.8795) Training 113: 28000 / 113424: Premsel loss 0.2202, acc 0.9062 (0.9337 / 0.8786) Training 113: 35000 / 113424: Premsel loss 0.2280, acc 0.9016 (0.9254 / 0.8777) Training 113: 42000 / 113424: Premsel loss 0.2208, acc 0.9058 (0.9306 / 0.8810) Training 113: 49000 / 113424: Premsel loss 0.2194, acc 0.9055 (0.9306 / 0.8804) Training 113: 56000 / 113424: Premsel loss 0.2233, acc 0.9043 (0.9303 / 0.8783) Training 113: 63000 / 113424: Premsel loss 0.2194, acc 0.9063 (0.9306 / 0.8819) Training 113: 70000 / 113424: Premsel loss 0.2232, acc 0.9039 (0.9299 / 0.8779) Training 113: 77000 / 113424: Premsel loss 0.2267, acc 0.9024 (0.9297 / 0.8750) Training 113: 84000 / 113424: Premsel loss 0.2280, acc 0.9021 (0.9278 / 0.8765) Training 113: 91000 / 113424: Premsel loss 0.2161, acc 0.9072 (0.9323 / 0.8820) Training 113: 98000 / 113424: Premsel loss 0.2207, acc 0.9057 (0.9303 / 0.8811) Training 113: 105000 / 113424: Premsel loss 0.2237, acc 0.9045 (0.9320 / 0.8769) Training 113: 112000 / 113424: Premsel loss 0.2184, acc 0.9068 (0.9252 / 0.8884) Evaluation 113: Premsel loss 0.2192, acc 0.9060 (0.9381 / 0.8740) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_09.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_19.pkl.gz Training 114: 0 / 113607: Premsel loss 0.2193, acc 0.9057 (0.9318 / 0.8796) Training 114: 7000 / 113607: Premsel loss 0.2372, acc 0.8971 (0.9242 / 0.8699) Training 114: 14000 / 113607: Premsel loss 0.2288, acc 0.9020 (0.9279 / 0.8762) Training 114: 21000 / 113607: Premsel loss 0.2358, acc 0.8977 (0.9198 / 0.8756) Training 114: 28000 / 113607: Premsel loss 0.2386, acc 0.8963 (0.9239 / 0.8686) Training 114: 35000 / 113607: Premsel loss 0.2365, acc 0.8979 (0.9181 / 0.8777) Training 114: 42000 / 113607: Premsel loss 0.2311, acc 0.8999 (0.9253 / 0.8744) Training 114: 49000 / 113607: Premsel loss 0.2341, acc 0.8993 (0.9219 / 0.8768) Training 114: 56000 / 113607: Premsel loss 0.2240, acc 0.9043 (0.9276 / 0.8809) Training 114: 63000 / 113607: Premsel loss 0.2283, acc 0.9015 (0.9237 / 0.8792) Training 114: 70000 / 113607: Premsel loss 0.2287, acc 0.9017 (0.9273 / 0.8761) Training 114: 77000 / 113607: Premsel loss 0.2293, acc 0.9003 (0.9272 / 0.8734) Training 114: 84000 / 113607: Premsel loss 0.2316, acc 0.9000 (0.9255 / 0.8746) Training 114: 91000 / 113607: Premsel loss 0.2280, acc 0.9020 (0.9249 / 0.8791) Training 114: 98000 / 113607: Premsel loss 0.2363, acc 0.8972 (0.9241 / 0.8702) Training 114: 105000 / 113607: Premsel loss 0.2410, acc 0.8958 (0.9198 / 0.8718) Training 114: 112000 / 113607: Premsel loss 0.2406, acc 0.8956 (0.9192 / 0.8721) Evaluation 114: Premsel loss 0.2292, acc 0.9016 (0.9276 / 0.8756) Training 115: 0 / 113607: Premsel loss 0.2399, acc 0.8962 (0.9204 / 0.8720) Training 115: 7000 / 113607: Premsel loss 0.2384, acc 0.8970 (0.9280 / 0.8659) Training 115: 14000 / 113607: Premsel loss 0.2295, acc 0.9016 (0.9255 / 0.8777) Training 115: 21000 / 113607: Premsel loss 0.2303, acc 0.9009 (0.9273 / 0.8745) Training 115: 28000 / 113607: Premsel loss 0.2366, acc 0.8973 (0.9284 / 0.8662) Training 115: 35000 / 113607: Premsel loss 0.2337, acc 0.8987 (0.9227 / 0.8747) Training 115: 42000 / 113607: Premsel loss 0.2288, acc 0.9025 (0.9277 / 0.8774) Training 115: 49000 / 113607: Premsel loss 0.2319, acc 0.9004 (0.9252 / 0.8755) Training 115: 56000 / 113607: Premsel loss 0.2253, acc 0.9047 (0.9290 / 0.8805) Training 115: 63000 / 113607: Premsel loss 0.2357, acc 0.8982 (0.9236 / 0.8729) Training 115: 70000 / 113607: Premsel loss 0.2371, acc 0.8980 (0.9262 / 0.8698) Training 115: 77000 / 113607: Premsel loss 0.2285, acc 0.9019 (0.9308 / 0.8730) Training 115: 84000 / 113607: Premsel loss 0.2313, acc 0.8999 (0.9235 / 0.8762) Training 115: 91000 / 113607: Premsel loss 0.2361, acc 0.8977 (0.9253 / 0.8700) Training 115: 98000 / 113607: Premsel loss 0.2284, acc 0.9013 (0.9282 / 0.8744) Training 115: 105000 / 113607: Premsel loss 0.2338, acc 0.8987 (0.9230 / 0.8744) Training 115: 112000 / 113607: Premsel loss 0.2338, acc 0.8988 (0.9213 / 0.8763) Evaluation 115: Premsel loss 0.2279, acc 0.9020 (0.9228 / 0.8811) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_17.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 116: 0 / 134070: Premsel loss 0.2326, acc 0.8994 (0.9220 / 0.8768) Training 116: 7000 / 134070: Premsel loss 0.2181, acc 0.9075 (0.9320 / 0.8829) Training 116: 14000 / 134070: Premsel loss 0.2187, acc 0.9067 (0.9361 / 0.8772) Training 116: 21000 / 134070: Premsel loss 0.2147, acc 0.9082 (0.9343 / 0.8821) Training 116: 28000 / 134070: Premsel loss 0.2250, acc 0.9032 (0.9284 / 0.8781) Training 116: 35000 / 134070: Premsel loss 0.2205, acc 0.9051 (0.9318 / 0.8785) Training 116: 42000 / 134070: Premsel loss 0.2172, acc 0.9076 (0.9329 / 0.8824) Training 116: 49000 / 134070: Premsel loss 0.2246, acc 0.9038 (0.9296 / 0.8779) Training 116: 56000 / 134070: Premsel loss 0.2158, acc 0.9078 (0.9338 / 0.8817) Training 116: 63000 / 134070: Premsel loss 0.2223, acc 0.9041 (0.9277 / 0.8805) Training 116: 70000 / 134070: Premsel loss 0.2235, acc 0.9042 (0.9285 / 0.8800) Training 116: 77000 / 134070: Premsel loss 0.2227, acc 0.9043 (0.9296 / 0.8791) Training 116: 84000 / 134070: Premsel loss 0.2233, acc 0.9048 (0.9311 / 0.8785) Training 116: 91000 / 134070: Premsel loss 0.2182, acc 0.9061 (0.9315 / 0.8808) Training 116: 98000 / 134070: Premsel loss 0.2239, acc 0.9048 (0.9277 / 0.8819) Training 116: 105000 / 134070: Premsel loss 0.2163, acc 0.9077 (0.9327 / 0.8828) Training 116: 112000 / 134070: Premsel loss 0.2195, acc 0.9057 (0.9305 / 0.8810) Training 116: 119000 / 134070: Premsel loss 0.2213, acc 0.9059 (0.9313 / 0.8805) Training 116: 126000 / 134070: Premsel loss 0.2279, acc 0.9013 (0.9286 / 0.8740) Training 116: 133000 / 134070: Premsel loss 0.2191, acc 0.9064 (0.9363 / 0.8765) Evaluation 116: Premsel loss 0.2200, acc 0.9059 (0.9422 / 0.8695) Training 117: 0 / 134070: Premsel loss 0.2240, acc 0.9038 (0.9298 / 0.8778) Training 117: 7000 / 134070: Premsel loss 0.2244, acc 0.9036 (0.9334 / 0.8737) Training 117: 14000 / 134070: Premsel loss 0.2182, acc 0.9068 (0.9322 / 0.8813) Training 117: 21000 / 134070: Premsel loss 0.2234, acc 0.9039 (0.9324 / 0.8754) Training 117: 28000 / 134070: Premsel loss 0.2246, acc 0.9032 (0.9331 / 0.8732) Training 117: 35000 / 134070: Premsel loss 0.2244, acc 0.9033 (0.9303 / 0.8762) Training 117: 42000 / 134070: Premsel loss 0.2260, acc 0.9027 (0.9254 / 0.8800) Training 117: 49000 / 134070: Premsel loss 0.2259, acc 0.9044 (0.9296 / 0.8792) Training 117: 56000 / 134070: Premsel loss 0.2303, acc 0.9001 (0.9302 / 0.8700) Training 117: 63000 / 134070: Premsel loss 0.2285, acc 0.9020 (0.9294 / 0.8746) Training 117: 70000 / 134070: Premsel loss 0.2237, acc 0.9037 (0.9261 / 0.8813) Training 117: 77000 / 134070: Premsel loss 0.2259, acc 0.9025 (0.9274 / 0.8776) Training 117: 84000 / 134070: Premsel loss 0.2255, acc 0.9025 (0.9269 / 0.8782) Training 117: 91000 / 134070: Premsel loss 0.2341, acc 0.8981 (0.9293 / 0.8668) Training 117: 98000 / 134070: Premsel loss 0.2230, acc 0.9042 (0.9316 / 0.8769) Training 117: 105000 / 134070: Premsel loss 0.2233, acc 0.9040 (0.9309 / 0.8771) Training 117: 112000 / 134070: Premsel loss 0.2211, acc 0.9042 (0.9288 / 0.8796) Training 117: 119000 / 134070: Premsel loss 0.2241, acc 0.9036 (0.9288 / 0.8785) Training 117: 126000 / 134070: Premsel loss 0.2236, acc 0.9043 (0.9293 / 0.8792) Training 117: 133000 / 134070: Premsel loss 0.2233, acc 0.9040 (0.9296 / 0.8784) Evaluation 117: Premsel loss 0.2196, acc 0.9059 (0.9270 / 0.8848) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t300-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_12.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz Training 118: 0 / 124844: Premsel loss 0.2203, acc 0.9054 (0.9316 / 0.8792) Training 118: 7000 / 124844: Premsel loss 0.2233, acc 0.9031 (0.9326 / 0.8737) Training 118: 14000 / 124844: Premsel loss 0.2302, acc 0.9004 (0.9310 / 0.8699) Training 118: 21000 / 124844: Premsel loss 0.2199, acc 0.9049 (0.9332 / 0.8766) Training 118: 28000 / 124844: Premsel loss 0.2232, acc 0.9028 (0.9297 / 0.8758) Training 118: 35000 / 124844: Premsel loss 0.2222, acc 0.9037 (0.9293 / 0.8781) Training 118: 42000 / 124844: Premsel loss 0.2300, acc 0.9001 (0.9248 / 0.8754) Training 118: 49000 / 124844: Premsel loss 0.2280, acc 0.9021 (0.9231 / 0.8811) Training 118: 56000 / 124844: Premsel loss 0.2294, acc 0.9004 (0.9256 / 0.8752) Training 118: 63000 / 124844: Premsel loss 0.2315, acc 0.8998 (0.9254 / 0.8742) Training 118: 70000 / 124844: Premsel loss 0.2360, acc 0.8973 (0.9228 / 0.8718) Training 118: 77000 / 124844: Premsel loss 0.2300, acc 0.8998 (0.9277 / 0.8718) Training 118: 84000 / 124844: Premsel loss 0.2310, acc 0.9012 (0.9248 / 0.8776) Training 118: 91000 / 124844: Premsel loss 0.2307, acc 0.8994 (0.9256 / 0.8732) Training 118: 98000 / 124844: Premsel loss 0.2331, acc 0.8988 (0.9284 / 0.8692) Training 118: 105000 / 124844: Premsel loss 0.2319, acc 0.9002 (0.9278 / 0.8726) Training 118: 112000 / 124844: Premsel loss 0.2250, acc 0.9024 (0.9284 / 0.8764) Training 118: 119000 / 124844: Premsel loss 0.2268, acc 0.9019 (0.9332 / 0.8706) Evaluation 118: Premsel loss 0.2305, acc 0.8999 (0.9216 / 0.8781) Training 119: 0 / 124844: Premsel loss 0.2316, acc 0.8992 (0.9289 / 0.8695) Training 119: 7000 / 124844: Premsel loss 0.2296, acc 0.9013 (0.9291 / 0.8735) Training 119: 14000 / 124844: Premsel loss 0.2280, acc 0.9016 (0.9319 / 0.8713) Training 119: 21000 / 124844: Premsel loss 0.2279, acc 0.9016 (0.9295 / 0.8738) Training 119: 28000 / 124844: Premsel loss 0.2401, acc 0.8949 (0.9238 / 0.8659) Training 119: 35000 / 124844: Premsel loss 0.2303, acc 0.9003 (0.9301 / 0.8706) Training 119: 42000 / 124844: Premsel loss 0.2331, acc 0.8988 (0.9246 / 0.8730) Training 119: 49000 / 124844: Premsel loss 0.2293, acc 0.9000 (0.9254 / 0.8745) Training 119: 56000 / 124844: Premsel loss 0.2270, acc 0.9012 (0.9264 / 0.8760) Training 119: 63000 / 124844: Premsel loss 0.2337, acc 0.8981 (0.9264 / 0.8698) Training 119: 70000 / 124844: Premsel loss 0.2346, acc 0.8984 (0.9262 / 0.8705) Training 119: 77000 / 124844: Premsel loss 0.2322, acc 0.8985 (0.9272 / 0.8697) Training 119: 84000 / 124844: Premsel loss 0.2333, acc 0.8988 (0.9244 / 0.8733) Training 119: 91000 / 124844: Premsel loss 0.2268, acc 0.9017 (0.9251 / 0.8782) Training 119: 98000 / 124844: Premsel loss 0.2344, acc 0.8980 (0.9315 / 0.8645) Training 119: 105000 / 124844: Premsel loss 0.2270, acc 0.9021 (0.9292 / 0.8750) Training 119: 112000 / 124844: Premsel loss 0.2309, acc 0.9004 (0.9291 / 0.8717) Training 119: 119000 / 124844: Premsel loss 0.2337, acc 0.8983 (0.9199 / 0.8768) Evaluation 119: Premsel loss 0.2287, acc 0.9010 (0.9275 / 0.8746) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_14.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz Training 120: 0 / 63687: Premsel loss 0.2314, acc 0.9002 (0.9270 / 0.8735) Training 120: 7000 / 63687: Premsel loss 0.2436, acc 0.8942 (0.9223 / 0.8660) Training 120: 14000 / 63687: Premsel loss 0.2460, acc 0.8927 (0.9215 / 0.8639) Training 120: 21000 / 63687: Premsel loss 0.2480, acc 0.8921 (0.9205 / 0.8636) Training 120: 28000 / 63687: Premsel loss 0.2429, acc 0.8942 (0.9186 / 0.8697) Training 120: 35000 / 63687: Premsel loss 0.2407, acc 0.8962 (0.9233 / 0.8692) Training 120: 42000 / 63687: Premsel loss 0.2325, acc 0.8992 (0.9232 / 0.8751) Training 120: 49000 / 63687: Premsel loss 0.2448, acc 0.8936 (0.9219 / 0.8652) Training 120: 56000 / 63687: Premsel loss 0.2356, acc 0.8977 (0.9257 / 0.8697) Training 120: 63000 / 63687: Premsel loss 0.2392, acc 0.8962 (0.9244 / 0.8680) Evaluation 120: Premsel loss 0.2365, acc 0.8973 (0.9390 / 0.8555) Training 121: 0 / 63687: Premsel loss 0.2383, acc 0.8971 (0.9246 / 0.8697) Training 121: 7000 / 63687: Premsel loss 0.2411, acc 0.8957 (0.9237 / 0.8677) Training 121: 14000 / 63687: Premsel loss 0.2398, acc 0.8955 (0.9207 / 0.8703) Training 121: 21000 / 63687: Premsel loss 0.2288, acc 0.9013 (0.9276 / 0.8751) Training 121: 28000 / 63687: Premsel loss 0.2374, acc 0.8973 (0.9240 / 0.8706) Training 121: 35000 / 63687: Premsel loss 0.2397, acc 0.8958 (0.9208 / 0.8708) Training 121: 42000 / 63687: Premsel loss 0.2434, acc 0.8935 (0.9190 / 0.8681) Training 121: 49000 / 63687: Premsel loss 0.2422, acc 0.8941 (0.9232 / 0.8649) Training 121: 56000 / 63687: Premsel loss 0.2391, acc 0.8963 (0.9212 / 0.8715) Training 121: 63000 / 63687: Premsel loss 0.2320, acc 0.9002 (0.9299 / 0.8705) Evaluation 121: Premsel loss 0.2306, acc 0.9006 (0.9283 / 0.8729) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_01.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_05.pkl.gz Training 122: 0 / 68670: Premsel loss 0.2328, acc 0.8990 (0.9300 / 0.8680) Training 122: 7000 / 68670: Premsel loss 0.2529, acc 0.8889 (0.9131 / 0.8647) Training 122: 14000 / 68670: Premsel loss 0.2565, acc 0.8874 (0.9108 / 0.8641) Training 122: 21000 / 68670: Premsel loss 0.2514, acc 0.8904 (0.9199 / 0.8609) Training 122: 28000 / 68670: Premsel loss 0.2494, acc 0.8924 (0.9190 / 0.8659) Training 122: 35000 / 68670: Premsel loss 0.2458, acc 0.8937 (0.9187 / 0.8688) Training 122: 42000 / 68670: Premsel loss 0.2488, acc 0.8916 (0.9156 / 0.8675) Training 122: 49000 / 68670: Premsel loss 0.2479, acc 0.8922 (0.9167 / 0.8678) Training 122: 56000 / 68670: Premsel loss 0.2504, acc 0.8919 (0.9189 / 0.8649) Training 122: 63000 / 68670: Premsel loss 0.2573, acc 0.8873 (0.9115 / 0.8631) Evaluation 122: Premsel loss 0.2469, acc 0.8933 (0.9294 / 0.8572) Training 123: 0 / 68670: Premsel loss 0.2467, acc 0.8930 (0.9172 / 0.8688) Training 123: 7000 / 68670: Premsel loss 0.2520, acc 0.8905 (0.9183 / 0.8627) Training 123: 14000 / 68670: Premsel loss 0.2468, acc 0.8934 (0.9219 / 0.8649) Training 123: 21000 / 68670: Premsel loss 0.2492, acc 0.8921 (0.9142 / 0.8701) Training 123: 28000 / 68670: Premsel loss 0.2483, acc 0.8921 (0.9181 / 0.8660) Training 123: 35000 / 68670: Premsel loss 0.2467, acc 0.8934 (0.9177 / 0.8691) Training 123: 42000 / 68670: Premsel loss 0.2501, acc 0.8913 (0.9151 / 0.8676) Training 123: 49000 / 68670: Premsel loss 0.2720, acc 0.8791 (0.9044 / 0.8537) Training 123: 56000 / 68670: Premsel loss 0.2667, acc 0.8820 (0.9042 / 0.8598) Training 123: 63000 / 68670: Premsel loss 0.2650, acc 0.8838 (0.9102 / 0.8574) Evaluation 123: Premsel loss 0.2572, acc 0.8883 (0.9370 / 0.8395) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_03.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz Training 124: 0 / 115885: Premsel loss 0.2539, acc 0.8894 (0.9124 / 0.8665) Training 124: 7000 / 115885: Premsel loss 0.2294, acc 0.9013 (0.9292 / 0.8733) Training 124: 14000 / 115885: Premsel loss 0.2358, acc 0.8979 (0.9266 / 0.8692) Training 124: 21000 / 115885: Premsel loss 0.2273, acc 0.9026 (0.9323 / 0.8729) Training 124: 28000 / 115885: Premsel loss 0.2286, acc 0.9011 (0.9238 / 0.8784) Training 124: 35000 / 115885: Premsel loss 0.2310, acc 0.9005 (0.9260 / 0.8750) Training 124: 42000 / 115885: Premsel loss 0.2228, acc 0.9039 (0.9258 / 0.8820) Training 124: 49000 / 115885: Premsel loss 0.2213, acc 0.9049 (0.9283 / 0.8814) Training 124: 56000 / 115885: Premsel loss 0.2186, acc 0.9068 (0.9302 / 0.8834) Training 124: 63000 / 115885: Premsel loss 0.2275, acc 0.9012 (0.9288 / 0.8737) Training 124: 70000 / 115885: Premsel loss 0.2190, acc 0.9061 (0.9300 / 0.8821) Training 124: 77000 / 115885: Premsel loss 0.2187, acc 0.9065 (0.9321 / 0.8810) Training 124: 84000 / 115885: Premsel loss 0.2207, acc 0.9047 (0.9310 / 0.8784) Training 124: 91000 / 115885: Premsel loss 0.2323, acc 0.8995 (0.9239 / 0.8750) Training 124: 98000 / 115885: Premsel loss 0.2248, acc 0.9020 (0.9299 / 0.8742) Training 124: 105000 / 115885: Premsel loss 0.2212, acc 0.9052 (0.9284 / 0.8819) Training 124: 112000 / 115885: Premsel loss 0.2222, acc 0.9045 (0.9313 / 0.8777) Evaluation 124: Premsel loss 0.2197, acc 0.9058 (0.9250 / 0.8867) Training 125: 0 / 115885: Premsel loss 0.2256, acc 0.9028 (0.9333 / 0.8722) Training 125: 7000 / 115885: Premsel loss 0.2197, acc 0.9053 (0.9336 / 0.8770) Training 125: 14000 / 115885: Premsel loss 0.2193, acc 0.9067 (0.9308 / 0.8825) Training 125: 21000 / 115885: Premsel loss 0.2185, acc 0.9067 (0.9286 / 0.8849) Training 125: 28000 / 115885: Premsel loss 0.2215, acc 0.9057 (0.9292 / 0.8821) Training 125: 35000 / 115885: Premsel loss 0.2176, acc 0.9072 (0.9346 / 0.8799) Training 125: 42000 / 115885: Premsel loss 0.2201, acc 0.9055 (0.9328 / 0.8782) Training 125: 49000 / 115885: Premsel loss 0.2218, acc 0.9050 (0.9335 / 0.8765) Training 125: 56000 / 115885: Premsel loss 0.2174, acc 0.9068 (0.9329 / 0.8807) Training 125: 63000 / 115885: Premsel loss 0.2183, acc 0.9063 (0.9314 / 0.8812) Training 125: 70000 / 115885: Premsel loss 0.2198, acc 0.9063 (0.9306 / 0.8820) Training 125: 77000 / 115885: Premsel loss 0.2232, acc 0.9042 (0.9307 / 0.8777) Training 125: 84000 / 115885: Premsel loss 0.2293, acc 0.9010 (0.9299 / 0.8721) Training 125: 91000 / 115885: Premsel loss 0.2231, acc 0.9044 (0.9325 / 0.8763) Training 125: 98000 / 115885: Premsel loss 0.2161, acc 0.9074 (0.9366 / 0.8783) Training 125: 105000 / 115885: Premsel loss 0.2201, acc 0.9052 (0.9296 / 0.8809) Training 125: 112000 / 115885: Premsel loss 0.2180, acc 0.9072 (0.9297 / 0.8847) Evaluation 125: Premsel loss 0.2167, acc 0.9072 (0.9301 / 0.8843) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_00.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz Training 126: 0 / 121459: Premsel loss 0.2173, acc 0.9079 (0.9345 / 0.8812) Training 126: 7000 / 121459: Premsel loss 0.2613, acc 0.8833 (0.9044 / 0.8622) Training 126: 14000 / 121459: Premsel loss 0.2500, acc 0.8903 (0.9098 / 0.8709) Training 126: 21000 / 121459: Premsel loss 0.2534, acc 0.8888 (0.9069 / 0.8706) Training 126: 28000 / 121459: Premsel loss 0.2578, acc 0.8868 (0.9053 / 0.8684) Training 126: 35000 / 121459: Premsel loss 0.2555, acc 0.8880 (0.9077 / 0.8683) Training 126: 42000 / 121459: Premsel loss 0.2565, acc 0.8867 (0.9015 / 0.8719) Training 126: 49000 / 121459: Premsel loss 0.2568, acc 0.8863 (0.9032 / 0.8695) Training 126: 56000 / 121459: Premsel loss 0.2573, acc 0.8871 (0.9072 / 0.8670) Training 126: 63000 / 121459: Premsel loss 0.2655, acc 0.8836 (0.8993 / 0.8679) Training 126: 70000 / 121459: Premsel loss 0.2543, acc 0.8893 (0.9020 / 0.8767) Training 126: 77000 / 121459: Premsel loss 0.2613, acc 0.8853 (0.9061 / 0.8644) Training 126: 84000 / 121459: Premsel loss 0.2519, acc 0.8903 (0.9110 / 0.8696) Training 126: 91000 / 121459: Premsel loss 0.2527, acc 0.8892 (0.9077 / 0.8707) Training 126: 98000 / 121459: Premsel loss 0.2593, acc 0.8851 (0.9031 / 0.8670) Training 126: 105000 / 121459: Premsel loss 0.2661, acc 0.8826 (0.8994 / 0.8659) Training 126: 112000 / 121459: Premsel loss 0.2525, acc 0.8897 (0.9035 / 0.8758) Training 126: 119000 / 121459: Premsel loss 0.2656, acc 0.8818 (0.9063 / 0.8572) Evaluation 126: Premsel loss 0.2570, acc 0.8868 (0.9140 / 0.8597) Training 127: 0 / 121459: Premsel loss 0.2581, acc 0.8845 (0.9051 / 0.8640) Training 127: 7000 / 121459: Premsel loss 0.2592, acc 0.8862 (0.9015 / 0.8708) Training 127: 14000 / 121459: Premsel loss 0.2547, acc 0.8868 (0.9080 / 0.8656) Training 127: 21000 / 121459: Premsel loss 0.2577, acc 0.8868 (0.9067 / 0.8669) Training 127: 28000 / 121459: Premsel loss 0.2538, acc 0.8875 (0.9063 / 0.8687) Training 127: 35000 / 121459: Premsel loss 0.2539, acc 0.8879 (0.9077 / 0.8681) Training 127: 42000 / 121459: Premsel loss 0.2564, acc 0.8866 (0.9134 / 0.8598) Training 127: 49000 / 121459: Premsel loss 0.2559, acc 0.8869 (0.9045 / 0.8693) Training 127: 56000 / 121459: Premsel loss 0.2571, acc 0.8863 (0.9094 / 0.8632) Training 127: 63000 / 121459: Premsel loss 0.2505, acc 0.8894 (0.9073 / 0.8715) Training 127: 70000 / 121459: Premsel loss 0.2534, acc 0.8886 (0.9071 / 0.8701) Training 127: 77000 / 121459: Premsel loss 0.2641, acc 0.8823 (0.8975 / 0.8671) Training 127: 84000 / 121459: Premsel loss 0.2494, acc 0.8904 (0.9100 / 0.8708) Training 127: 91000 / 121459: Premsel loss 0.2549, acc 0.8879 (0.9099 / 0.8659) Training 127: 98000 / 121459: Premsel loss 0.2560, acc 0.8874 (0.9040 / 0.8708) Training 127: 105000 / 121459: Premsel loss 0.2506, acc 0.8898 (0.9063 / 0.8732) Training 127: 112000 / 121459: Premsel loss 0.2585, acc 0.8856 (0.9010 / 0.8702) Training 127: 119000 / 121459: Premsel loss 0.2612, acc 0.8841 (0.9025 / 0.8657) Evaluation 127: Premsel loss 0.2596, acc 0.8858 (0.8978 / 0.8739) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_10.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l16000-e0.15+coop-mzr02___out1.pkl.gz Training 128: 0 / 128449: Premsel loss 0.2590, acc 0.8862 (0.8999 / 0.8724) Training 128: 7000 / 128449: Premsel loss 0.2473, acc 0.8911 (0.9164 / 0.8657) Training 128: 14000 / 128449: Premsel loss 0.2522, acc 0.8887 (0.9099 / 0.8676) Training 128: 21000 / 128449: Premsel loss 0.2442, acc 0.8932 (0.9197 / 0.8668) Training 128: 28000 / 128449: Premsel loss 0.2471, acc 0.8919 (0.9126 / 0.8712) Training 128: 35000 / 128449: Premsel loss 0.2462, acc 0.8928 (0.9153 / 0.8703) Training 128: 42000 / 128449: Premsel loss 0.2460, acc 0.8927 (0.9142 / 0.8712) Training 128: 49000 / 128449: Premsel loss 0.2414, acc 0.8941 (0.9166 / 0.8715) Training 128: 56000 / 128449: Premsel loss 0.2400, acc 0.8948 (0.9120 / 0.8776) Training 128: 63000 / 128449: Premsel loss 0.2486, acc 0.8904 (0.9159 / 0.8649) Training 128: 70000 / 128449: Premsel loss 0.2492, acc 0.8907 (0.9123 / 0.8691) Training 128: 77000 / 128449: Premsel loss 0.2497, acc 0.8904 (0.9089 / 0.8718) Training 128: 84000 / 128449: Premsel loss 0.2479, acc 0.8917 (0.9086 / 0.8747) Training 128: 91000 / 128449: Premsel loss 0.2468, acc 0.8920 (0.9165 / 0.8676) Training 128: 98000 / 128449: Premsel loss 0.2479, acc 0.8910 (0.9152 / 0.8669) Training 128: 105000 / 128449: Premsel loss 0.2406, acc 0.8959 (0.9187 / 0.8732) Training 128: 112000 / 128449: Premsel loss 0.2451, acc 0.8919 (0.9117 / 0.8721) Training 128: 119000 / 128449: Premsel loss 0.2427, acc 0.8943 (0.9163 / 0.8722) Training 128: 126000 / 128449: Premsel loss 0.2439, acc 0.8929 (0.9172 / 0.8686) Evaluation 128: Premsel loss 0.2450, acc 0.8928 (0.8939 / 0.8917) Training 129: 0 / 128449: Premsel loss 0.2489, acc 0.8910 (0.9143 / 0.8677) Training 129: 7000 / 128449: Premsel loss 0.2453, acc 0.8929 (0.9151 / 0.8706) Training 129: 14000 / 128449: Premsel loss 0.2513, acc 0.8895 (0.9196 / 0.8593) Training 129: 21000 / 128449: Premsel loss 0.2422, acc 0.8950 (0.9149 / 0.8751) Training 129: 28000 / 128449: Premsel loss 0.2409, acc 0.8947 (0.9170 / 0.8725) Training 129: 35000 / 128449: Premsel loss 0.2405, acc 0.8945 (0.9188 / 0.8701) Training 129: 42000 / 128449: Premsel loss 0.2453, acc 0.8926 (0.9108 / 0.8744) Training 129: 49000 / 128449: Premsel loss 0.2484, acc 0.8902 (0.9127 / 0.8678) Training 129: 56000 / 128449: Premsel loss 0.2495, acc 0.8898 (0.9138 / 0.8658) ^[^[[BTraining 129: 63000 / 128449: Premsel loss 0.2435, acc 0.8940 (0.9128 / 0.8751) Training 129: 70000 / 128449: Premsel loss 0.2459, acc 0.8927 (0.9153 / 0.8701) Training 129: 77000 / 128449: Premsel loss 0.2410, acc 0.8943 (0.9180 / 0.8706) Training 129: 84000 / 128449: Premsel loss 0.2524, acc 0.8897 (0.9112 / 0.8681) Training 129: 91000 / 128449: Premsel loss 0.2471, acc 0.8917 (0.9132 / 0.8703) Training 129: 98000 / 128449: Premsel loss 0.2527, acc 0.8883 (0.9050 / 0.8717) Training 129: 105000 / 128449: Premsel loss 0.2493, acc 0.8905 (0.9108 / 0.8702) Training 129: 112000 / 128449: Premsel loss 0.2464, acc 0.8921 (0.9090 / 0.8753) Training 129: 119000 / 128449: Premsel loss 0.2428, acc 0.8938 (0.9168 / 0.8708) Training 129: 126000 / 128449: Premsel loss 0.2376, acc 0.8960 (0.9154 / 0.8766) Evaluation 129: Premsel loss 0.2429, acc 0.8931 (0.9081 / 0.8780) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_13.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_02.pkl.gz Training 130: 0 / 125122: Premsel loss 0.2461, acc 0.8916 (0.9132 / 0.8700) Training 130: 7000 / 125122: Premsel loss 0.2521, acc 0.8892 (0.9165 / 0.8620) Training 130: 14000 / 125122: Premsel loss 0.2742, acc 0.8790 (0.9072 / 0.8507) Training 130: 21000 / 125122: Premsel loss 0.2798, acc 0.8751 (0.8998 / 0.8505) Training 130: 28000 / 125122: Premsel loss 0.2676, acc 0.8811 (0.9101 / 0.8521) Training 130: 35000 / 125122: Premsel loss 0.2693, acc 0.8810 (0.9075 / 0.8546) Training 130: 42000 / 125122: Premsel loss 0.2586, acc 0.8862 (0.9092 / 0.8632) Training 130: 49000 / 125122: Premsel loss 0.2551, acc 0.8880 (0.9139 / 0.8621) Training 130: 56000 / 125122: Premsel loss 0.2578, acc 0.8867 (0.9149 / 0.8584) Training 130: 63000 / 125122: Premsel loss 0.2511, acc 0.8892 (0.9171 / 0.8613) Training 130: 70000 / 125122: Premsel loss 0.2503, acc 0.8905 (0.9201 / 0.8608) Training 130: 77000 / 125122: Premsel loss 0.2508, acc 0.8898 (0.9167 / 0.8629) Training 130: 84000 / 125122: Premsel loss 0.2521, acc 0.8895 (0.9118 / 0.8672) Training 130: 91000 / 125122: Premsel loss 0.2523, acc 0.8901 (0.9150 / 0.8651) Training 130: 98000 / 125122: Premsel loss 0.2487, acc 0.8917 (0.9198 / 0.8636) Training 130: 105000 / 125122: Premsel loss 0.2537, acc 0.8892 (0.9186 / 0.8599) Training 130: 112000 / 125122: Premsel loss 0.2548, acc 0.8878 (0.9155 / 0.8602) Training 130: 119000 / 125122: Premsel loss 0.2482, acc 0.8917 (0.9180 / 0.8654) Evaluation 130: Premsel loss 0.2483, acc 0.8918 (0.9059 / 0.8777) Training 131: 0 / 125122: Premsel loss 0.2503, acc 0.8906 (0.9202 / 0.8609) Training 131: 7000 / 125122: Premsel loss 0.2523, acc 0.8891 (0.9158 / 0.8624) Training 131: 14000 / 125122: Premsel loss 0.2505, acc 0.8908 (0.9182 / 0.8633) Training 131: 21000 / 125122: Premsel loss 0.2488, acc 0.8909 (0.9175 / 0.8644) Training 131: 28000 / 125122: Premsel loss 0.2469, acc 0.8922 (0.9211 / 0.8634) Training 131: 35000 / 125122: Premsel loss 0.2439, acc 0.8938 (0.9195 / 0.8682) Training 131: 42000 / 125122: Premsel loss 0.2506, acc 0.8908 (0.9217 / 0.8599) Training 131: 49000 / 125122: Premsel loss 0.2490, acc 0.8911 (0.9191 / 0.8630) Training 131: 56000 / 125122: Premsel loss 0.2448, acc 0.8933 (0.9194 / 0.8671) Training 131: 63000 / 125122: Premsel loss 0.2484, acc 0.8911 (0.9144 / 0.8678) Training 131: 70000 / 125122: Premsel loss 0.2470, acc 0.8926 (0.9200 / 0.8651) Training 131: 77000 / 125122: Premsel loss 0.2542, acc 0.8889 (0.9165 / 0.8614) Training 131: 84000 / 125122: Premsel loss 0.2531, acc 0.8892 (0.9167 / 0.8618) Training 131: 91000 / 125122: Premsel loss 0.2552, acc 0.8887 (0.9174 / 0.8601) Training 131: 98000 / 125122: Premsel loss 0.2479, acc 0.8919 (0.9228 / 0.8610) Training 131: 105000 / 125122: Premsel loss 0.2503, acc 0.8914 (0.9164 / 0.8664) Training 131: 112000 / 125122: Premsel loss 0.2585, acc 0.8866 (0.9117 / 0.8615) Training 131: 119000 / 125122: Premsel loss 0.2450, acc 0.8942 (0.9206 / 0.8679) Evaluation 131: Premsel loss 0.2474, acc 0.8925 (0.9255 / 0.8595) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_15.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz Training 132: 0 / 128581: Premsel loss 0.2460, acc 0.8933 (0.9211 / 0.8656) Training 132: 7000 / 128581: Premsel loss 0.2431, acc 0.8939 (0.9209 / 0.8668) Training 132: 14000 / 128581: Premsel loss 0.2382, acc 0.8966 (0.9194 / 0.8739) Training 132: 21000 / 128581: Premsel loss 0.2526, acc 0.8886 (0.9096 / 0.8676) Training 132: 28000 / 128581: Premsel loss 0.2705, acc 0.8805 (0.9020 / 0.8590) Training 132: 35000 / 128581: Premsel loss 0.2545, acc 0.8891 (0.9129 / 0.8652) Training 132: 42000 / 128581: Premsel loss 0.2481, acc 0.8910 (0.9132 / 0.8688) Training 132: 49000 / 128581: Premsel loss 0.2397, acc 0.8953 (0.9189 / 0.8717) Training 132: 56000 / 128581: Premsel loss 0.2393, acc 0.8957 (0.9202 / 0.8712) Training 132: 63000 / 128581: Premsel loss 0.2453, acc 0.8937 (0.9189 / 0.8685) Training 132: 70000 / 128581: Premsel loss 0.2483, acc 0.8908 (0.9175 / 0.8640) Training 132: 77000 / 128581: Premsel loss 0.2441, acc 0.8936 (0.9217 / 0.8654) Training 132: 84000 / 128581: Premsel loss 0.2435, acc 0.8942 (0.9153 / 0.8732) Training 132: 91000 / 128581: Premsel loss 0.2455, acc 0.8927 (0.9173 / 0.8680) Training 132: 98000 / 128581: Premsel loss 0.2489, acc 0.8916 (0.9177 / 0.8655) Training 132: 105000 / 128581: Premsel loss 0.2383, acc 0.8966 (0.9200 / 0.8731) Training 132: 112000 / 128581: Premsel loss 0.2449, acc 0.8927 (0.9167 / 0.8686) Training 132: 119000 / 128581: Premsel loss 0.2444, acc 0.8938 (0.9231 / 0.8644) Training 132: 126000 / 128581: Premsel loss 0.2479, acc 0.8919 (0.9115 / 0.8724) Evaluation 132: Premsel loss 0.2512, acc 0.8898 (0.9291 / 0.8505) Training 133: 0 / 128581: Premsel loss 0.2504, acc 0.8899 (0.9148 / 0.8650) Training 133: 7000 / 128581: Premsel loss 0.2465, acc 0.8927 (0.9176 / 0.8677) Training 133: 14000 / 128581: Premsel loss 0.2413, acc 0.8948 (0.9211 / 0.8684) Training 133: 21000 / 128581: Premsel loss 0.2409, acc 0.8962 (0.9180 / 0.8745) Training 133: 28000 / 128581: Premsel loss 0.2440, acc 0.8942 (0.9175 / 0.8708) Training 133: 35000 / 128581: Premsel loss 0.2444, acc 0.8930 (0.9187 / 0.8673) Training 133: 42000 / 128581: Premsel loss 0.2446, acc 0.8931 (0.9199 / 0.8663) Training 133: 49000 / 128581: Premsel loss 0.2489, acc 0.8907 (0.9173 / 0.8642) Training 133: 56000 / 128581: Premsel loss 0.2405, acc 0.8951 (0.9195 / 0.8707) Training 133: 63000 / 128581: Premsel loss 0.2364, acc 0.8972 (0.9204 / 0.8739) Training 133: 70000 / 128581: Premsel loss 0.2417, acc 0.8939 (0.9215 / 0.8662) Training 133: 77000 / 128581: Premsel loss 0.2330, acc 0.8994 (0.9248 / 0.8740) Training 133: 84000 / 128581: Premsel loss 0.2409, acc 0.8947 (0.9215 / 0.8679) Training 133: 91000 / 128581: Premsel loss 0.2409, acc 0.8942 (0.9225 / 0.8659) Training 133: 98000 / 128581: Premsel loss 0.2450, acc 0.8932 (0.9169 / 0.8695) Training 133: 105000 / 128581: Premsel loss 0.2495, acc 0.8902 (0.9100 / 0.8705) Training 133: 112000 / 128581: Premsel loss 0.2445, acc 0.8940 (0.9198 / 0.8682) Training 133: 119000 / 128581: Premsel loss 0.2437, acc 0.8927 (0.9137 / 0.8718) Training 133: 126000 / 128581: Premsel loss 0.2451, acc 0.8932 (0.9180 / 0.8684) Evaluation 133: Premsel loss 0.2425, acc 0.8943 (0.9051 / 0.8834) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo2___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_25.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l700-e0.20+coop-mzr02___out1.pkl.gz Training 134: 0 / 86443: Premsel loss 0.2406, acc 0.8950 (0.9231 / 0.8669) Training 134: 7000 / 86443: Premsel loss 0.2481, acc 0.8908 (0.9171 / 0.8645) Training 134: 14000 / 86443: Premsel loss 0.2457, acc 0.8926 (0.9190 / 0.8663) Training 134: 21000 / 86443: Premsel loss 0.2516, acc 0.8910 (0.9179 / 0.8642) Training 134: 28000 / 86443: Premsel loss 0.2453, acc 0.8931 (0.9184 / 0.8678) Training 134: 35000 / 86443: Premsel loss 0.2483, acc 0.8908 (0.9168 / 0.8648) Training 134: 42000 / 86443: Premsel loss 0.2468, acc 0.8928 (0.9175 / 0.8681) Training 134: 49000 / 86443: Premsel loss 0.2493, acc 0.8905 (0.9161 / 0.8649) Training 134: 56000 / 86443: Premsel loss 0.2436, acc 0.8936 (0.9197 / 0.8675) Training 134: 63000 / 86443: Premsel loss 0.2472, acc 0.8920 (0.9161 / 0.8680) Training 134: 70000 / 86443: Premsel loss 0.2487, acc 0.8916 (0.9173 / 0.8658) Training 134: 77000 / 86443: Premsel loss 0.2558, acc 0.8868 (0.9173 / 0.8563) Training 134: 84000 / 86443: Premsel loss 0.2526, acc 0.8894 (0.9146 / 0.8643) Evaluation 134: Premsel loss 0.2474, acc 0.8922 (0.9179 / 0.8666) Training 135: 0 / 86443: Premsel loss 0.2494, acc 0.8911 (0.9216 / 0.8607) Training 135: 7000 / 86443: Premsel loss 0.2486, acc 0.8917 (0.9167 / 0.8666) Training 135: 14000 / 86443: Premsel loss 0.2392, acc 0.8960 (0.9221 / 0.8700) Training 135: 21000 / 86443: Premsel loss 0.2486, acc 0.8909 (0.9168 / 0.8649) Training 135: 28000 / 86443: Premsel loss 0.2495, acc 0.8909 (0.9182 / 0.8636) Training 135: 35000 / 86443: Premsel loss 0.2491, acc 0.8912 (0.9169 / 0.8655) Training 135: 42000 / 86443: Premsel loss 0.2782, acc 0.8768 (0.9064 / 0.8471) Training 135: 49000 / 86443: Premsel loss 0.2609, acc 0.8849 (0.9108 / 0.8589) Training 135: 56000 / 86443: Premsel loss 0.2659, acc 0.8822 (0.9111 / 0.8534) Training 135: 63000 / 86443: Premsel loss 0.2572, acc 0.8874 (0.9188 / 0.8560) Training 135: 70000 / 86443: Premsel loss 0.2526, acc 0.8889 (0.9179 / 0.8598) Training 135: 77000 / 86443: Premsel loss 0.2526, acc 0.8899 (0.9181 / 0.8618) Training 135: 84000 / 86443: Premsel loss 0.2585, acc 0.8860 (0.9125 / 0.8594) Evaluation 135: Premsel loss 0.2543, acc 0.8883 (0.9226 / 0.8540) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_11.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d50-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz Training 136: 0 / 126682: Premsel loss 0.2546, acc 0.8880 (0.9143 / 0.8617) Training 136: 7000 / 126682: Premsel loss 0.2624, acc 0.8842 (0.9049 / 0.8635) Training 136: 14000 / 126682: Premsel loss 0.2621, acc 0.8836 (0.9110 / 0.8562) Training 136: 21000 / 126682: Premsel loss 0.2640, acc 0.8826 (0.9079 / 0.8573) Training 136: 28000 / 126682: Premsel loss 0.2661, acc 0.8813 (0.9014 / 0.8612) Training 136: 35000 / 126682: Premsel loss 0.2612, acc 0.8842 (0.9027 / 0.8657) Training 136: 42000 / 126682: Premsel loss 0.2615, acc 0.8834 (0.9071 / 0.8596) Training 136: 49000 / 126682: Premsel loss 0.2529, acc 0.8879 (0.9109 / 0.8649) Training 136: 56000 / 126682: Premsel loss 0.2616, acc 0.8840 (0.9054 / 0.8627) Training 136: 63000 / 126682: Premsel loss 0.2483, acc 0.8917 (0.9126 / 0.8707) Training 136: 70000 / 126682: Premsel loss 0.2631, acc 0.8838 (0.9033 / 0.8644) Training 136: 77000 / 126682: Premsel loss 0.2541, acc 0.8886 (0.9090 / 0.8681) Training 136: 84000 / 126682: Premsel loss 0.2639, acc 0.8838 (0.9072 / 0.8605) Training 136: 91000 / 126682: Premsel loss 0.2699, acc 0.8808 (0.9011 / 0.8604) Training 136: 98000 / 126682: Premsel loss 0.2670, acc 0.8819 (0.9028 / 0.8610) Training 136: 105000 / 126682: Premsel loss 0.2678, acc 0.8806 (0.9019 / 0.8592) Training 136: 112000 / 126682: Premsel loss 0.2595, acc 0.8854 (0.9047 / 0.8660) Training 136: 119000 / 126682: Premsel loss 0.2587, acc 0.8852 (0.9035 / 0.8668) Training 136: 126000 / 126682: Premsel loss 0.2666, acc 0.8810 (0.9079 / 0.8541) Evaluation 136: Premsel loss 0.2578, acc 0.8859 (0.9124 / 0.8593) Training 137: 0 / 126682: Premsel loss 0.2672, acc 0.8808 (0.8996 / 0.8621) Training 137: 7000 / 126682: Premsel loss 0.2529, acc 0.8884 (0.9097 / 0.8670) Training 137: 14000 / 126682: Premsel loss 0.2625, acc 0.8827 (0.9081 / 0.8573) Training 137: 21000 / 126682: Premsel loss 0.2673, acc 0.8804 (0.9062 / 0.8547) Training 137: 28000 / 126682: Premsel loss 0.2634, acc 0.8837 (0.9052 / 0.8622) Training 137: 35000 / 126682: Premsel loss 0.2631, acc 0.8834 (0.8986 / 0.8682) Training 137: 42000 / 126682: Premsel loss 0.2657, acc 0.8829 (0.9044 / 0.8614) Training 137: 49000 / 126682: Premsel loss 0.2655, acc 0.8819 (0.9069 / 0.8568) Training 137: 56000 / 126682: Premsel loss 0.2649, acc 0.8827 (0.9048 / 0.8606) Training 137: 63000 / 126682: Premsel loss 0.2703, acc 0.8806 (0.9032 / 0.8579) Training 137: 70000 / 126682: Premsel loss 0.2581, acc 0.8870 (0.9090 / 0.8649) Training 137: 77000 / 126682: Premsel loss 0.2545, acc 0.8879 (0.9154 / 0.8605) Training 137: 84000 / 126682: Premsel loss 0.2678, acc 0.8810 (0.9023 / 0.8598) Training 137: 91000 / 126682: Premsel loss 0.2597, acc 0.8842 (0.9066 / 0.8618) Training 137: 98000 / 126682: Premsel loss 0.2566, acc 0.8871 (0.9061 / 0.8681) Training 137: 105000 / 126682: Premsel loss 0.2519, acc 0.8884 (0.9068 / 0.8699) Training 137: 112000 / 126682: Premsel loss 0.2552, acc 0.8861 (0.9040 / 0.8682) Training 137: 119000 / 126682: Premsel loss 0.2603, acc 0.8851 (0.9091 / 0.8611) Training 137: 126000 / 126682: Premsel loss 0.2680, acc 0.8808 (0.8989 / 0.8627) Evaluation 137: Premsel loss 0.2656, acc 0.8818 (0.9120 / 0.8516) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_27.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_08.pkl.gz Training 138: 0 / 113991: Premsel loss 0.2714, acc 0.8789 (0.9005 / 0.8573) Training 138: 7000 / 113991: Premsel loss 0.2656, acc 0.8815 (0.9073 / 0.8558) Training 138: 14000 / 113991: Premsel loss 0.2603, acc 0.8851 (0.9086 / 0.8616) Training 138: 21000 / 113991: Premsel loss 0.2514, acc 0.8899 (0.9151 / 0.8647) Training 138: 28000 / 113991: Premsel loss 0.2576, acc 0.8861 (0.9123 / 0.8599) Training 138: 35000 / 113991: Premsel loss 0.2660, acc 0.8816 (0.9088 / 0.8545) Training 138: 42000 / 113991: Premsel loss 0.2642, acc 0.8832 (0.9083 / 0.8581) Training 138: 49000 / 113991: Premsel loss 0.2568, acc 0.8865 (0.9119 / 0.8612) Training 138: 56000 / 113991: Premsel loss 0.2586, acc 0.8854 (0.9141 / 0.8567) Training 138: 63000 / 113991: Premsel loss 0.2552, acc 0.8875 (0.9099 / 0.8652) Training 138: 70000 / 113991: Premsel loss 0.2576, acc 0.8863 (0.9106 / 0.8620) Training 138: 77000 / 113991: Premsel loss 0.2545, acc 0.8880 (0.9135 / 0.8624) Training 138: 84000 / 113991: Premsel loss 0.2566, acc 0.8871 (0.9122 / 0.8620) Training 138: 91000 / 113991: Premsel loss 0.2610, acc 0.8848 (0.9110 / 0.8586) Training 138: 98000 / 113991: Premsel loss 0.2616, acc 0.8839 (0.9069 / 0.8608) Training 138: 105000 / 113991: Premsel loss 0.2590, acc 0.8858 (0.9109 / 0.8608) Training 138: 112000 / 113991: Premsel loss 0.2551, acc 0.8872 (0.9155 / 0.8589) Evaluation 138: Premsel loss 0.2590, acc 0.8859 (0.9072 / 0.8645) Training 139: 0 / 113991: Premsel loss 0.2586, acc 0.8855 (0.9124 / 0.8586) Training 139: 7000 / 113991: Premsel loss 0.2560, acc 0.8868 (0.9142 / 0.8594) Training 139: 14000 / 113991: Premsel loss 0.2572, acc 0.8873 (0.9156 / 0.8589) Training 139: 21000 / 113991: Premsel loss 0.2527, acc 0.8884 (0.9168 / 0.8601) Training 139: 28000 / 113991: Premsel loss 0.2645, acc 0.8826 (0.9115 / 0.8537) Training 139: 35000 / 113991: Premsel loss 0.2584, acc 0.8859 (0.9100 / 0.8618) Training 139: 42000 / 113991: Premsel loss 0.2517, acc 0.8887 (0.9166 / 0.8607) Training 139: 49000 / 113991: Premsel loss 0.2494, acc 0.8912 (0.9195 / 0.8628) Training 139: 56000 / 113991: Premsel loss 0.2552, acc 0.8872 (0.9135 / 0.8608) Training 139: 63000 / 113991: Premsel loss 0.2509, acc 0.8903 (0.9169 / 0.8637) Training 139: 70000 / 113991: Premsel loss 0.2558, acc 0.8872 (0.9126 / 0.8617) Training 139: 77000 / 113991: Premsel loss 0.2674, acc 0.8811 (0.9062 / 0.8560) Training 139: 84000 / 113991: Premsel loss 0.2859, acc 0.8710 (0.8969 / 0.8451) Training 139: 91000 / 113991: Premsel loss 0.2770, acc 0.8766 (0.9015 / 0.8517) Training 139: 98000 / 113991: Premsel loss 0.2577, acc 0.8866 (0.9162 / 0.8570) Training 139: 105000 / 113991: Premsel loss 0.2661, acc 0.8820 (0.9127 / 0.8514) Training 139: 112000 / 113991: Premsel loss 0.2547, acc 0.8869 (0.9120 / 0.8618) Evaluation 139: Premsel loss 0.2522, acc 0.8892 (0.9163 / 0.8621) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_24.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-coop___out1.pkl.gz Training 140: 0 / 127841: Premsel loss 0.2526, acc 0.8889 (0.9185 / 0.8594) Training 140: 7000 / 127841: Premsel loss 0.2545, acc 0.8880 (0.9164 / 0.8596) Training 140: 14000 / 127841: Premsel loss 0.2544, acc 0.8879 (0.9144 / 0.8614) Training 140: 21000 / 127841: Premsel loss 0.2597, acc 0.8858 (0.9085 / 0.8631) Training 140: 28000 / 127841: Premsel loss 0.2540, acc 0.8876 (0.9093 / 0.8659) Training 140: 35000 / 127841: Premsel loss 0.2678, acc 0.8815 (0.9073 / 0.8556) Training 140: 42000 / 127841: Premsel loss 0.2677, acc 0.8803 (0.9076 / 0.8529) Training 140: 49000 / 127841: Premsel loss 0.2615, acc 0.8843 (0.9121 / 0.8564) Training 140: 56000 / 127841: Premsel loss 0.2577, acc 0.8864 (0.9142 / 0.8587) Training 140: 63000 / 127841: Premsel loss 0.2635, acc 0.8829 (0.9086 / 0.8573) Training 140: 70000 / 127841: Premsel loss 0.2643, acc 0.8822 (0.9023 / 0.8622) Training 140: 77000 / 127841: Premsel loss 0.2642, acc 0.8831 (0.9111 / 0.8550) Training 140: 84000 / 127841: Premsel loss 0.2569, acc 0.8864 (0.9121 / 0.8606) Training 140: 91000 / 127841: Premsel loss 0.2849, acc 0.8710 (0.8997 / 0.8424) Training 140: 98000 / 127841: Premsel loss 0.2742, acc 0.8769 (0.8999 / 0.8539) Training 140: 105000 / 127841: Premsel loss 0.2728, acc 0.8778 (0.8985 / 0.8570) Training 140: 112000 / 127841: Premsel loss 0.2702, acc 0.8787 (0.9011 / 0.8562) Training 140: 119000 / 127841: Premsel loss 0.2652, acc 0.8818 (0.9078 / 0.8558) Training 140: 126000 / 127841: Premsel loss 0.2625, acc 0.8837 (0.9099 / 0.8575) Evaluation 140: Premsel loss 0.2601, acc 0.8856 (0.9239 / 0.8472) Training 141: 0 / 127841: Premsel loss 0.2628, acc 0.8843 (0.9073 / 0.8612) Training 141: 7000 / 127841: Premsel loss 0.2581, acc 0.8855 (0.9101 / 0.8610) Training 141: 14000 / 127841: Premsel loss 0.2596, acc 0.8850 (0.9086 / 0.8615) Training 141: 21000 / 127841: Premsel loss 0.2627, acc 0.8824 (0.9062 / 0.8586) Training 141: 28000 / 127841: Premsel loss 0.2671, acc 0.8810 (0.9117 / 0.8504) Training 141: 35000 / 127841: Premsel loss 0.2641, acc 0.8828 (0.9113 / 0.8544) Training 141: 42000 / 127841: Premsel loss 0.2661, acc 0.8807 (0.9032 / 0.8581) Training 141: 49000 / 127841: Premsel loss 0.2669, acc 0.8814 (0.9100 / 0.8529) Training 141: 56000 / 127841: Premsel loss 0.2655, acc 0.8815 (0.9061 / 0.8568) Training 141: 63000 / 127841: Premsel loss 0.2648, acc 0.8817 (0.9073 / 0.8560) Training 141: 70000 / 127841: Premsel loss 0.2579, acc 0.8853 (0.9121 / 0.8586) Training 141: 77000 / 127841: Premsel loss 0.2700, acc 0.8800 (0.9095 / 0.8505) Training 141: 84000 / 127841: Premsel loss 0.2688, acc 0.8804 (0.9029 / 0.8578) Training 141: 91000 / 127841: Premsel loss 0.2708, acc 0.8781 (0.9132 / 0.8430) Training 141: 98000 / 127841: Premsel loss 0.2685, acc 0.8799 (0.9006 / 0.8592) Training 141: 105000 / 127841: Premsel loss 0.2662, acc 0.8819 (0.9101 / 0.8536) Training 141: 112000 / 127841: Premsel loss 0.2610, acc 0.8832 (0.9086 / 0.8578) Training 141: 119000 / 127841: Premsel loss 0.2677, acc 0.8810 (0.9023 / 0.8596) Training 141: 126000 / 127841: Premsel loss 0.2655, acc 0.8814 (0.9077 / 0.8551) Evaluation 141: Premsel loss 0.2602, acc 0.8852 (0.9070 / 0.8633) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_06.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l8000-e0.15+coop-mzr02___out1.pkl.gz Training 142: 0 / 129584: Premsel loss 0.2620, acc 0.8833 (0.9062 / 0.8604) Training 142: 7000 / 129584: Premsel loss 0.2735, acc 0.8780 (0.9040 / 0.8521) Training 142: 14000 / 129584: Premsel loss 0.2721, acc 0.8795 (0.9028 / 0.8561) Training 142: 21000 / 129584: Premsel loss 0.2853, acc 0.8713 (0.8996 / 0.8430) Training 142: 28000 / 129584: Premsel loss 0.2849, acc 0.8720 (0.8955 / 0.8486) Training 142: 35000 / 129584: Premsel loss 0.2914, acc 0.8694 (0.8926 / 0.8462) Training 142: 42000 / 129584: Premsel loss 0.2957, acc 0.8654 (0.8901 / 0.8407) Training 142: 49000 / 129584: Premsel loss 0.2794, acc 0.8745 (0.8996 / 0.8495) Training 142: 56000 / 129584: Premsel loss 0.2823, acc 0.8738 (0.9044 / 0.8432) Training 142: 63000 / 129584: Premsel loss 0.2832, acc 0.8737 (0.8972 / 0.8502) Training 142: 70000 / 129584: Premsel loss 0.2796, acc 0.8746 (0.9029 / 0.8463) Training 142: 77000 / 129584: Premsel loss 0.2732, acc 0.8789 (0.9025 / 0.8553) Training 142: 84000 / 129584: Premsel loss 0.2689, acc 0.8803 (0.9042 / 0.8565) Training 142: 91000 / 129584: Premsel loss 0.2768, acc 0.8767 (0.9020 / 0.8513) Training 142: 98000 / 129584: Premsel loss 0.2794, acc 0.8744 (0.9060 / 0.8429) Training 142: 105000 / 129584: Premsel loss 0.2673, acc 0.8819 (0.9044 / 0.8594) Training 142: 112000 / 129584: Premsel loss 0.2780, acc 0.8762 (0.8979 / 0.8544) Training 142: 119000 / 129584: Premsel loss 0.2800, acc 0.8743 (0.8982 / 0.8504) Training 142: 126000 / 129584: Premsel loss 0.2853, acc 0.8713 (0.8993 / 0.8433) Evaluation 142: Premsel loss 0.2844, acc 0.8723 (0.8661 / 0.8786) Training 143: 0 / 129584: Premsel loss 0.2796, acc 0.8744 (0.8969 / 0.8518) Training 143: 7000 / 129584: Premsel loss 0.2789, acc 0.8757 (0.9037 / 0.8476) Training 143: 14000 / 129584: Premsel loss 0.2765, acc 0.8768 (0.9013 / 0.8523) Training 143: 21000 / 129584: Premsel loss 0.2706, acc 0.8799 (0.9010 / 0.8588) Training 143: 28000 / 129584: Premsel loss 0.2698, acc 0.8789 (0.8978 / 0.8599) Training 143: 35000 / 129584: Premsel loss 0.2811, acc 0.8746 (0.9031 / 0.8460) Training 143: 42000 / 129584: Premsel loss 0.2789, acc 0.8756 (0.9011 / 0.8501) Training 143: 49000 / 129584: Premsel loss 0.2708, acc 0.8797 (0.9024 / 0.8571) Training 143: 56000 / 129584: Premsel loss 0.2690, acc 0.8815 (0.9066 / 0.8564) Training 143: 63000 / 129584: Premsel loss 0.2753, acc 0.8766 (0.8990 / 0.8543) Training 143: 70000 / 129584: Premsel loss 0.2735, acc 0.8785 (0.9046 / 0.8523) Training 143: 77000 / 129584: Premsel loss 0.2924, acc 0.8687 (0.8985 / 0.8389) Training 143: 84000 / 129584: Premsel loss 0.2902, acc 0.8690 (0.8968 / 0.8413) Training 143: 91000 / 129584: Premsel loss 0.2764, acc 0.8757 (0.8957 / 0.8557) Training 143: 98000 / 129584: Premsel loss 0.2796, acc 0.8747 (0.9043 / 0.8451) Training 143: 105000 / 129584: Premsel loss 0.2878, acc 0.8712 (0.8901 / 0.8524) Training 143: 112000 / 129584: Premsel loss 0.2782, acc 0.8758 (0.9002 / 0.8514) Training 143: 119000 / 129584: Premsel loss 0.2737, acc 0.8785 (0.9041 / 0.8529) Training 143: 126000 / 129584: Premsel loss 0.2748, acc 0.8780 (0.8967 / 0.8594) Evaluation 143: Premsel loss 0.2817, acc 0.8742 (0.9156 / 0.8328) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_26.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+coop___out1.pkl.gz Training 144: 0 / 123082: Premsel loss 0.2768, acc 0.8770 (0.9002 / 0.8538) Training 144: 7000 / 123082: Premsel loss 0.2668, acc 0.8817 (0.9108 / 0.8526) Training 144: 14000 / 123082: Premsel loss 0.2663, acc 0.8828 (0.9081 / 0.8576) Training 144: 21000 / 123082: Premsel loss 0.2627, acc 0.8835 (0.9118 / 0.8552) Training 144: 28000 / 123082: Premsel loss 0.2677, acc 0.8807 (0.9113 / 0.8501) Training 144: 35000 / 123082: Premsel loss 0.2610, acc 0.8851 (0.9161 / 0.8541) Training 144: 42000 / 123082: Premsel loss 0.2667, acc 0.8820 (0.9097 / 0.8543) Training 144: 49000 / 123082: Premsel loss 0.2755, acc 0.8769 (0.9082 / 0.8456) Training 144: 56000 / 123082: Premsel loss 0.2645, acc 0.8834 (0.9076 / 0.8591) Training 144: 63000 / 123082: Premsel loss 0.2640, acc 0.8822 (0.9076 / 0.8567) Training 144: 70000 / 123082: Premsel loss 0.2781, acc 0.8750 (0.8956 / 0.8545) Training 144: 77000 / 123082: Premsel loss 0.2702, acc 0.8807 (0.9134 / 0.8480) Training 144: 84000 / 123082: Premsel loss 0.2658, acc 0.8826 (0.9077 / 0.8576) Training 144: 91000 / 123082: Premsel loss 0.2691, acc 0.8805 (0.9041 / 0.8569) Training 144: 98000 / 123082: Premsel loss 0.2668, acc 0.8822 (0.9056 / 0.8588) Training 144: 105000 / 123082: Premsel loss 0.2644, acc 0.8822 (0.9098 / 0.8547) Training 144: 112000 / 123082: Premsel loss 0.2726, acc 0.8784 (0.9047 / 0.8521) Training 144: 119000 / 123082: Premsel loss 0.2715, acc 0.8792 (0.9049 / 0.8534) Evaluation 144: Premsel loss 0.2655, acc 0.8826 (0.8954 / 0.8698) Training 145: 0 / 123082: Premsel loss 0.2706, acc 0.8795 (0.9081 / 0.8508) Training 145: 7000 / 123082: Premsel loss 0.2626, acc 0.8833 (0.9149 / 0.8517) Training 145: 14000 / 123082: Premsel loss 0.2667, acc 0.8811 (0.9091 / 0.8531) Training 145: 21000 / 123082: Premsel loss 0.2636, acc 0.8835 (0.9098 / 0.8573) Training 145: 28000 / 123082: Premsel loss 0.2615, acc 0.8853 (0.9092 / 0.8613) Training 145: 35000 / 123082: Premsel loss 0.2570, acc 0.8860 (0.9122 / 0.8598) Training 145: 42000 / 123082: Premsel loss 0.2711, acc 0.8798 (0.9078 / 0.8519) Training 145: 49000 / 123082: Premsel loss 0.2674, acc 0.8812 (0.9076 / 0.8547) Training 145: 56000 / 123082: Premsel loss 0.2672, acc 0.8809 (0.9068 / 0.8551) Training 145: 63000 / 123082: Premsel loss 0.2711, acc 0.8781 (0.9023 / 0.8540) Training 145: 70000 / 123082: Premsel loss 0.2680, acc 0.8804 (0.9048 / 0.8560) Training 145: 77000 / 123082: Premsel loss 0.2623, acc 0.8837 (0.9102 / 0.8572) Training 145: 84000 / 123082: Premsel loss 0.2609, acc 0.8859 (0.9100 / 0.8618) Training 145: 91000 / 123082: Premsel loss 0.2689, acc 0.8804 (0.9088 / 0.8520) Training 145: 98000 / 123082: Premsel loss 0.2673, acc 0.8809 (0.9084 / 0.8534) Training 145: 105000 / 123082: Premsel loss 0.2784, acc 0.8757 (0.9018 / 0.8495) Training 145: 112000 / 123082: Premsel loss 0.2659, acc 0.8824 (0.9107 / 0.8542) Training 145: 119000 / 123082: Premsel loss 0.2767, acc 0.8764 (0.9009 / 0.8519) Evaluation 145: Premsel loss 0.2702, acc 0.8797 (0.8928 / 0.8666) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_74avg-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz Training 146: 0 / 97413: Premsel loss 0.2732, acc 0.8787 (0.9045 / 0.8530) Training 146: 7000 / 97413: Premsel loss 0.2415, acc 0.8950 (0.9216 / 0.8684) Training 146: 14000 / 97413: Premsel loss 0.2362, acc 0.8974 (0.9228 / 0.8719) Training 146: 21000 / 97413: Premsel loss 0.2447, acc 0.8915 (0.9132 / 0.8697) Training 146: 28000 / 97413: Premsel loss 0.2433, acc 0.8936 (0.9229 / 0.8644) Training 146: 35000 / 97413: Premsel loss 0.2608, acc 0.8844 (0.9117 / 0.8572) Training 146: 42000 / 97413: Premsel loss 0.2548, acc 0.8875 (0.9099 / 0.8652) Training 146: 49000 / 97413: Premsel loss 0.2503, acc 0.8895 (0.9190 / 0.8600) Training 146: 56000 / 97413: Premsel loss 0.2584, acc 0.8853 (0.9088 / 0.8619) Training 146: 63000 / 97413: Premsel loss 0.2438, acc 0.8941 (0.9209 / 0.8673) Training 146: 70000 / 97413: Premsel loss 0.2475, acc 0.8916 (0.9210 / 0.8622) Training 146: 77000 / 97413: Premsel loss 0.2374, acc 0.8973 (0.9211 / 0.8735) Training 146: 84000 / 97413: Premsel loss 0.2503, acc 0.8906 (0.9137 / 0.8675) Training 146: 91000 / 97413: Premsel loss 0.2368, acc 0.8968 (0.9237 / 0.8699) Evaluation 146: Premsel loss 0.2424, acc 0.8940 (0.9077 / 0.8803) Training 147: 0 / 97413: Premsel loss 0.2433, acc 0.8929 (0.9192 / 0.8665) Training 147: 7000 / 97413: Premsel loss 0.2411, acc 0.8952 (0.9215 / 0.8689) Training 147: 14000 / 97413: Premsel loss 0.2384, acc 0.8960 (0.9232 / 0.8687) Training 147: 21000 / 97413: Premsel loss 0.2347, acc 0.8968 (0.9189 / 0.8746) Training 147: 28000 / 97413: Premsel loss 0.2356, acc 0.8969 (0.9239 / 0.8698) Training 147: 35000 / 97413: Premsel loss 0.2317, acc 0.9005 (0.9323 / 0.8686) Training 147: 42000 / 97413: Premsel loss 0.2328, acc 0.8988 (0.9276 / 0.8700) Training 147: 49000 / 97413: Premsel loss 0.2310, acc 0.8998 (0.9282 / 0.8714) Training 147: 56000 / 97413: Premsel loss 0.2255, acc 0.9032 (0.9266 / 0.8797) Training 147: 63000 / 97413: Premsel loss 0.2394, acc 0.8956 (0.9200 / 0.8713) Training 147: 70000 / 97413: Premsel loss 0.2343, acc 0.8981 (0.9218 / 0.8744) Training 147: 77000 / 97413: Premsel loss 0.2318, acc 0.9004 (0.9278 / 0.8730) Training 147: 84000 / 97413: Premsel loss 0.2308, acc 0.8990 (0.9284 / 0.8695) Training 147: 91000 / 97413: Premsel loss 0.2410, acc 0.8930 (0.9125 / 0.8736) Evaluation 147: Premsel loss 0.2388, acc 0.8959 (0.9113 / 0.8804) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz Training 148: 0 / 78964: Premsel loss 0.2408, acc 0.8941 (0.9208 / 0.8674) Training 148: 7000 / 78964: Premsel loss 0.2729, acc 0.8754 (0.8999 / 0.8509) Training 148: 14000 / 78964: Premsel loss 0.2654, acc 0.8829 (0.9016 / 0.8642) Training 148: 21000 / 78964: Premsel loss 0.2577, acc 0.8862 (0.9158 / 0.8566) Training 148: 28000 / 78964: Premsel loss 0.2600, acc 0.8849 (0.9104 / 0.8594) Training 148: 35000 / 78964: Premsel loss 0.2575, acc 0.8864 (0.9139 / 0.8589) Training 148: 42000 / 78964: Premsel loss 0.2569, acc 0.8865 (0.9170 / 0.8560) Training 148: 49000 / 78964: Premsel loss 0.2573, acc 0.8872 (0.9055 / 0.8688) Training 148: 56000 / 78964: Premsel loss 0.2606, acc 0.8852 (0.9049 / 0.8654) Training 148: 63000 / 78964: Premsel loss 0.2394, acc 0.8964 (0.9185 / 0.8743) Training 148: 70000 / 78964: Premsel loss 0.2510, acc 0.8892 (0.9228 / 0.8556) Training 148: 77000 / 78964: Premsel loss 0.2707, acc 0.8793 (0.9037 / 0.8549) Evaluation 148: Premsel loss 0.2598, acc 0.8846 (0.9087 / 0.8605) Training 149: 0 / 78964: Premsel loss 0.2821, acc 0.8716 (0.9097 / 0.8335) Training 149: 7000 / 78964: Premsel loss 0.2534, acc 0.8886 (0.9137 / 0.8634) Training 149: 14000 / 78964: Premsel loss 0.2495, acc 0.8907 (0.9153 / 0.8662) Training 149: 21000 / 78964: Premsel loss 0.2523, acc 0.8863 (0.9135 / 0.8592) Training 149: 28000 / 78964: Premsel loss 0.2554, acc 0.8879 (0.9151 / 0.8606) Training 149: 35000 / 78964: Premsel loss 0.2463, acc 0.8912 (0.9183 / 0.8642) Training 149: 42000 / 78964: Premsel loss 0.2427, acc 0.8933 (0.9197 / 0.8668) Training 149: 49000 / 78964: Premsel loss 0.2574, acc 0.8851 (0.9147 / 0.8555) Training 149: 56000 / 78964: Premsel loss 0.2636, acc 0.8838 (0.9071 / 0.8606) Training 149: 63000 / 78964: Premsel loss 0.2497, acc 0.8886 (0.9168 / 0.8603) Training 149: 70000 / 78964: Premsel loss 0.2633, acc 0.8829 (0.9064 / 0.8595) Training 149: 77000 / 78964: Premsel loss 0.2637, acc 0.8833 (0.9049 / 0.8616) Evaluation 149: Premsel loss 0.2542, acc 0.8878 (0.9094 / 0.8662) Loading data... Full data reached, reshuffling... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_09.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz Training 150: 0 / 96970: Premsel loss 0.2602, acc 0.8849 (0.9154 / 0.8544) Training 150: 7000 / 96970: Premsel loss 0.2709, acc 0.8788 (0.9017 / 0.8559) Training 150: 14000 / 96970: Premsel loss 0.2673, acc 0.8795 (0.9028 / 0.8561) Training 150: 21000 / 96970: Premsel loss 0.2679, acc 0.8803 (0.9041 / 0.8565) Training 150: 28000 / 96970: Premsel loss 0.2684, acc 0.8797 (0.9073 / 0.8522) Training 150: 35000 / 96970: Premsel loss 0.2697, acc 0.8797 (0.9070 / 0.8523) Training 150: 42000 / 96970: Premsel loss 0.2608, acc 0.8842 (0.9044 / 0.8639) Training 150: 49000 / 96970: Premsel loss 0.2686, acc 0.8793 (0.8997 / 0.8588) Training 150: 56000 / 96970: Premsel loss 0.2644, acc 0.8819 (0.9064 / 0.8574) Training 150: 63000 / 96970: Premsel loss 0.2675, acc 0.8804 (0.9084 / 0.8524) Training 150: 70000 / 96970: Premsel loss 0.2622, acc 0.8826 (0.9109 / 0.8543) Training 150: 77000 / 96970: Premsel loss 0.2569, acc 0.8864 (0.9108 / 0.8619) Training 150: 84000 / 96970: Premsel loss 0.2600, acc 0.8839 (0.9066 / 0.8612) Training 150: 91000 / 96970: Premsel loss 0.2641, acc 0.8821 (0.9092 / 0.8550) Evaluation 150: Premsel loss 0.2589, acc 0.8856 (0.9096 / 0.8616) Training 151: 0 / 96970: Premsel loss 0.2619, acc 0.8842 (0.9053 / 0.8632) Training 151: 7000 / 96970: Premsel loss 0.2613, acc 0.8834 (0.9065 / 0.8602) Training 151: 14000 / 96970: Premsel loss 0.2721, acc 0.8774 (0.9006 / 0.8542) Training 151: 21000 / 96970: Premsel loss 0.2540, acc 0.8882 (0.9094 / 0.8669) Training 151: 28000 / 96970: Premsel loss 0.2618, acc 0.8848 (0.9069 / 0.8627) Training 151: 35000 / 96970: Premsel loss 0.2634, acc 0.8835 (0.9083 / 0.8587) Training 151: 42000 / 96970: Premsel loss 0.2652, acc 0.8829 (0.9053 / 0.8605) Training 151: 49000 / 96970: Premsel loss 0.2638, acc 0.8823 (0.9064 / 0.8581) Training 151: 56000 / 96970: Premsel loss 0.2642, acc 0.8832 (0.9083 / 0.8581) Training 151: 63000 / 96970: Premsel loss 0.2604, acc 0.8843 (0.9106 / 0.8580) Training 151: 70000 / 96970: Premsel loss 0.2655, acc 0.8819 (0.9088 / 0.8551) Training 151: 77000 / 96970: Premsel loss 0.2594, acc 0.8849 (0.9079 / 0.8619) Training 151: 84000 / 96970: Premsel loss 0.2666, acc 0.8820 (0.9071 / 0.8570) Training 151: 91000 / 96970: Premsel loss 0.2601, acc 0.8840 (0.9072 / 0.8609) Evaluation 151: Premsel loss 0.2577, acc 0.8862 (0.9139 / 0.8585) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_18.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 152: 0 / 108933: Premsel loss 0.2658, acc 0.8815 (0.9036 / 0.8593) Training 152: 7000 / 108933: Premsel loss 0.2763, acc 0.8767 (0.8994 / 0.8539) Training 152: 14000 / 108933: Premsel loss 0.2779, acc 0.8758 (0.9080 / 0.8436) Training 152: 21000 / 108933: Premsel loss 0.2990, acc 0.8657 (0.8892 / 0.8422) Training 152: 28000 / 108933: Premsel loss 0.2940, acc 0.8676 (0.8917 / 0.8435) Training 152: 35000 / 108933: Premsel loss 0.2844, acc 0.8723 (0.9012 / 0.8434) Training 152: 42000 / 108933: Premsel loss 0.2801, acc 0.8749 (0.8973 / 0.8525) Training 152: 49000 / 108933: Premsel loss 0.2750, acc 0.8771 (0.9044 / 0.8498) Training 152: 56000 / 108933: Premsel loss 0.2785, acc 0.8755 (0.8987 / 0.8522) Training 152: 63000 / 108933: Premsel loss 0.2796, acc 0.8750 (0.9075 / 0.8425) Training 152: 70000 / 108933: Premsel loss 0.2802, acc 0.8753 (0.9025 / 0.8480) Training 152: 77000 / 108933: Premsel loss 0.2726, acc 0.8789 (0.9043 / 0.8534) Training 152: 84000 / 108933: Premsel loss 0.2731, acc 0.8776 (0.8991 / 0.8562) Training 152: 91000 / 108933: Premsel loss 0.2765, acc 0.8758 (0.9004 / 0.8512) Training 152: 98000 / 108933: Premsel loss 0.2692, acc 0.8792 (0.9126 / 0.8458) Training 152: 105000 / 108933: Premsel loss 0.2809, acc 0.8744 (0.9048 / 0.8441) Evaluation 152: Premsel loss 0.3071, acc 0.8609 (0.8766 / 0.8452) Training 153: 0 / 108933: Premsel loss 0.3079, acc 0.8610 (0.8930 / 0.8289) Training 153: 7000 / 108933: Premsel loss 0.2906, acc 0.8682 (0.8962 / 0.8402) Training 153: 14000 / 108933: Premsel loss 0.2824, acc 0.8741 (0.9014 / 0.8468) Training 153: 21000 / 108933: Premsel loss 0.2822, acc 0.8734 (0.8982 / 0.8486) Training 153: 28000 / 108933: Premsel loss 0.2809, acc 0.8740 (0.8994 / 0.8487) Training 153: 35000 / 108933: Premsel loss 0.2793, acc 0.8755 (0.8996 / 0.8514) Training 153: 42000 / 108933: Premsel loss 0.2797, acc 0.8746 (0.9014 / 0.8477) Training 153: 49000 / 108933: Premsel loss 0.2809, acc 0.8744 (0.8968 / 0.8520) Training 153: 56000 / 108933: Premsel loss 0.2817, acc 0.8745 (0.8966 / 0.8523) Training 153: 63000 / 108933: Premsel loss 0.2690, acc 0.8813 (0.9121 / 0.8506) Training 153: 70000 / 108933: Premsel loss 0.2874, acc 0.8721 (0.8940 / 0.8501) Training 153: 77000 / 108933: Premsel loss 0.2835, acc 0.8742 (0.9007 / 0.8477) Training 153: 84000 / 108933: Premsel loss 0.2850, acc 0.8734 (0.8959 / 0.8508) Training 153: 91000 / 108933: Premsel loss 0.2843, acc 0.8720 (0.9002 / 0.8438) Training 153: 98000 / 108933: Premsel loss 0.2810, acc 0.8752 (0.9022 / 0.8482) Training 153: 105000 / 108933: Premsel loss 0.2883, acc 0.8705 (0.8963 / 0.8446) Evaluation 153: Premsel loss 0.2876, acc 0.8703 (0.8934 / 0.8472) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_08.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l4800-e0.15+coop-mzr02___out1.pkl.gz Training 154: 0 / 128946: Premsel loss 0.2830, acc 0.8753 (0.8997 / 0.8508) Training 154: 7000 / 128946: Premsel loss 0.3091, acc 0.8594 (0.8750 / 0.8437) Training 154: 14000 / 128946: Premsel loss 0.3040, acc 0.8621 (0.8781 / 0.8460) Training 154: 21000 / 128946: Premsel loss 0.3077, acc 0.8591 (0.8729 / 0.8454) Training 154: 28000 / 128946: Premsel loss 0.3069, acc 0.8614 (0.8799 / 0.8429) Training 154: 35000 / 128946: Premsel loss 0.3041, acc 0.8631 (0.8812 / 0.8449) Training 154: 42000 / 128946: Premsel loss 0.2983, acc 0.8646 (0.8898 / 0.8394) Training 154: 49000 / 128946: Premsel loss 0.3038, acc 0.8616 (0.8845 / 0.8386) Training 154: 56000 / 128946: Premsel loss 0.2991, acc 0.8637 (0.8823 / 0.8452) Training 154: 63000 / 128946: Premsel loss 0.2955, acc 0.8665 (0.8910 / 0.8421) Training 154: 70000 / 128946: Premsel loss 0.2978, acc 0.8658 (0.8807 / 0.8509) Training 154: 77000 / 128946: Premsel loss 0.2984, acc 0.8656 (0.8842 / 0.8471) Training 154: 84000 / 128946: Premsel loss 0.2964, acc 0.8663 (0.8853 / 0.8472) Training 154: 91000 / 128946: Premsel loss 0.2969, acc 0.8646 (0.8873 / 0.8419) Training 154: 98000 / 128946: Premsel loss 0.3149, acc 0.8555 (0.8716 / 0.8393) Training 154: 105000 / 128946: Premsel loss 0.3140, acc 0.8568 (0.8793 / 0.8343) Training 154: 112000 / 128946: Premsel loss 0.3158, acc 0.8552 (0.8758 / 0.8347) Training 154: 119000 / 128946: Premsel loss 0.3223, acc 0.8520 (0.8645 / 0.8396) Training 154: 126000 / 128946: Premsel loss 0.3132, acc 0.8558 (0.8715 / 0.8401) Evaluation 154: Premsel loss 0.3162, acc 0.8552 (0.8387 / 0.8718) Training 155: 0 / 128946: Premsel loss 0.3195, acc 0.8530 (0.8742 / 0.8318) Training 155: 7000 / 128946: Premsel loss 0.3089, acc 0.8588 (0.8736 / 0.8441) Training 155: 14000 / 128946: Premsel loss 0.3077, acc 0.8601 (0.8732 / 0.8469) Training 155: 21000 / 128946: Premsel loss 0.3072, acc 0.8604 (0.8737 / 0.8471) Training 155: 28000 / 128946: Premsel loss 0.3153, acc 0.8554 (0.8713 / 0.8395) Training 155: 35000 / 128946: Premsel loss 0.2988, acc 0.8656 (0.8809 / 0.8503) Training 155: 42000 / 128946: Premsel loss 0.2978, acc 0.8652 (0.8796 / 0.8508) Training 155: 49000 / 128946: Premsel loss 0.3000, acc 0.8639 (0.8783 / 0.8495) Training 155: 56000 / 128946: Premsel loss 0.2937, acc 0.8670 (0.8857 / 0.8483) Training 155: 63000 / 128946: Premsel loss 0.3001, acc 0.8643 (0.8760 / 0.8525) Training 155: 70000 / 128946: Premsel loss 0.2946, acc 0.8666 (0.8812 / 0.8520) Training 155: 77000 / 128946: Premsel loss 0.2988, acc 0.8661 (0.8895 / 0.8427) Training 155: 84000 / 128946: Premsel loss 0.2970, acc 0.8653 (0.8851 / 0.8455) Training 155: 91000 / 128946: Premsel loss 0.3035, acc 0.8620 (0.8814 / 0.8426) Training 155: 98000 / 128946: Premsel loss 0.2975, acc 0.8652 (0.8839 / 0.8465) Training 155: 105000 / 128946: Premsel loss 0.2983, acc 0.8645 (0.8847 / 0.8444) Training 155: 112000 / 128946: Premsel loss 0.3022, acc 0.8619 (0.8779 / 0.8459) Training 155: 119000 / 128946: Premsel loss 0.2947, acc 0.8676 (0.8847 / 0.8505) Training 155: 126000 / 128946: Premsel loss 0.2951, acc 0.8668 (0.8813 / 0.8524) Evaluation 155: Premsel loss 0.2966, acc 0.8658 (0.8768 / 0.8547) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_14.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_07.pkl.gz Training 156: 0 / 123220: Premsel loss 0.3220, acc 0.8533 (0.8834 / 0.8232) Training 156: 7000 / 123220: Premsel loss 0.2879, acc 0.8706 (0.8962 / 0.8451) Training 156: 14000 / 123220: Premsel loss 0.2879, acc 0.8706 (0.8983 / 0.8429) Training 156: 21000 / 123220: Premsel loss 0.2853, acc 0.8716 (0.8978 / 0.8453) Training 156: 28000 / 123220: Premsel loss 0.2850, acc 0.8723 (0.9009 / 0.8438) Training 156: 35000 / 123220: Premsel loss 0.2931, acc 0.8673 (0.8927 / 0.8419) Training 156: 42000 / 123220: Premsel loss 0.2904, acc 0.8699 (0.8935 / 0.8463) Training 156: 49000 / 123220: Premsel loss 0.2911, acc 0.8682 (0.8924 / 0.8439) Training 156: 56000 / 123220: Premsel loss 0.2900, acc 0.8691 (0.8935 / 0.8447) Training 156: 63000 / 123220: Premsel loss 0.2833, acc 0.8737 (0.8992 / 0.8481) Training 156: 70000 / 123220: Premsel loss 0.2939, acc 0.8683 (0.8934 / 0.8432) Training 156: 77000 / 123220: Premsel loss 0.2920, acc 0.8688 (0.8915 / 0.8461) Training 156: 84000 / 123220: Premsel loss 0.2901, acc 0.8690 (0.8924 / 0.8457) Training 156: 91000 / 123220: Premsel loss 0.2872, acc 0.8715 (0.8995 / 0.8435) Training 156: 98000 / 123220: Premsel loss 0.2953, acc 0.8665 (0.8942 / 0.8387) Training 156: 105000 / 123220: Premsel loss 0.2927, acc 0.8679 (0.8940 / 0.8417) Training 156: 112000 / 123220: Premsel loss 0.2928, acc 0.8682 (0.8956 / 0.8408) Training 156: 119000 / 123220: Premsel loss 0.2965, acc 0.8654 (0.8950 / 0.8358) Evaluation 156: Premsel loss 0.2969, acc 0.8659 (0.9276 / 0.8042) Training 157: 0 / 123220: Premsel loss 0.2863, acc 0.8715 (0.8963 / 0.8467) Training 157: 7000 / 123220: Premsel loss 0.2845, acc 0.8727 (0.8970 / 0.8484) Training 157: 14000 / 123220: Premsel loss 0.2860, acc 0.8717 (0.8973 / 0.8461) Training 157: 21000 / 123220: Premsel loss 0.2897, acc 0.8698 (0.9003 / 0.8394) Training 157: 28000 / 123220: Premsel loss 0.2919, acc 0.8680 (0.8980 / 0.8381) Training 157: 35000 / 123220: Premsel loss 0.2833, acc 0.8729 (0.9009 / 0.8448) Training 157: 42000 / 123220: Premsel loss 0.2748, acc 0.8776 (0.9018 / 0.8534) Training 157: 49000 / 123220: Premsel loss 0.2925, acc 0.8667 (0.8948 / 0.8387) Training 157: 56000 / 123220: Premsel loss 0.2803, acc 0.8750 (0.9043 / 0.8457) Training 157: 63000 / 123220: Premsel loss 0.2764, acc 0.8773 (0.9005 / 0.8542) Training 157: 70000 / 123220: Premsel loss 0.2846, acc 0.8723 (0.8962 / 0.8485) Training 157: 77000 / 123220: Premsel loss 0.2796, acc 0.8750 (0.9002 / 0.8498) Training 157: 84000 / 123220: Premsel loss 0.2803, acc 0.8751 (0.9030 / 0.8472) Training 157: 91000 / 123220: Premsel loss 0.2830, acc 0.8727 (0.8964 / 0.8491) Training 157: 98000 / 123220: Premsel loss 0.2793, acc 0.8750 (0.9029 / 0.8470) Training 157: 105000 / 123220: Premsel loss 0.2792, acc 0.8749 (0.9002 / 0.8496) Training 157: 112000 / 123220: Premsel loss 0.3119, acc 0.8583 (0.8841 / 0.8325) Training 157: 119000 / 123220: Premsel loss 0.3107, acc 0.8581 (0.8843 / 0.8319) Evaluation 157: Premsel loss 0.3009, acc 0.8635 (0.9064 / 0.8207) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_05.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz Training 158: 0 / 113689: Premsel loss 0.2998, acc 0.8646 (0.8895 / 0.8396) Training 158: 7000 / 113689: Premsel loss 0.2836, acc 0.8723 (0.8970 / 0.8477) Training 158: 14000 / 113689: Premsel loss 0.3013, acc 0.8632 (0.8939 / 0.8324) Training 158: 21000 / 113689: Premsel loss 0.2878, acc 0.8717 (0.8933 / 0.8501) Training 158: 28000 / 113689: Premsel loss 0.2944, acc 0.8675 (0.8937 / 0.8413) Training 158: 35000 / 113689: Premsel loss 0.2875, acc 0.8708 (0.8973 / 0.8443) Training 158: 42000 / 113689: Premsel loss 0.2844, acc 0.8735 (0.9026 / 0.8444) Training 158: 49000 / 113689: Premsel loss 0.2775, acc 0.8756 (0.9035 / 0.8478) Training 158: 56000 / 113689: Premsel loss 0.2768, acc 0.8760 (0.9028 / 0.8491) Training 158: 63000 / 113689: Premsel loss 0.2815, acc 0.8746 (0.8988 / 0.8505) Training 158: 70000 / 113689: Premsel loss 0.2776, acc 0.8764 (0.9026 / 0.8503) Training 158: 77000 / 113689: Premsel loss 0.2867, acc 0.8706 (0.8976 / 0.8437) Training 158: 84000 / 113689: Premsel loss 0.2902, acc 0.8709 (0.8939 / 0.8479) Training 158: 91000 / 113689: Premsel loss 0.2846, acc 0.8716 (0.8926 / 0.8506) Training 158: 98000 / 113689: Premsel loss 0.2795, acc 0.8742 (0.8968 / 0.8515) Training 158: 105000 / 113689: Premsel loss 0.2814, acc 0.8734 (0.8955 / 0.8512) Training 158: 112000 / 113689: Premsel loss 0.2751, acc 0.8764 (0.9019 / 0.8508) Evaluation 158: Premsel loss 0.2746, acc 0.8775 (0.8826 / 0.8723) Training 159: 0 / 113689: Premsel loss 0.2629, acc 0.8828 (0.9079 / 0.8576) Training 159: 7000 / 113689: Premsel loss 0.2708, acc 0.8795 (0.9007 / 0.8583) Training 159: 14000 / 113689: Premsel loss 0.2780, acc 0.8756 (0.9049 / 0.8464) Training 159: 21000 / 113689: Premsel loss 0.2743, acc 0.8772 (0.9027 / 0.8517) Training 159: 28000 / 113689: Premsel loss 0.2755, acc 0.8768 (0.9006 / 0.8530) Training 159: 35000 / 113689: Premsel loss 0.2713, acc 0.8807 (0.9058 / 0.8555) Training 159: 42000 / 113689: Premsel loss 0.2749, acc 0.8772 (0.9055 / 0.8490) Training 159: 49000 / 113689: Premsel loss 0.2688, acc 0.8802 (0.9061 / 0.8542) Training 159: 56000 / 113689: Premsel loss 0.2794, acc 0.8761 (0.9023 / 0.8500) Training 159: 63000 / 113689: Premsel loss 0.2691, acc 0.8799 (0.9024 / 0.8574) Training 159: 70000 / 113689: Premsel loss 0.2747, acc 0.8778 (0.8984 / 0.8572) Training 159: 77000 / 113689: Premsel loss 0.2819, acc 0.8729 (0.8954 / 0.8504) Training 159: 84000 / 113689: Premsel loss 0.2759, acc 0.8762 (0.9048 / 0.8476) Training 159: 91000 / 113689: Premsel loss 0.2812, acc 0.8739 (0.9000 / 0.8478) Training 159: 98000 / 113689: Premsel loss 0.2705, acc 0.8786 (0.8963 / 0.8608) Training 159: 105000 / 113689: Premsel loss 0.2705, acc 0.8799 (0.9061 / 0.8536) Training 159: 112000 / 113689: Premsel loss 0.2726, acc 0.8787 (0.9050 / 0.8524) Evaluation 159: Premsel loss 0.2721, acc 0.8788 (0.8909 / 0.8668) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_16.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz Training 160: 0 / 117434: Premsel loss 0.2662, acc 0.8815 (0.9073 / 0.8556) Training 160: 7000 / 117434: Premsel loss 0.2760, acc 0.8777 (0.9023 / 0.8531) Training 160: 14000 / 117434: Premsel loss 0.2696, acc 0.8809 (0.9041 / 0.8578) Training 160: 21000 / 117434: Premsel loss 0.2738, acc 0.8780 (0.9073 / 0.8488) Training 160: 28000 / 117434: Premsel loss 0.2738, acc 0.8787 (0.9040 / 0.8533) Training 160: 35000 / 117434: Premsel loss 0.2764, acc 0.8764 (0.9046 / 0.8482) Training 160: 42000 / 117434: Premsel loss 0.2687, acc 0.8807 (0.9062 / 0.8551) Training 160: 49000 / 117434: Premsel loss 0.2727, acc 0.8789 (0.9082 / 0.8495) Training 160: 56000 / 117434: Premsel loss 0.2711, acc 0.8795 (0.9060 / 0.8530) Training 160: 63000 / 117434: Premsel loss 0.2626, acc 0.8827 (0.9103 / 0.8552) Training 160: 70000 / 117434: Premsel loss 0.2723, acc 0.8778 (0.9027 / 0.8528) Training 160: 77000 / 117434: Premsel loss 0.2760, acc 0.8766 (0.9005 / 0.8528) Training 160: 84000 / 117434: Premsel loss 0.2720, acc 0.8801 (0.9086 / 0.8515) Training 160: 91000 / 117434: Premsel loss 0.2684, acc 0.8809 (0.9044 / 0.8574) Training 160: 98000 / 117434: Premsel loss 0.2731, acc 0.8781 (0.9026 / 0.8537) Training 160: 105000 / 117434: Premsel loss 0.2676, acc 0.8814 (0.9099 / 0.8529) Training 160: 112000 / 117434: Premsel loss 0.2708, acc 0.8790 (0.8994 / 0.8587) Evaluation 160: Premsel loss 0.2677, acc 0.8812 (0.9234 / 0.8390) Training 161: 0 / 117434: Premsel loss 0.2637, acc 0.8849 (0.9096 / 0.8602) Training 161: 7000 / 117434: Premsel loss 0.2707, acc 0.8794 (0.9095 / 0.8493) Training 161: 14000 / 117434: Premsel loss 0.2725, acc 0.8780 (0.9047 / 0.8514) Training 161: 21000 / 117434: Premsel loss 0.2657, acc 0.8825 (0.9048 / 0.8602) Training 161: 28000 / 117434: Premsel loss 0.2674, acc 0.8819 (0.9104 / 0.8534) Training 161: 35000 / 117434: Premsel loss 0.2625, acc 0.8849 (0.9092 / 0.8605) Training 161: 42000 / 117434: Premsel loss 0.2638, acc 0.8834 (0.9065 / 0.8603) Training 161: 49000 / 117434: Premsel loss 0.2650, acc 0.8817 (0.9085 / 0.8549) Training 161: 56000 / 117434: Premsel loss 0.2646, acc 0.8817 (0.9058 / 0.8576) Training 161: 63000 / 117434: Premsel loss 0.2697, acc 0.8808 (0.9103 / 0.8513) Training 161: 70000 / 117434: Premsel loss 0.2611, acc 0.8845 (0.9083 / 0.8606) Training 161: 77000 / 117434: Premsel loss 0.2665, acc 0.8810 (0.9076 / 0.8543) Training 161: 84000 / 117434: Premsel loss 0.2605, acc 0.8853 (0.9094 / 0.8612) Training 161: 91000 / 117434: Premsel loss 0.2739, acc 0.8787 (0.9018 / 0.8556) Training 161: 98000 / 117434: Premsel loss 0.2682, acc 0.8806 (0.9042 / 0.8571) Training 161: 105000 / 117434: Premsel loss 0.2684, acc 0.8808 (0.9082 / 0.8535) Training 161: 112000 / 117434: Premsel loss 0.2633, acc 0.8834 (0.9086 / 0.8582) Evaluation 161: Premsel loss 0.2645, acc 0.8828 (0.8897 / 0.8759) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_15.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz Training 162: 0 / 128043: Premsel loss 0.2692, acc 0.8805 (0.9060 / 0.8551) Training 162: 7000 / 128043: Premsel loss 0.2907, acc 0.8698 (0.8886 / 0.8510) Training 162: 14000 / 128043: Premsel loss 0.2886, acc 0.8699 (0.8925 / 0.8472) Training 162: 21000 / 128043: Premsel loss 0.2893, acc 0.8696 (0.8849 / 0.8543) Training 162: 28000 / 128043: Premsel loss 0.2847, acc 0.8730 (0.8863 / 0.8596) Training 162: 35000 / 128043: Premsel loss 0.2863, acc 0.8713 (0.8902 / 0.8525) Training 162: 42000 / 128043: Premsel loss 0.2824, acc 0.8728 (0.8870 / 0.8587) Training 162: 49000 / 128043: Premsel loss 0.2781, acc 0.8750 (0.8875 / 0.8626) Training 162: 56000 / 128043: Premsel loss 0.2791, acc 0.8734 (0.8863 / 0.8605) Training 162: 63000 / 128043: Premsel loss 0.2822, acc 0.8735 (0.8834 / 0.8636) Training 162: 70000 / 128043: Premsel loss 0.2837, acc 0.8720 (0.8833 / 0.8607) Training 162: 77000 / 128043: Premsel loss 0.2829, acc 0.8722 (0.8873 / 0.8571) Training 162: 84000 / 128043: Premsel loss 0.2928, acc 0.8667 (0.8767 / 0.8568) Training 162: 91000 / 128043: Premsel loss 0.2825, acc 0.8740 (0.8897 / 0.8583) Training 162: 98000 / 128043: Premsel loss 0.2854, acc 0.8711 (0.8878 / 0.8544) Training 162: 105000 / 128043: Premsel loss 0.2839, acc 0.8723 (0.8912 / 0.8535) Training 162: 112000 / 128043: Premsel loss 0.2833, acc 0.8734 (0.8966 / 0.8502) Training 162: 119000 / 128043: Premsel loss 0.2860, acc 0.8707 (0.8910 / 0.8504) Training 162: 126000 / 128043: Premsel loss 0.2841, acc 0.8722 (0.8869 / 0.8575) Evaluation 162: Premsel loss 0.2827, acc 0.8730 (0.8961 / 0.8499) Training 163: 0 / 128043: Premsel loss 0.2937, acc 0.8681 (0.8857 / 0.8505) Training 163: 7000 / 128043: Premsel loss 0.2768, acc 0.8758 (0.8952 / 0.8564) Training 163: 14000 / 128043: Premsel loss 0.2817, acc 0.8741 (0.8904 / 0.8579) Training 163: 21000 / 128043: Premsel loss 0.2800, acc 0.8739 (0.8906 / 0.8572) Training 163: 28000 / 128043: Premsel loss 0.2802, acc 0.8734 (0.8853 / 0.8615) Training 163: 35000 / 128043: Premsel loss 0.2808, acc 0.8737 (0.8922 / 0.8552) Training 163: 42000 / 128043: Premsel loss 0.2817, acc 0.8739 (0.8908 / 0.8571) Training 163: 49000 / 128043: Premsel loss 0.2814, acc 0.8739 (0.8916 / 0.8563) Training 163: 56000 / 128043: Premsel loss 0.2845, acc 0.8720 (0.8906 / 0.8534) Training 163: 63000 / 128043: Premsel loss 0.2842, acc 0.8715 (0.8907 / 0.8523) Training 163: 70000 / 128043: Premsel loss 0.2785, acc 0.8750 (0.8907 / 0.8593) Training 163: 77000 / 128043: Premsel loss 0.2801, acc 0.8738 (0.8870 / 0.8606) Training 163: 84000 / 128043: Premsel loss 0.2789, acc 0.8740 (0.8898 / 0.8583) Training 163: 91000 / 128043: Premsel loss 0.2757, acc 0.8767 (0.8887 / 0.8648) Training 163: 98000 / 128043: Premsel loss 0.2781, acc 0.8752 (0.8916 / 0.8587) Training 163: 105000 / 128043: Premsel loss 0.2737, acc 0.8781 (0.8924 / 0.8638) Training 163: 112000 / 128043: Premsel loss 0.2854, acc 0.8700 (0.8838 / 0.8561) Training 163: 119000 / 128043: Premsel loss 0.2832, acc 0.8723 (0.8832 / 0.8615) Training 163: 126000 / 128043: Premsel loss 0.2761, acc 0.8761 (0.8902 / 0.8620) Evaluation 163: Premsel loss 0.2827, acc 0.8729 (0.8662 / 0.8795) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_04.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_06.pkl.gz Training 164: 0 / 99007: Premsel loss 0.2834, acc 0.8716 (0.8851 / 0.8581) Training 164: 7000 / 99007: Premsel loss 0.3042, acc 0.8617 (0.8827 / 0.8408) Training 164: 14000 / 99007: Premsel loss 0.3086, acc 0.8586 (0.8807 / 0.8365) Training 164: 21000 / 99007: Premsel loss 0.2997, acc 0.8640 (0.8888 / 0.8392) Training 164: 28000 / 99007: Premsel loss 0.3015, acc 0.8630 (0.8843 / 0.8418) Training 164: 35000 / 99007: Premsel loss 0.2952, acc 0.8665 (0.8864 / 0.8466) Training 164: 42000 / 99007: Premsel loss 0.2958, acc 0.8659 (0.8889 / 0.8430) Training 164: 49000 / 99007: Premsel loss 0.2988, acc 0.8641 (0.8886 / 0.8396) Training 164: 56000 / 99007: Premsel loss 0.2991, acc 0.8644 (0.8903 / 0.8386) Training 164: 63000 / 99007: Premsel loss 0.2977, acc 0.8650 (0.8897 / 0.8403) Training 164: 70000 / 99007: Premsel loss 0.2956, acc 0.8665 (0.8894 / 0.8436) Training 164: 77000 / 99007: Premsel loss 0.3062, acc 0.8601 (0.8822 / 0.8381) Training 164: 84000 / 99007: Premsel loss 0.2955, acc 0.8666 (0.8867 / 0.8465) Training 164: 91000 / 99007: Premsel loss 0.3088, acc 0.8586 (0.8833 / 0.8340) Training 164: 98000 / 99007: Premsel loss 0.3089, acc 0.8584 (0.8844 / 0.8324) Evaluation 164: Premsel loss 0.3076, acc 0.8594 (0.8793 / 0.8394) Training 165: 0 / 99007: Premsel loss 0.3058, acc 0.8603 (0.8789 / 0.8417) Training 165: 7000 / 99007: Premsel loss 0.3054, acc 0.8614 (0.8835 / 0.8393) Training 165: 14000 / 99007: Premsel loss 0.3015, acc 0.8635 (0.8908 / 0.8362) Training 165: 21000 / 99007: Premsel loss 0.3062, acc 0.8603 (0.8785 / 0.8422) Training 165: 28000 / 99007: Premsel loss 0.3089, acc 0.8600 (0.8816 / 0.8384) Training 165: 35000 / 99007: Premsel loss 0.3008, acc 0.8640 (0.8849 / 0.8431) Training 165: 42000 / 99007: Premsel loss 0.3048, acc 0.8618 (0.8803 / 0.8433) Training 165: 49000 / 99007: Premsel loss 0.2977, acc 0.8651 (0.8901 / 0.8402) Training 165: 56000 / 99007: Premsel loss 0.3003, acc 0.8642 (0.8891 / 0.8392) Training 165: 63000 / 99007: Premsel loss 0.2938, acc 0.8669 (0.8896 / 0.8441) Training 165: 70000 / 99007: Premsel loss 0.3017, acc 0.8630 (0.8903 / 0.8358) Training 165: 77000 / 99007: Premsel loss 0.3052, acc 0.8616 (0.8787 / 0.8445) Training 165: 84000 / 99007: Premsel loss 0.2995, acc 0.8646 (0.8894 / 0.8398) Training 165: 91000 / 99007: Premsel loss 0.3026, acc 0.8623 (0.8844 / 0.8401) Training 165: 98000 / 99007: Premsel loss 0.3066, acc 0.8611 (0.8888 / 0.8334) Evaluation 165: Premsel loss 0.2990, acc 0.8645 (0.8739 / 0.8550) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_01.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_88-query512-ctx768-w0-coop___out1.pkl.gz Training 166: 0 / 125452: Premsel loss 0.2960, acc 0.8677 (0.8865 / 0.8490) Training 166: 7000 / 125452: Premsel loss 0.2825, acc 0.8731 (0.8931 / 0.8532) Training 166: 14000 / 125452: Premsel loss 0.2843, acc 0.8730 (0.8931 / 0.8528) Training 166: 21000 / 125452: Premsel loss 0.2877, acc 0.8710 (0.8966 / 0.8455) Training 166: 28000 / 125452: Premsel loss 0.2953, acc 0.8670 (0.8837 / 0.8502) Training 166: 35000 / 125452: Premsel loss 0.2944, acc 0.8679 (0.8836 / 0.8521) Training 166: 42000 / 125452: Premsel loss 0.3103, acc 0.8591 (0.8717 / 0.8466) Training 166: 49000 / 125452: Premsel loss 0.3160, acc 0.8556 (0.8714 / 0.8397) Training 166: 56000 / 125452: Premsel loss 0.3004, acc 0.8637 (0.8767 / 0.8508) Training 166: 63000 / 125452: Premsel loss 0.2912, acc 0.8681 (0.8882 / 0.8480) Training 166: 70000 / 125452: Premsel loss 0.2934, acc 0.8684 (0.8879 / 0.8489) Training 166: 77000 / 125452: Premsel loss 0.3006, acc 0.8651 (0.8850 / 0.8453) Training 166: 84000 / 125452: Premsel loss 0.3042, acc 0.8612 (0.8827 / 0.8397) Training 166: 91000 / 125452: Premsel loss 0.3011, acc 0.8629 (0.8808 / 0.8451) Training 166: 98000 / 125452: Premsel loss 0.3103, acc 0.8585 (0.8737 / 0.8433) Training 166: 105000 / 125452: Premsel loss 0.2933, acc 0.8687 (0.8870 / 0.8505) Training 166: 112000 / 125452: Premsel loss 0.2864, acc 0.8714 (0.8855 / 0.8574) Training 166: 119000 / 125452: Premsel loss 0.3089, acc 0.8599 (0.8758 / 0.8440) Evaluation 166: Premsel loss 0.2950, acc 0.8668 (0.9000 / 0.8336) Training 167: 0 / 125452: Premsel loss 0.2999, acc 0.8651 (0.8791 / 0.8512) Training 167: 7000 / 125452: Premsel loss 0.2912, acc 0.8683 (0.8834 / 0.8532) Training 167: 14000 / 125452: Premsel loss 0.2832, acc 0.8732 (0.8895 / 0.8569) Training 167: 21000 / 125452: Premsel loss 0.3045, acc 0.8623 (0.8861 / 0.8386) Training 167: 28000 / 125452: Premsel loss 0.2974, acc 0.8647 (0.8859 / 0.8434) Training 167: 35000 / 125452: Premsel loss 0.2992, acc 0.8640 (0.8899 / 0.8381) Training 167: 42000 / 125452: Premsel loss 0.2932, acc 0.8683 (0.8839 / 0.8526) Training 167: 49000 / 125452: Premsel loss 0.2935, acc 0.8674 (0.8850 / 0.8497) Training 167: 56000 / 125452: Premsel loss 0.3061, acc 0.8608 (0.8825 / 0.8391) Training 167: 63000 / 125452: Premsel loss 0.3028, acc 0.8638 (0.8825 / 0.8451) Training 167: 70000 / 125452: Premsel loss 0.3031, acc 0.8636 (0.8841 / 0.8430) Training 167: 77000 / 125452: Premsel loss 0.2981, acc 0.8644 (0.8836 / 0.8451) Training 167: 84000 / 125452: Premsel loss 0.2988, acc 0.8659 (0.8839 / 0.8480) Training 167: 91000 / 125452: Premsel loss 0.2973, acc 0.8650 (0.8779 / 0.8520) Training 167: 98000 / 125452: Premsel loss 0.2990, acc 0.8656 (0.8880 / 0.8432) Training 167: 105000 / 125452: Premsel loss 0.2902, acc 0.8697 (0.8908 / 0.8485) Training 167: 112000 / 125452: Premsel loss 0.3021, acc 0.8631 (0.8822 / 0.8439) Training 167: 119000 / 125452: Premsel loss 0.2948, acc 0.8677 (0.8838 / 0.8516) Evaluation 167: Premsel loss 0.2988, acc 0.8648 (0.9183 / 0.8112) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_19.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz Training 168: 0 / 123415: Premsel loss 0.2950, acc 0.8659 (0.8792 / 0.8527) Training 168: 7000 / 123415: Premsel loss 0.2871, acc 0.8711 (0.8922 / 0.8501) Training 168: 14000 / 123415: Premsel loss 0.2911, acc 0.8680 (0.8941 / 0.8419) Training 168: 21000 / 123415: Premsel loss 0.2828, acc 0.8729 (0.8929 / 0.8528) Training 168: 28000 / 123415: Premsel loss 0.2839, acc 0.8715 (0.8910 / 0.8520) Training 168: 35000 / 123415: Premsel loss 0.2979, acc 0.8644 (0.8827 / 0.8460) Training 168: 42000 / 123415: Premsel loss 0.2826, acc 0.8730 (0.9004 / 0.8455) Training 168: 49000 / 123415: Premsel loss 0.2887, acc 0.8704 (0.8995 / 0.8413) Training 168: 56000 / 123415: Premsel loss 0.3159, acc 0.8555 (0.8812 / 0.8297) Training 168: 63000 / 123415: Premsel loss 0.2938, acc 0.8676 (0.8911 / 0.8440) Training 168: 70000 / 123415: Premsel loss 0.2850, acc 0.8710 (0.8937 / 0.8482) Training 168: 77000 / 123415: Premsel loss 0.2876, acc 0.8707 (0.8922 / 0.8492) Training 168: 84000 / 123415: Premsel loss 0.2925, acc 0.8677 (0.8868 / 0.8487) Training 168: 91000 / 123415: Premsel loss 0.2792, acc 0.8736 (0.8981 / 0.8491) Training 168: 98000 / 123415: Premsel loss 0.2838, acc 0.8728 (0.8961 / 0.8494) Training 168: 105000 / 123415: Premsel loss 0.2865, acc 0.8709 (0.8898 / 0.8521) Training 168: 112000 / 123415: Premsel loss 0.2779, acc 0.8758 (0.9012 / 0.8504) Training 168: 119000 / 123415: Premsel loss 0.2832, acc 0.8738 (0.8961 / 0.8516) Evaluation 168: Premsel loss 0.2834, acc 0.8728 (0.9194 / 0.8262) Training 169: 0 / 123415: Premsel loss 0.2773, acc 0.8746 (0.8992 / 0.8500) Training 169: 7000 / 123415: Premsel loss 0.2773, acc 0.8754 (0.8963 / 0.8545) Training 169: 14000 / 123415: Premsel loss 0.2865, acc 0.8700 (0.8915 / 0.8485) Training 169: 21000 / 123415: Premsel loss 0.2866, acc 0.8712 (0.9004 / 0.8421) Training 169: 28000 / 123415: Premsel loss 0.2904, acc 0.8700 (0.8947 / 0.8454) Training 169: 35000 / 123415: Premsel loss 0.2811, acc 0.8741 (0.9074 / 0.8408) Training 169: 42000 / 123415: Premsel loss 0.2784, acc 0.8759 (0.9050 / 0.8469) Training 169: 49000 / 123415: Premsel loss 0.2791, acc 0.8754 (0.9020 / 0.8488) Training 169: 56000 / 123415: Premsel loss 0.2835, acc 0.8729 (0.8988 / 0.8471) Training 169: 63000 / 123415: Premsel loss 0.2846, acc 0.8721 (0.8973 / 0.8468) Training 169: 70000 / 123415: Premsel loss 0.2888, acc 0.8692 (0.8973 / 0.8412) Training 169: 77000 / 123415: Premsel loss 0.2783, acc 0.8771 (0.9004 / 0.8537) Training 169: 84000 / 123415: Premsel loss 0.2820, acc 0.8736 (0.8974 / 0.8499) Training 169: 91000 / 123415: Premsel loss 0.2819, acc 0.8738 (0.9001 / 0.8476) Training 169: 98000 / 123415: Premsel loss 0.2801, acc 0.8749 (0.9006 / 0.8492) Training 169: 105000 / 123415: Premsel loss 0.2803, acc 0.8733 (0.8955 / 0.8512) Training 169: 112000 / 123415: Premsel loss 0.2836, acc 0.8721 (0.8945 / 0.8498) Training 169: 119000 / 123415: Premsel loss 0.2896, acc 0.8692 (0.8950 / 0.8435) Evaluation 169: Premsel loss 0.2791, acc 0.8750 (0.9068 / 0.8431) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_13.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz Training 170: 0 / 124344: Premsel loss 0.2876, acc 0.8701 (0.9046 / 0.8356) Training 170: 7000 / 124344: Premsel loss 0.2699, acc 0.8802 (0.9106 / 0.8497) Training 170: 14000 / 124344: Premsel loss 0.2718, acc 0.8784 (0.9053 / 0.8516) Training 170: 21000 / 124344: Premsel loss 0.2700, acc 0.8802 (0.9071 / 0.8533) Training 170: 28000 / 124344: Premsel loss 0.2739, acc 0.8777 (0.9107 / 0.8446) Training 170: 35000 / 124344: Premsel loss 0.2691, acc 0.8804 (0.9116 / 0.8493) Training 170: 42000 / 124344: Premsel loss 0.2654, acc 0.8826 (0.9105 / 0.8548) Training 170: 49000 / 124344: Premsel loss 0.2654, acc 0.8827 (0.9099 / 0.8555) Training 170: 56000 / 124344: Premsel loss 0.2727, acc 0.8785 (0.9066 / 0.8504) Training 170: 63000 / 124344: Premsel loss 0.2747, acc 0.8766 (0.9059 / 0.8472) Training 170: 70000 / 124344: Premsel loss 0.2801, acc 0.8742 (0.8979 / 0.8505) Training 170: 77000 / 124344: Premsel loss 0.2757, acc 0.8771 (0.9064 / 0.8479) Training 170: 84000 / 124344: Premsel loss 0.2780, acc 0.8762 (0.9021 / 0.8502) Training 170: 91000 / 124344: Premsel loss 0.2736, acc 0.8783 (0.9075 / 0.8490) Training 170: 98000 / 124344: Premsel loss 0.2715, acc 0.8795 (0.9071 / 0.8519) Training 170: 105000 / 124344: Premsel loss 0.2733, acc 0.8788 (0.9087 / 0.8489) Training 170: 112000 / 124344: Premsel loss 0.2743, acc 0.8786 (0.9060 / 0.8511) Training 170: 119000 / 124344: Premsel loss 0.2710, acc 0.8792 (0.9082 / 0.8501) Evaluation 170: Premsel loss 0.2744, acc 0.8774 (0.9023 / 0.8525) Training 171: 0 / 124344: Premsel loss 0.2695, acc 0.8813 (0.9126 / 0.8500) Training 171: 7000 / 124344: Premsel loss 0.2703, acc 0.8794 (0.9053 / 0.8535) Training 171: 14000 / 124344: Premsel loss 0.2730, acc 0.8784 (0.9074 / 0.8495) Training 171: 21000 / 124344: Premsel loss 0.2784, acc 0.8751 (0.9019 / 0.8482) Training 171: 28000 / 124344: Premsel loss 0.2741, acc 0.8775 (0.9139 / 0.8412) Training 171: 35000 / 124344: Premsel loss 0.2723, acc 0.8788 (0.9096 / 0.8480) Training 171: 42000 / 124344: Premsel loss 0.2712, acc 0.8789 (0.9063 / 0.8516) Training 171: 49000 / 124344: Premsel loss 0.2693, acc 0.8805 (0.9132 / 0.8477) Training 171: 56000 / 124344: Premsel loss 0.2702, acc 0.8793 (0.9056 / 0.8531) Training 171: 63000 / 124344: Premsel loss 0.2626, acc 0.8837 (0.9106 / 0.8568) Training 171: 70000 / 124344: Premsel loss 0.2654, acc 0.8821 (0.9097 / 0.8544) Training 171: 77000 / 124344: Premsel loss 0.2643, acc 0.8824 (0.9135 / 0.8514) Training 171: 84000 / 124344: Premsel loss 0.2711, acc 0.8800 (0.9086 / 0.8514) Training 171: 91000 / 124344: Premsel loss 0.2663, acc 0.8814 (0.9089 / 0.8538) Training 171: 98000 / 124344: Premsel loss 0.2669, acc 0.8814 (0.9110 / 0.8518) Training 171: 105000 / 124344: Premsel loss 0.2725, acc 0.8787 (0.9065 / 0.8508) Training 171: 112000 / 124344: Premsel loss 0.2692, acc 0.8807 (0.9087 / 0.8527) Training 171: 119000 / 124344: Premsel loss 0.2631, acc 0.8834 (0.9121 / 0.8546) Evaluation 171: Premsel loss 0.2701, acc 0.8800 (0.9315 / 0.8286) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_11.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo2___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_26.pkl.gz Training 172: 0 / 112082: Premsel loss 0.2646, acc 0.8823 (0.9064 / 0.8583) Training 172: 7000 / 112082: Premsel loss 0.3010, acc 0.8636 (0.8875 / 0.8398) Training 172: 14000 / 112082: Premsel loss 0.3064, acc 0.8614 (0.8793 / 0.8436) Training 172: 21000 / 112082: Premsel loss 0.3078, acc 0.8594 (0.8796 / 0.8392) Training 172: 28000 / 112082: Premsel loss 0.3214, acc 0.8537 (0.8766 / 0.8308) Training 172: 35000 / 112082: Premsel loss 0.3073, acc 0.8599 (0.8783 / 0.8414) Training 172: 42000 / 112082: Premsel loss 0.3074, acc 0.8596 (0.8770 / 0.8423) Training 172: 49000 / 112082: Premsel loss 0.3050, acc 0.8623 (0.8847 / 0.8400) Training 172: 56000 / 112082: Premsel loss 0.3060, acc 0.8607 (0.8829 / 0.8386) Training 172: 63000 / 112082: Premsel loss 0.3131, acc 0.8570 (0.8799 / 0.8340) Training 172: 70000 / 112082: Premsel loss 0.3127, acc 0.8567 (0.8852 / 0.8282) Training 172: 77000 / 112082: Premsel loss 0.3026, acc 0.8632 (0.8845 / 0.8419) Training 172: 84000 / 112082: Premsel loss 0.3105, acc 0.8591 (0.8823 / 0.8359) Training 172: 91000 / 112082: Premsel loss 0.3023, acc 0.8638 (0.8877 / 0.8398) Training 172: 98000 / 112082: Premsel loss 0.3013, acc 0.8638 (0.8850 / 0.8427) Training 172: 105000 / 112082: Premsel loss 0.3075, acc 0.8598 (0.8783 / 0.8412) Training 172: 112000 / 112082: Premsel loss 0.3049, acc 0.8611 (0.8858 / 0.8363) Evaluation 172: Premsel loss 0.3063, acc 0.8609 (0.8918 / 0.8300) Training 173: 0 / 112082: Premsel loss 0.3130, acc 0.8566 (0.8783 / 0.8349) Training 173: 7000 / 112082: Premsel loss 0.3019, acc 0.8633 (0.8830 / 0.8436) Training 173: 14000 / 112082: Premsel loss 0.3051, acc 0.8613 (0.8900 / 0.8326) Training 173: 21000 / 112082: Premsel loss 0.2988, acc 0.8659 (0.8904 / 0.8413) Training 173: 28000 / 112082: Premsel loss 0.2996, acc 0.8647 (0.8887 / 0.8408) Training 173: 35000 / 112082: Premsel loss 0.2983, acc 0.8650 (0.8885 / 0.8414) Training 173: 42000 / 112082: Premsel loss 0.3013, acc 0.8643 (0.8879 / 0.8408) Training 173: 49000 / 112082: Premsel loss 0.3032, acc 0.8630 (0.8871 / 0.8389) Training 173: 56000 / 112082: Premsel loss 0.3139, acc 0.8581 (0.8761 / 0.8400) Training 173: 63000 / 112082: Premsel loss 0.3054, acc 0.8618 (0.8804 / 0.8432) Training 173: 70000 / 112082: Premsel loss 0.3050, acc 0.8614 (0.8847 / 0.8380) Training 173: 77000 / 112082: Premsel loss 0.3072, acc 0.8610 (0.8830 / 0.8389) Training 173: 84000 / 112082: Premsel loss 0.3195, acc 0.8537 (0.8763 / 0.8310) Training 173: 91000 / 112082: Premsel loss 0.3163, acc 0.8554 (0.8815 / 0.8294) Training 173: 98000 / 112082: Premsel loss 0.3131, acc 0.8564 (0.8769 / 0.8360) Training 173: 105000 / 112082: Premsel loss 0.3109, acc 0.8576 (0.8818 / 0.8333) Training 173: 112000 / 112082: Premsel loss 0.3043, acc 0.8617 (0.8869 / 0.8366) Evaluation 173: Premsel loss 0.3073, acc 0.8605 (0.8897 / 0.8312) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d50-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_03.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop01-VHSLCAXPh+lgb-d50-l900-e0.15loop01+coop___out1.pkl.gz Training 174: 0 / 97568: Premsel loss 0.3116, acc 0.8576 (0.8758 / 0.8394) Training 174: 7000 / 97568: Premsel loss 0.2944, acc 0.8658 (0.8904 / 0.8411) Training 174: 14000 / 97568: Premsel loss 0.2981, acc 0.8649 (0.8861 / 0.8438) Training 174: 21000 / 97568: Premsel loss 0.2938, acc 0.8669 (0.8968 / 0.8370) Training 174: 28000 / 97568: Premsel loss 0.2932, acc 0.8667 (0.8940 / 0.8394) Training 174: 35000 / 97568: Premsel loss 0.2864, acc 0.8715 (0.8939 / 0.8490) Training 174: 42000 / 97568: Premsel loss 0.2941, acc 0.8682 (0.8971 / 0.8393) Training 174: 49000 / 97568: Premsel loss 0.2866, acc 0.8709 (0.8956 / 0.8462) Training 174: 56000 / 97568: Premsel loss 0.2945, acc 0.8658 (0.8922 / 0.8395) Training 174: 63000 / 97568: Premsel loss 0.2896, acc 0.8695 (0.8901 / 0.8490) Training 174: 70000 / 97568: Premsel loss 0.2928, acc 0.8683 (0.8949 / 0.8417) Training 174: 77000 / 97568: Premsel loss 0.2928, acc 0.8681 (0.8973 / 0.8390) Training 174: 84000 / 97568: Premsel loss 0.2867, acc 0.8703 (0.8988 / 0.8418) Training 174: 91000 / 97568: Premsel loss 0.2892, acc 0.8702 (0.8975 / 0.8429) Evaluation 174: Premsel loss 0.2863, acc 0.8715 (0.8954 / 0.8476) Training 175: 0 / 97568: Premsel loss 0.2881, acc 0.8707 (0.8966 / 0.8449) Training 175: 7000 / 97568: Premsel loss 0.2905, acc 0.8698 (0.8963 / 0.8433) Training 175: 14000 / 97568: Premsel loss 0.2846, acc 0.8717 (0.8974 / 0.8460) Training 175: 21000 / 97568: Premsel loss 0.2879, acc 0.8706 (0.8998 / 0.8414) Training 175: 28000 / 97568: Premsel loss 0.2948, acc 0.8666 (0.8916 / 0.8415) Training 175: 35000 / 97568: Premsel loss 0.2869, acc 0.8709 (0.8972 / 0.8446) Training 175: 42000 / 97568: Premsel loss 0.2937, acc 0.8667 (0.8884 / 0.8450) Training 175: 49000 / 97568: Premsel loss 0.2857, acc 0.8702 (0.8927 / 0.8477) Training 175: 56000 / 97568: Premsel loss 0.2866, acc 0.8715 (0.8975 / 0.8454) Training 175: 63000 / 97568: Premsel loss 0.2927, acc 0.8682 (0.8913 / 0.8451) Training 175: 70000 / 97568: Premsel loss 0.2961, acc 0.8656 (0.8963 / 0.8348) Training 175: 77000 / 97568: Premsel loss 0.3195, acc 0.8528 (0.8819 / 0.8237) Training 175: 84000 / 97568: Premsel loss 0.3039, acc 0.8616 (0.8845 / 0.8388) Training 175: 91000 / 97568: Premsel loss 0.2965, acc 0.8651 (0.8933 / 0.8369) Evaluation 175: Premsel loss 0.2998, acc 0.8640 (0.8880 / 0.8399) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_15-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_21.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz Training 176: 0 / 128869: Premsel loss 0.2992, acc 0.8636 (0.8869 / 0.8404) Training 176: 7000 / 128869: Premsel loss 0.3024, acc 0.8627 (0.8855 / 0.8399) Training 176: 14000 / 128869: Premsel loss 0.2943, acc 0.8675 (0.8955 / 0.8394) Training 176: 21000 / 128869: Premsel loss 0.2909, acc 0.8685 (0.8890 / 0.8480) Training 176: 28000 / 128869: Premsel loss 0.2901, acc 0.8699 (0.8936 / 0.8463) Training 176: 35000 / 128869: Premsel loss 0.2943, acc 0.8682 (0.8924 / 0.8439) Training 176: 42000 / 128869: Premsel loss 0.2850, acc 0.8720 (0.8952 / 0.8488) Training 176: 49000 / 128869: Premsel loss 0.2913, acc 0.8697 (0.8967 / 0.8427) Training 176: 56000 / 128869: Premsel loss 0.2822, acc 0.8740 (0.8983 / 0.8496) Training 176: 63000 / 128869: Premsel loss 0.2853, acc 0.8716 (0.8953 / 0.8479) Training 176: 70000 / 128869: Premsel loss 0.2927, acc 0.8686 (0.8959 / 0.8414) Training 176: 77000 / 128869: Premsel loss 0.2887, acc 0.8711 (0.8933 / 0.8489) Training 176: 84000 / 128869: Premsel loss 0.2825, acc 0.8736 (0.8993 / 0.8478) Training 176: 91000 / 128869: Premsel loss 0.2883, acc 0.8705 (0.8914 / 0.8495) Training 176: 98000 / 128869: Premsel loss 0.2887, acc 0.8683 (0.8941 / 0.8426) Training 176: 105000 / 128869: Premsel loss 0.2871, acc 0.8703 (0.8957 / 0.8449) Training 176: 112000 / 128869: Premsel loss 0.2953, acc 0.8669 (0.8980 / 0.8358) Training 176: 119000 / 128869: Premsel loss 0.2909, acc 0.8694 (0.8974 / 0.8415) Training 176: 126000 / 128869: Premsel loss 0.2873, acc 0.8695 (0.8940 / 0.8449) Evaluation 176: Premsel loss 0.2895, acc 0.8691 (0.9117 / 0.8266) Training 177: 0 / 128869: Premsel loss 0.2859, acc 0.8719 (0.8932 / 0.8507) Training 177: 7000 / 128869: Premsel loss 0.2780, acc 0.8762 (0.9021 / 0.8503) Training 177: 14000 / 128869: Premsel loss 0.2864, acc 0.8716 (0.8956 / 0.8476) Training 177: 21000 / 128869: Premsel loss 0.2820, acc 0.8729 (0.8953 / 0.8505) Training 177: 28000 / 128869: Premsel loss 0.2843, acc 0.8722 (0.8986 / 0.8457) Training 177: 35000 / 128869: Premsel loss 0.2780, acc 0.8764 (0.9016 / 0.8512) Training 177: 42000 / 128869: Premsel loss 0.2927, acc 0.8681 (0.8952 / 0.8410) Training 177: 49000 / 128869: Premsel loss 0.2877, acc 0.8713 (0.8940 / 0.8486) Training 177: 56000 / 128869: Premsel loss 0.2859, acc 0.8720 (0.8982 / 0.8458) Training 177: 63000 / 128869: Premsel loss 0.2897, acc 0.8691 (0.8906 / 0.8476) Training 177: 70000 / 128869: Premsel loss 0.2852, acc 0.8723 (0.8939 / 0.8508) Training 177: 77000 / 128869: Premsel loss 0.2851, acc 0.8714 (0.8934 / 0.8493) Training 177: 84000 / 128869: Premsel loss 0.2863, acc 0.8709 (0.8978 / 0.8439) Training 177: 91000 / 128869: Premsel loss 0.2827, acc 0.8731 (0.8966 / 0.8496) Training 177: 98000 / 128869: Premsel loss 0.2863, acc 0.8703 (0.9005 / 0.8401) Training 177: 105000 / 128869: Premsel loss 0.2881, acc 0.8704 (0.8929 / 0.8479) Training 177: 112000 / 128869: Premsel loss 0.2824, acc 0.8717 (0.8911 / 0.8523) Training 177: 119000 / 128869: Premsel loss 0.2906, acc 0.8696 (0.8904 / 0.8487) Training 177: 126000 / 128869: Premsel loss 0.2877, acc 0.8710 (0.8941 / 0.8478) Evaluation 177: Premsel loss 0.2815, acc 0.8736 (0.8999 / 0.8472) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_24.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t300-d60-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-coop___out1.pkl.gz Training 178: 0 / 117792: Premsel loss 0.2894, acc 0.8702 (0.8944 / 0.8460) Training 178: 7000 / 117792: Premsel loss 0.2870, acc 0.8708 (0.8951 / 0.8464) Training 178: 14000 / 117792: Premsel loss 0.2788, acc 0.8747 (0.9029 / 0.8465) Training 178: 21000 / 117792: Premsel loss 0.2806, acc 0.8752 (0.9020 / 0.8484) Training 178: 28000 / 117792: Premsel loss 0.2783, acc 0.8752 (0.8978 / 0.8527) Training 178: 35000 / 117792: Premsel loss 0.2851, acc 0.8723 (0.8975 / 0.8470) Training 178: 42000 / 117792: Premsel loss 0.2725, acc 0.8781 (0.9033 / 0.8529) Training 178: 49000 / 117792: Premsel loss 0.2816, acc 0.8738 (0.8971 / 0.8506) Training 178: 56000 / 117792: Premsel loss 0.2758, acc 0.8764 (0.9004 / 0.8525) Training 178: 63000 / 117792: Premsel loss 0.2796, acc 0.8756 (0.8985 / 0.8528) Training 178: 70000 / 117792: Premsel loss 0.2862, acc 0.8724 (0.8965 / 0.8484) Training 178: 77000 / 117792: Premsel loss 0.2785, acc 0.8761 (0.8979 / 0.8543) Training 178: 84000 / 117792: Premsel loss 0.2868, acc 0.8713 (0.8955 / 0.8470) Training 178: 91000 / 117792: Premsel loss 0.2921, acc 0.8699 (0.8945 / 0.8454) Training 178: 98000 / 117792: Premsel loss 0.2893, acc 0.8694 (0.8956 / 0.8432) Training 178: 105000 / 117792: Premsel loss 0.2835, acc 0.8728 (0.8993 / 0.8463) Training 178: 112000 / 117792: Premsel loss 0.2834, acc 0.8739 (0.9005 / 0.8474) Evaluation 178: Premsel loss 0.2780, acc 0.8760 (0.9017 / 0.8504) Training 179: 0 / 117792: Premsel loss 0.2816, acc 0.8735 (0.8979 / 0.8492) Training 179: 7000 / 117792: Premsel loss 0.2723, acc 0.8794 (0.9024 / 0.8564) Training 179: 14000 / 117792: Premsel loss 0.2794, acc 0.8759 (0.9015 / 0.8502) Training 179: 21000 / 117792: Premsel loss 0.2823, acc 0.8735 (0.9027 / 0.8443) Training 179: 28000 / 117792: Premsel loss 0.2787, acc 0.8768 (0.9053 / 0.8484) Training 179: 35000 / 117792: Premsel loss 0.2762, acc 0.8757 (0.8975 / 0.8540) Training 179: 42000 / 117792: Premsel loss 0.2830, acc 0.8723 (0.8996 / 0.8449) Training 179: 49000 / 117792: Premsel loss 0.2949, acc 0.8668 (0.8926 / 0.8411) Training 179: 56000 / 117792: Premsel loss 0.2849, acc 0.8723 (0.9019 / 0.8426) Training 179: 63000 / 117792: Premsel loss 0.2839, acc 0.8728 (0.9009 / 0.8447) Training 179: 70000 / 117792: Premsel loss 0.2872, acc 0.8720 (0.9017 / 0.8423) Training 179: 77000 / 117792: Premsel loss 0.2877, acc 0.8698 (0.8938 / 0.8459) Training 179: 84000 / 117792: Premsel loss 0.2892, acc 0.8700 (0.8989 / 0.8412) Training 179: 91000 / 117792: Premsel loss 0.3438, acc 0.8456 (0.8928 / 0.7985) Training 179: 98000 / 117792: Premsel loss 0.3045, acc 0.8623 (0.8932 / 0.8314) Training 179: 105000 / 117792: Premsel loss 0.2966, acc 0.8657 (0.8975 / 0.8338) Training 179: 112000 / 117792: Premsel loss 0.2914, acc 0.8695 (0.8966 / 0.8425) Evaluation 179: Premsel loss 0.2927, acc 0.8681 (0.8751 / 0.8611) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_20.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+mzr02-VHSLCAXPh+xgb-d12-e0.2-loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_74avg-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_22.pkl.gz Training 180: 0 / 114227: Premsel loss 0.2987, acc 0.8650 (0.8978 / 0.8323) Training 180: 7000 / 114227: Premsel loss 0.3132, acc 0.8570 (0.8875 / 0.8265) Training 180: 14000 / 114227: Premsel loss 0.3109, acc 0.8573 (0.8878 / 0.8269) Training 180: 21000 / 114227: Premsel loss 0.3046, acc 0.8610 (0.8898 / 0.8322) Training 180: 28000 / 114227: Premsel loss 0.3128, acc 0.8572 (0.8870 / 0.8274) Training 180: 35000 / 114227: Premsel loss 0.3032, acc 0.8623 (0.8867 / 0.8380) Training 180: 42000 / 114227: Premsel loss 0.3010, acc 0.8636 (0.8897 / 0.8376) Training 180: 49000 / 114227: Premsel loss 0.3013, acc 0.8635 (0.8897 / 0.8373) Training 180: 56000 / 114227: Premsel loss 0.3027, acc 0.8619 (0.8887 / 0.8351) Training 180: 63000 / 114227: Premsel loss 0.2971, acc 0.8659 (0.8928 / 0.8390) Training 180: 70000 / 114227: Premsel loss 0.2979, acc 0.8664 (0.8944 / 0.8384) Training 180: 77000 / 114227: Premsel loss 0.2975, acc 0.8663 (0.8923 / 0.8402) Training 180: 84000 / 114227: Premsel loss 0.2942, acc 0.8672 (0.8925 / 0.8418) Training 180: 91000 / 114227: Premsel loss 0.2938, acc 0.8674 (0.8985 / 0.8363) Training 180: 98000 / 114227: Premsel loss 0.2914, acc 0.8687 (0.8979 / 0.8396) Training 180: 105000 / 114227: Premsel loss 0.3061, acc 0.8609 (0.8912 / 0.8305) Training 180: 112000 / 114227: Premsel loss 0.2960, acc 0.8663 (0.8945 / 0.8381) Evaluation 180: Premsel loss 0.2923, acc 0.8689 (0.8999 / 0.8379) Training 181: 0 / 114227: Premsel loss 0.2933, acc 0.8679 (0.8963 / 0.8394) Training 181: 7000 / 114227: Premsel loss 0.3008, acc 0.8643 (0.8943 / 0.8342) Training 181: 14000 / 114227: Premsel loss 0.2959, acc 0.8665 (0.8906 / 0.8425) Training 181: 21000 / 114227: Premsel loss 0.3003, acc 0.8651 (0.8944 / 0.8358) Training 181: 28000 / 114227: Premsel loss 0.3087, acc 0.8589 (0.8889 / 0.8288) Training 181: 35000 / 114227: Premsel loss 0.3089, acc 0.8595 (0.8907 / 0.8283) Training 181: 42000 / 114227: Premsel loss 0.2998, acc 0.8641 (0.8912 / 0.8370) Training 181: 49000 / 114227: Premsel loss 0.3019, acc 0.8619 (0.8864 / 0.8374) Training 181: 56000 / 114227: Premsel loss 0.2981, acc 0.8650 (0.9015 / 0.8286) Training 181: 63000 / 114227: Premsel loss 0.3024, acc 0.8632 (0.8927 / 0.8336) Training 181: 70000 / 114227: Premsel loss 0.3006, acc 0.8646 (0.8928 / 0.8364) Training 181: 77000 / 114227: Premsel loss 0.2992, acc 0.8642 (0.8926 / 0.8358) Training 181: 84000 / 114227: Premsel loss 0.3044, acc 0.8621 (0.8883 / 0.8359) Training 181: 91000 / 114227: Premsel loss 0.2986, acc 0.8634 (0.8898 / 0.8370) Training 181: 98000 / 114227: Premsel loss 0.2898, acc 0.8689 (0.8962 / 0.8417) Training 181: 105000 / 114227: Premsel loss 0.2944, acc 0.8670 (0.8943 / 0.8396) Training 181: 112000 / 114227: Premsel loss 0.3008, acc 0.8623 (0.8882 / 0.8365) Evaluation 181: Premsel loss 0.2976, acc 0.8661 (0.8958 / 0.8363) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_00.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+solo___out1.pkl.gz Training 182: 0 / 112700: Premsel loss 0.3001, acc 0.8649 (0.8918 / 0.8379) Training 182: 7000 / 112700: Premsel loss 0.3001, acc 0.8640 (0.8844 / 0.8436) Training 182: 14000 / 112700: Premsel loss 0.3089, acc 0.8595 (0.8849 / 0.8342) Training 182: 21000 / 112700: Premsel loss 0.2973, acc 0.8653 (0.8891 / 0.8414) Training 182: 28000 / 112700: Premsel loss 0.3118, acc 0.8590 (0.8770 / 0.8410) Training 182: 35000 / 112700: Premsel loss 0.3074, acc 0.8611 (0.8819 / 0.8403) Training 182: 42000 / 112700: Premsel loss 0.3194, acc 0.8547 (0.8703 / 0.8391) Training 182: 49000 / 112700: Premsel loss 0.3358, acc 0.8444 (0.8780 / 0.8107) Training 182: 56000 / 112700: Premsel loss 0.3062, acc 0.8617 (0.8843 / 0.8392) Training 182: 63000 / 112700: Premsel loss 0.3155, acc 0.8556 (0.8729 / 0.8382) Training 182: 70000 / 112700: Premsel loss 0.3051, acc 0.8623 (0.8855 / 0.8392) Training 182: 77000 / 112700: Premsel loss 0.3078, acc 0.8599 (0.8816 / 0.8382) Training 182: 84000 / 112700: Premsel loss 0.3215, acc 0.8519 (0.8759 / 0.8280) Training 182: 91000 / 112700: Premsel loss 0.3113, acc 0.8563 (0.8789 / 0.8336) Training 182: 98000 / 112700: Premsel loss 0.3112, acc 0.8589 (0.8810 / 0.8367) Training 182: 105000 / 112700: Premsel loss 0.3041, acc 0.8624 (0.8803 / 0.8445) Training 182: 112000 / 112700: Premsel loss 0.3090, acc 0.8587 (0.8859 / 0.8315) Evaluation 182: Premsel loss 0.3126, acc 0.8574 (0.8896 / 0.8251) Training 183: 0 / 112700: Premsel loss 0.3169, acc 0.8549 (0.8794 / 0.8303) Training 183: 7000 / 112700: Premsel loss 0.3104, acc 0.8592 (0.8861 / 0.8323) Training 183: 14000 / 112700: Premsel loss 0.3082, acc 0.8602 (0.8866 / 0.8338) Training 183: 21000 / 112700: Premsel loss 0.2991, acc 0.8654 (0.8906 / 0.8401) Training 183: 28000 / 112700: Premsel loss 0.3089, acc 0.8581 (0.8889 / 0.8273) Training 183: 35000 / 112700: Premsel loss 0.2987, acc 0.8642 (0.8869 / 0.8415) Training 183: 42000 / 112700: Premsel loss 0.3151, acc 0.8546 (0.8788 / 0.8303) Training 183: 49000 / 112700: Premsel loss 0.3048, acc 0.8605 (0.8847 / 0.8362) Training 183: 56000 / 112700: Premsel loss 0.3102, acc 0.8580 (0.8810 / 0.8350) Training 183: 63000 / 112700: Premsel loss 0.3165, acc 0.8555 (0.8821 / 0.8289) Training 183: 70000 / 112700: Premsel loss 0.3125, acc 0.8578 (0.8818 / 0.8337) Training 183: 77000 / 112700: Premsel loss 0.3071, acc 0.8600 (0.8887 / 0.8312) Training 183: 84000 / 112700: Premsel loss 0.3262, acc 0.8485 (0.8724 / 0.8246) Training 183: 91000 / 112700: Premsel loss 0.3137, acc 0.8575 (0.8771 / 0.8378) Training 183: 98000 / 112700: Premsel loss 0.3040, acc 0.8619 (0.8854 / 0.8384) Training 183: 105000 / 112700: Premsel loss 0.3107, acc 0.8588 (0.8753 / 0.8423) Training 183: 112000 / 112700: Premsel loss 0.3145, acc 0.8560 (0.8797 / 0.8323) Evaluation 183: Premsel loss 0.3045, acc 0.8619 (0.8769 / 0.8469) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_23.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz Training 184: 0 / 93973: Premsel loss 0.3103, acc 0.8590 (0.8865 / 0.8315) Training 184: 7000 / 93973: Premsel loss 0.2995, acc 0.8628 (0.8913 / 0.8342) Training 184: 14000 / 93973: Premsel loss 0.3107, acc 0.8579 (0.8903 / 0.8255) Training 184: 21000 / 93973: Premsel loss 0.3082, acc 0.8600 (0.8854 / 0.8346) Training 184: 28000 / 93973: Premsel loss 0.3148, acc 0.8556 (0.8806 / 0.8305) Training 184: 35000 / 93973: Premsel loss 0.3149, acc 0.8557 (0.8776 / 0.8338) Training 184: 42000 / 93973: Premsel loss 0.3064, acc 0.8611 (0.8868 / 0.8354) Training 184: 49000 / 93973: Premsel loss 0.2980, acc 0.8654 (0.8905 / 0.8403) Training 184: 56000 / 93973: Premsel loss 0.3073, acc 0.8592 (0.8871 / 0.8314) Training 184: 63000 / 93973: Premsel loss 0.3018, acc 0.8632 (0.8915 / 0.8349) Training 184: 70000 / 93973: Premsel loss 0.3052, acc 0.8617 (0.8951 / 0.8284) Training 184: 77000 / 93973: Premsel loss 0.2948, acc 0.8670 (0.8965 / 0.8375) Training 184: 84000 / 93973: Premsel loss 0.3062, acc 0.8605 (0.8902 / 0.8308) Training 184: 91000 / 93973: Premsel loss 0.2968, acc 0.8655 (0.8910 / 0.8401) Evaluation 184: Premsel loss 0.2954, acc 0.8668 (0.8981 / 0.8355) Training 185: 0 / 93973: Premsel loss 0.2991, acc 0.8649 (0.8901 / 0.8397) Training 185: 7000 / 93973: Premsel loss 0.2963, acc 0.8658 (0.8923 / 0.8394) Training 185: 14000 / 93973: Premsel loss 0.3056, acc 0.8619 (0.8856 / 0.8382) Training 185: 21000 / 93973: Premsel loss 0.2988, acc 0.8646 (0.9007 / 0.8285) Training 185: 28000 / 93973: Premsel loss 0.2968, acc 0.8659 (0.8942 / 0.8376) Training 185: 35000 / 93973: Premsel loss 0.2945, acc 0.8666 (0.8967 / 0.8365) Training 185: 42000 / 93973: Premsel loss 0.2943, acc 0.8673 (0.8959 / 0.8387) Training 185: 49000 / 93973: Premsel loss 0.2984, acc 0.8656 (0.8962 / 0.8350) Training 185: 56000 / 93973: Premsel loss 0.3077, acc 0.8597 (0.8912 / 0.8283) Training 185: 63000 / 93973: Premsel loss 0.3041, acc 0.8611 (0.8874 / 0.8348) Training 185: 70000 / 93973: Premsel loss 0.3031, acc 0.8625 (0.8986 / 0.8263) Training 185: 77000 / 93973: Premsel loss 0.3053, acc 0.8604 (0.8877 / 0.8332) Training 185: 84000 / 93973: Premsel loss 0.2974, acc 0.8656 (0.8898 / 0.8415) Training 185: 91000 / 93973: Premsel loss 0.2959, acc 0.8659 (0.8933 / 0.8385) Evaluation 185: Premsel loss 0.2991, acc 0.8651 (0.8938 / 0.8364) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_25.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l900-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+coop___out1.pkl.gz Training 186: 0 / 98899: Premsel loss 0.2995, acc 0.8650 (0.8932 / 0.8368) Training 186: 7000 / 98899: Premsel loss 0.3026, acc 0.8617 (0.8858 / 0.8375) Training 186: 14000 / 98899: Premsel loss 0.2975, acc 0.8664 (0.8940 / 0.8387) Training 186: 21000 / 98899: Premsel loss 0.2918, acc 0.8692 (0.8942 / 0.8441) Training 186: 28000 / 98899: Premsel loss 0.2935, acc 0.8671 (0.8955 / 0.8387) Training 186: 35000 / 98899: Premsel loss 0.2863, acc 0.8719 (0.8995 / 0.8443) Training 186: 42000 / 98899: Premsel loss 0.3002, acc 0.8634 (0.8913 / 0.8355) Training 186: 49000 / 98899: Premsel loss 0.2983, acc 0.8651 (0.8930 / 0.8373) Training 186: 56000 / 98899: Premsel loss 0.2986, acc 0.8650 (0.8902 / 0.8398) Training 186: 63000 / 98899: Premsel loss 0.2931, acc 0.8683 (0.8900 / 0.8466) Training 186: 70000 / 98899: Premsel loss 0.2965, acc 0.8664 (0.8968 / 0.8361) Training 186: 77000 / 98899: Premsel loss 0.2974, acc 0.8649 (0.8904 / 0.8394) Training 186: 84000 / 98899: Premsel loss 0.2936, acc 0.8678 (0.8915 / 0.8442) Training 186: 91000 / 98899: Premsel loss 0.2821, acc 0.8743 (0.8978 / 0.8507) Training 186: 98000 / 98899: Premsel loss 0.2882, acc 0.8710 (0.8954 / 0.8466) Evaluation 186: Premsel loss 0.2897, acc 0.8694 (0.8798 / 0.8590) Training 187: 0 / 98899: Premsel loss 0.2953, acc 0.8664 (0.8939 / 0.8390) Training 187: 7000 / 98899: Premsel loss 0.2946, acc 0.8678 (0.8973 / 0.8384) Training 187: 14000 / 98899: Premsel loss 0.2935, acc 0.8676 (0.8965 / 0.8387) Training 187: 21000 / 98899: Premsel loss 0.2998, acc 0.8648 (0.8878 / 0.8417) Training 187: 28000 / 98899: Premsel loss 0.2967, acc 0.8656 (0.8909 / 0.8402) Training 187: 35000 / 98899: Premsel loss 0.2908, acc 0.8690 (0.8967 / 0.8412) Training 187: 42000 / 98899: Premsel loss 0.2875, acc 0.8716 (0.9035 / 0.8397) Training 187: 49000 / 98899: Premsel loss 0.2912, acc 0.8684 (0.8919 / 0.8449) Training 187: 56000 / 98899: Premsel loss 0.2960, acc 0.8670 (0.8920 / 0.8420) Training 187: 63000 / 98899: Premsel loss 0.2887, acc 0.8708 (0.8982 / 0.8435) Training 187: 70000 / 98899: Premsel loss 0.2828, acc 0.8734 (0.8999 / 0.8469) Training 187: 77000 / 98899: Premsel loss 0.2933, acc 0.8667 (0.8939 / 0.8394) Training 187: 84000 / 98899: Premsel loss 0.2942, acc 0.8675 (0.8894 / 0.8456) Training 187: 91000 / 98899: Premsel loss 0.3006, acc 0.8642 (0.8840 / 0.8444) Training 187: 98000 / 98899: Premsel loss 0.2971, acc 0.8660 (0.8920 / 0.8400) Evaluation 187: Premsel loss 0.2913, acc 0.8687 (0.9050 / 0.8324) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_17.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l1800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_10.pkl.gz Training 188: 0 / 100171: Premsel loss 0.2919, acc 0.8699 (0.8991 / 0.8407) Training 188: 7000 / 100171: Premsel loss 0.3147, acc 0.8569 (0.8863 / 0.8275) Training 188: 14000 / 100171: Premsel loss 0.3090, acc 0.8600 (0.8869 / 0.8331) Training 188: 21000 / 100171: Premsel loss 0.3076, acc 0.8597 (0.8803 / 0.8390) Training 188: 28000 / 100171: Premsel loss 0.3229, acc 0.8539 (0.8865 / 0.8214) Training 188: 35000 / 100171: Premsel loss 0.3163, acc 0.8565 (0.8810 / 0.8320) Training 188: 42000 / 100171: Premsel loss 0.3100, acc 0.8586 (0.8850 / 0.8321) Training 188: 49000 / 100171: Premsel loss 0.3180, acc 0.8551 (0.8847 / 0.8255) Training 188: 56000 / 100171: Premsel loss 0.3154, acc 0.8560 (0.8832 / 0.8289) Training 188: 63000 / 100171: Premsel loss 0.3105, acc 0.8593 (0.8861 / 0.8326) Training 188: 70000 / 100171: Premsel loss 0.3145, acc 0.8567 (0.8834 / 0.8299) Training 188: 77000 / 100171: Premsel loss 0.3129, acc 0.8579 (0.8815 / 0.8343) Training 188: 84000 / 100171: Premsel loss 0.3137, acc 0.8581 (0.8868 / 0.8294) Training 188: 91000 / 100171: Premsel loss 0.3205, acc 0.8530 (0.8747 / 0.8313) Training 188: 98000 / 100171: Premsel loss 0.3156, acc 0.8555 (0.8813 / 0.8297) Evaluation 188: Premsel loss 0.3215, acc 0.8528 (0.8322 / 0.8734) Training 189: 0 / 100171: Premsel loss 0.3158, acc 0.8561 (0.8868 / 0.8254) Training 189: 7000 / 100171: Premsel loss 0.3126, acc 0.8578 (0.8817 / 0.8338) Training 189: 14000 / 100171: Premsel loss 0.3064, acc 0.8611 (0.8852 / 0.8371) Training 189: 21000 / 100171: Premsel loss 0.3044, acc 0.8628 (0.8841 / 0.8414) Training 189: 28000 / 100171: Premsel loss 0.3062, acc 0.8614 (0.8851 / 0.8377) Training 189: 35000 / 100171: Premsel loss 0.3055, acc 0.8620 (0.8917 / 0.8323) Training 189: 42000 / 100171: Premsel loss 0.3102, acc 0.8592 (0.8837 / 0.8347) Training 189: 49000 / 100171: Premsel loss 0.3098, acc 0.8601 (0.8875 / 0.8327) Training 189: 56000 / 100171: Premsel loss 0.3023, acc 0.8638 (0.8921 / 0.8356) Training 189: 63000 / 100171: Premsel loss 0.3086, acc 0.8601 (0.8894 / 0.8308) Training 189: 70000 / 100171: Premsel loss 0.3047, acc 0.8626 (0.8904 / 0.8347) Training 189: 77000 / 100171: Premsel loss 0.3074, acc 0.8607 (0.8837 / 0.8378) Training 189: 84000 / 100171: Premsel loss 0.2987, acc 0.8652 (0.8895 / 0.8409) Training 189: 91000 / 100171: Premsel loss 0.3049, acc 0.8624 (0.8841 / 0.8407) Training 189: 98000 / 100171: Premsel loss 0.3020, acc 0.8634 (0.8908 / 0.8360) Evaluation 189: Premsel loss 0.3091, acc 0.8597 (0.8555 / 0.8638) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo1___bb_preds__0.05___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_27.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l6400-e0.15+coop-mzr02___out1.pkl.gz Training 190: 0 / 128456: Premsel loss 0.2970, acc 0.8602 (0.8987 / 0.8217) Training 190: 7000 / 128456: Premsel loss 0.3010, acc 0.8616 (0.8919 / 0.8313) Training 190: 14000 / 128456: Premsel loss 0.2982, acc 0.8642 (0.8925 / 0.8360) Training 190: 21000 / 128456: Premsel loss 0.3029, acc 0.8616 (0.8914 / 0.8319) Training 190: 28000 / 128456: Premsel loss 0.3026, acc 0.8620 (0.8898 / 0.8341) Training 190: 35000 / 128456: Premsel loss 0.3021, acc 0.8619 (0.8840 / 0.8397) Training 190: 42000 / 128456: Premsel loss 0.3104, acc 0.8570 (0.8794 / 0.8347) Training 190: 49000 / 128456: Premsel loss 0.3124, acc 0.8570 (0.8802 / 0.8338) Training 190: 56000 / 128456: Premsel loss 0.3198, acc 0.8522 (0.8764 / 0.8279) Training 190: 63000 / 128456: Premsel loss 0.3182, acc 0.8530 (0.8832 / 0.8228) Training 190: 70000 / 128456: Premsel loss 0.3133, acc 0.8573 (0.8865 / 0.8281) Training 190: 77000 / 128456: Premsel loss 0.3152, acc 0.8550 (0.8836 / 0.8264) Training 190: 84000 / 128456: Premsel loss 0.3064, acc 0.8610 (0.8861 / 0.8359) Training 190: 91000 / 128456: Premsel loss 0.3076, acc 0.8597 (0.8826 / 0.8368) Training 190: 98000 / 128456: Premsel loss 0.3021, acc 0.8624 (0.8934 / 0.8314) Training 190: 105000 / 128456: Premsel loss 0.3023, acc 0.8618 (0.8884 / 0.8352) Training 190: 112000 / 128456: Premsel loss 0.2997, acc 0.8627 (0.8877 / 0.8378) Training 190: 119000 / 128456: Premsel loss 0.3083, acc 0.8597 (0.8872 / 0.8322) Training 190: 126000 / 128456: Premsel loss 0.3060, acc 0.8612 (0.8884 / 0.8341) Evaluation 190: Premsel loss 0.3079, acc 0.8600 (0.8910 / 0.8290) Training 191: 0 / 128456: Premsel loss 0.3033, acc 0.8616 (0.8845 / 0.8387) Training 191: 7000 / 128456: Premsel loss 0.3053, acc 0.8613 (0.8846 / 0.8380) Training 191: 14000 / 128456: Premsel loss 0.3069, acc 0.8609 (0.8866 / 0.8351) Training 191: 21000 / 128456: Premsel loss 0.3116, acc 0.8571 (0.8896 / 0.8246) Training 191: 28000 / 128456: Premsel loss 0.3090, acc 0.8594 (0.8836 / 0.8351) Training 191: 35000 / 128456: Premsel loss 0.3135, acc 0.8566 (0.8824 / 0.8308) Training 191: 42000 / 128456: Premsel loss 0.2975, acc 0.8648 (0.8931 / 0.8366) Training 191: 49000 / 128456: Premsel loss 0.3043, acc 0.8615 (0.8839 / 0.8392) Training 191: 56000 / 128456: Premsel loss 0.3057, acc 0.8616 (0.8856 / 0.8375) Training 191: 63000 / 128456: Premsel loss 0.2999, acc 0.8632 (0.8873 / 0.8392) Training 191: 70000 / 128456: Premsel loss 0.3049, acc 0.8613 (0.8832 / 0.8394) Training 191: 77000 / 128456: Premsel loss 0.3065, acc 0.8602 (0.8879 / 0.8324) Training 191: 84000 / 128456: Premsel loss 0.3147, acc 0.8565 (0.8833 / 0.8296) Training 191: 91000 / 128456: Premsel loss 0.3216, acc 0.8520 (0.8806 / 0.8234) Training 191: 98000 / 128456: Premsel loss 0.3127, acc 0.8570 (0.8784 / 0.8356) Training 191: 105000 / 128456: Premsel loss 0.3044, acc 0.8609 (0.8878 / 0.8341) Training 191: 112000 / 128456: Premsel loss 0.3129, acc 0.8570 (0.8779 / 0.8362) Training 191: 119000 / 128456: Premsel loss 0.3090, acc 0.8590 (0.8840 / 0.8341) Training 191: 126000 / 128456: Premsel loss 0.3044, acc 0.8615 (0.8898 / 0.8331) Evaluation 191: Premsel loss 0.3139, acc 0.8557 (0.9049 / 0.8065) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_12.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz Training 192: 0 / 101042: Premsel loss 0.3216, acc 0.8504 (0.8762 / 0.8245) Training 192: 7000 / 101042: Premsel loss 0.2906, acc 0.8687 (0.8941 / 0.8432) Training 192: 14000 / 101042: Premsel loss 0.2955, acc 0.8668 (0.8900 / 0.8436) Training 192: 21000 / 101042: Premsel loss 0.2925, acc 0.8685 (0.8881 / 0.8489) Training 192: 28000 / 101042: Premsel loss 0.2828, acc 0.8744 (0.8941 / 0.8547) Training 192: 35000 / 101042: Premsel loss 0.2967, acc 0.8663 (0.8887 / 0.8440) Training 192: 42000 / 101042: Premsel loss 0.2944, acc 0.8662 (0.8889 / 0.8435) Training 192: 49000 / 101042: Premsel loss 0.2869, acc 0.8698 (0.8903 / 0.8493) Training 192: 56000 / 101042: Premsel loss 0.2870, acc 0.8709 (0.8923 / 0.8494) Training 192: 63000 / 101042: Premsel loss 0.2911, acc 0.8695 (0.8904 / 0.8486) Training 192: 70000 / 101042: Premsel loss 0.2878, acc 0.8700 (0.8910 / 0.8490) Training 192: 77000 / 101042: Premsel loss 0.2875, acc 0.8703 (0.8912 / 0.8494) Training 192: 84000 / 101042: Premsel loss 0.2894, acc 0.8689 (0.8980 / 0.8398) Training 192: 91000 / 101042: Premsel loss 0.2891, acc 0.8693 (0.8935 / 0.8450) Training 192: 98000 / 101042: Premsel loss 0.2927, acc 0.8672 (0.8881 / 0.8462) Evaluation 192: Premsel loss 0.2932, acc 0.8673 (0.9139 / 0.8207) Training 193: 0 / 101042: Premsel loss 0.2923, acc 0.8682 (0.8937 / 0.8428) Training 193: 7000 / 101042: Premsel loss 0.2907, acc 0.8695 (0.8925 / 0.8465) Training 193: 14000 / 101042: Premsel loss 0.2814, acc 0.8743 (0.8918 / 0.8568) Training 193: 21000 / 101042: Premsel loss 0.2808, acc 0.8744 (0.8976 / 0.8512) Training 193: 28000 / 101042: Premsel loss 0.2814, acc 0.8741 (0.8987 / 0.8495) Training 193: 35000 / 101042: Premsel loss 0.2878, acc 0.8700 (0.8984 / 0.8416) Training 193: 42000 / 101042: Premsel loss 0.2904, acc 0.8685 (0.8935 / 0.8435) Training 193: 49000 / 101042: Premsel loss 0.2920, acc 0.8677 (0.8911 / 0.8443) Training 193: 56000 / 101042: Premsel loss 0.2839, acc 0.8716 (0.8941 / 0.8490) Training 193: 63000 / 101042: Premsel loss 0.2881, acc 0.8700 (0.8966 / 0.8433) Training 193: 70000 / 101042: Premsel loss 0.2893, acc 0.8691 (0.8950 / 0.8431) Training 193: 77000 / 101042: Premsel loss 0.2885, acc 0.8694 (0.8981 / 0.8408) Training 193: 84000 / 101042: Premsel loss 0.2971, acc 0.8650 (0.8864 / 0.8436) Training 193: 91000 / 101042: Premsel loss 0.2877, acc 0.8709 (0.8944 / 0.8474) Training 193: 98000 / 101042: Premsel loss 0.2885, acc 0.8694 (0.8932 / 0.8457) Evaluation 193: Premsel loss 0.2907, acc 0.8690 (0.8799 / 0.8582) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l700-e0.20+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_02.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz Training 194: 0 / 123762: Premsel loss 0.2875, acc 0.8702 (0.8946 / 0.8459) Training 194: 7000 / 123762: Premsel loss 0.3117, acc 0.8591 (0.8717 / 0.8465) Training 194: 14000 / 123762: Premsel loss 0.3243, acc 0.8531 (0.8698 / 0.8363) Training 194: 21000 / 123762: Premsel loss 0.3193, acc 0.8552 (0.8667 / 0.8438) Training 194: 28000 / 123762: Premsel loss 0.3060, acc 0.8612 (0.8726 / 0.8499) Training 194: 35000 / 123762: Premsel loss 0.3088, acc 0.8595 (0.8746 / 0.8445) Training 194: 42000 / 123762: Premsel loss 0.3084, acc 0.8596 (0.8722 / 0.8471) Training 194: 49000 / 123762: Premsel loss 0.3190, acc 0.8539 (0.8731 / 0.8348) Training 194: 56000 / 123762: Premsel loss 0.3140, acc 0.8574 (0.8720 / 0.8428) Training 194: 63000 / 123762: Premsel loss 0.3173, acc 0.8551 (0.8715 / 0.8388) Training 194: 70000 / 123762: Premsel loss 0.3037, acc 0.8626 (0.8718 / 0.8534) Training 194: 77000 / 123762: Premsel loss 0.3055, acc 0.8620 (0.8780 / 0.8461) Training 194: 84000 / 123762: Premsel loss 0.3061, acc 0.8621 (0.8835 / 0.8408) Training 194: 91000 / 123762: Premsel loss 0.3066, acc 0.8617 (0.8771 / 0.8464) Training 194: 98000 / 123762: Premsel loss 0.3058, acc 0.8611 (0.8782 / 0.8439) Training 194: 105000 / 123762: Premsel loss 0.5954, acc 0.7940 (0.8649 / 0.7232) Training 194: 112000 / 123762: Premsel loss 0.3583, acc 0.8335 (0.8476 / 0.8194) Training 194: 119000 / 123762: Premsel loss 0.3418, acc 0.8430 (0.8507 / 0.8353) Evaluation 194: Premsel loss 0.3414, acc 0.8466 (0.8369 / 0.8564) Training 195: 0 / 123762: Premsel loss 0.3414, acc 0.8441 (0.8777 / 0.8105) Training 195: 7000 / 123762: Premsel loss 0.3276, acc 0.8505 (0.8718 / 0.8292) Training 195: 14000 / 123762: Premsel loss 0.3250, acc 0.8506 (0.8586 / 0.8427) Training 195: 21000 / 123762: Premsel loss 0.3283, acc 0.8493 (0.8755 / 0.8231) Training 195: 28000 / 123762: Premsel loss 0.3245, acc 0.8519 (0.8664 / 0.8374) Training 195: 35000 / 123762: Premsel loss 0.3173, acc 0.8562 (0.8725 / 0.8399) Training 195: 42000 / 123762: Premsel loss 0.3222, acc 0.8544 (0.8678 / 0.8410) Training 195: 49000 / 123762: Premsel loss 0.3128, acc 0.8581 (0.8751 / 0.8410) Training 195: 56000 / 123762: Premsel loss 0.3085, acc 0.8600 (0.8746 / 0.8453) Training 195: 63000 / 123762: Premsel loss 0.3167, acc 0.8554 (0.8739 / 0.8368) Training 195: 70000 / 123762: Premsel loss 0.3149, acc 0.8558 (0.8776 / 0.8340) Training 195: 77000 / 123762: Premsel loss 0.3167, acc 0.8554 (0.8707 / 0.8401) Training 195: 84000 / 123762: Premsel loss 0.3190, acc 0.8549 (0.8638 / 0.8459) Training 195: 91000 / 123762: Premsel loss 0.3095, acc 0.8604 (0.8773 / 0.8435) Training 195: 98000 / 123762: Premsel loss 0.3200, acc 0.8540 (0.8735 / 0.8346) Training 195: 105000 / 123762: Premsel loss 0.3205, acc 0.8543 (0.8711 / 0.8374) Training 195: 112000 / 123762: Premsel loss 0.3132, acc 0.8575 (0.8756 / 0.8395) Training 195: 119000 / 123762: Premsel loss 0.3109, acc 0.8586 (0.8734 / 0.8437) Evaluation 195: Premsel loss 0.3176, acc 0.8560 (0.8405 / 0.8714) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz Training 196: 0 / 121824: Premsel loss 0.3295, acc 0.8503 (0.8773 / 0.8233) Training 196: 7000 / 121824: Premsel loss 0.2640, acc 0.8849 (0.9120 / 0.8578) Training 196: 14000 / 121824: Premsel loss 0.2550, acc 0.8897 (0.9152 / 0.8642) Training 196: 21000 / 121824: Premsel loss 0.2591, acc 0.8863 (0.9172 / 0.8555) Training 196: 28000 / 121824: Premsel loss 0.2551, acc 0.8875 (0.9179 / 0.8570) Training 196: 35000 / 121824: Premsel loss 0.2813, acc 0.8764 (0.9128 / 0.8400) Training 196: 42000 / 121824: Premsel loss 0.2642, acc 0.8843 (0.9128 / 0.8557) Training 196: 49000 / 121824: Premsel loss 0.2601, acc 0.8863 (0.9149 / 0.8576) Training 196: 56000 / 121824: Premsel loss 0.2533, acc 0.8897 (0.9171 / 0.8624) Training 196: 63000 / 121824: Premsel loss 0.2592, acc 0.8867 (0.9115 / 0.8620) Training 196: 70000 / 121824: Premsel loss 0.2552, acc 0.8890 (0.9187 / 0.8594) Training 196: 77000 / 121824: Premsel loss 0.2593, acc 0.8865 (0.9145 / 0.8585) Training 196: 84000 / 121824: Premsel loss 0.2501, acc 0.8914 (0.9190 / 0.8637) Training 196: 91000 / 121824: Premsel loss 0.2542, acc 0.8898 (0.9175 / 0.8622) Training 196: 98000 / 121824: Premsel loss 0.2548, acc 0.8877 (0.9116 / 0.8638) Training 196: 105000 / 121824: Premsel loss 0.2514, acc 0.8913 (0.9161 / 0.8665) Training 196: 112000 / 121824: Premsel loss 0.2482, acc 0.8917 (0.9158 / 0.8676) Training 196: 119000 / 121824: Premsel loss 0.2590, acc 0.8861 (0.9218 / 0.8505) Evaluation 196: Premsel loss 0.2481, acc 0.8924 (0.9041 / 0.8808) Training 197: 0 / 121824: Premsel loss 0.2477, acc 0.8921 (0.9220 / 0.8623) Training 197: 7000 / 121824: Premsel loss 0.2463, acc 0.8929 (0.9182 / 0.8675) Training 197: 14000 / 121824: Premsel loss 0.2500, acc 0.8905 (0.9179 / 0.8632) Training 197: 21000 / 121824: Premsel loss 0.2554, acc 0.8884 (0.9155 / 0.8613) Training 197: 28000 / 121824: Premsel loss 0.2596, acc 0.8863 (0.9127 / 0.8598) Training 197: 35000 / 121824: Premsel loss 0.2590, acc 0.8864 (0.9120 / 0.8607) Training 197: 42000 / 121824: Premsel loss 0.2489, acc 0.8922 (0.9221 / 0.8622) Training 197: 49000 / 121824: Premsel loss 0.2497, acc 0.8923 (0.9253 / 0.8593) Training 197: 56000 / 121824: Premsel loss 0.2548, acc 0.8896 (0.9171 / 0.8620) Training 197: 63000 / 121824: Premsel loss 0.2534, acc 0.8894 (0.9128 / 0.8660) Training 197: 70000 / 121824: Premsel loss 0.2530, acc 0.8891 (0.9151 / 0.8631) Training 197: 77000 / 121824: Premsel loss 0.2526, acc 0.8894 (0.9188 / 0.8600) Training 197: 84000 / 121824: Premsel loss 0.2585, acc 0.8865 (0.9183 / 0.8547) Training 197: 91000 / 121824: Premsel loss 0.2506, acc 0.8919 (0.9229 / 0.8610) Training 197: 98000 / 121824: Premsel loss 0.2594, acc 0.8864 (0.9122 / 0.8606) Training 197: 105000 / 121824: Premsel loss 0.2622, acc 0.8858 (0.9105 / 0.8611) Training 197: 112000 / 121824: Premsel loss 0.2592, acc 0.8870 (0.9115 / 0.8625) Training 197: 119000 / 121824: Premsel loss 0.2531, acc 0.8907 (0.9118 / 0.8696) Evaluation 197: Premsel loss 0.2508, acc 0.8908 (0.9121 / 0.8696) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_73-query128-ctx256-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz Training 198: 0 / 67629: Premsel loss 0.2659, acc 0.8828 (0.9111 / 0.8546) Training 198: 7000 / 67629: Premsel loss 0.2984, acc 0.8662 (0.8740 / 0.8583) Training 198: 14000 / 67629: Premsel loss 0.3058, acc 0.8632 (0.8729 / 0.8534) Training 198: 21000 / 67629: Premsel loss 0.2978, acc 0.8673 (0.8751 / 0.8595) Training 198: 28000 / 67629: Premsel loss 0.3017, acc 0.8650 (0.8786 / 0.8513) Training 198: 35000 / 67629: Premsel loss 0.2935, acc 0.8699 (0.8838 / 0.8560) Training 198: 42000 / 67629: Premsel loss 0.2949, acc 0.8679 (0.8753 / 0.8605) Training 198: 49000 / 67629: Premsel loss 0.2923, acc 0.8689 (0.8808 / 0.8569) Training 198: 56000 / 67629: Premsel loss 0.2832, acc 0.8742 (0.8865 / 0.8618) Training 198: 63000 / 67629: Premsel loss 0.2902, acc 0.8712 (0.8862 / 0.8562) Evaluation 198: Premsel loss 0.2864, acc 0.8722 (0.8721 / 0.8722) Training 199: 0 / 67629: Premsel loss 0.2892, acc 0.8695 (0.8778 / 0.8612) Training 199: 7000 / 67629: Premsel loss 0.2829, acc 0.8740 (0.8873 / 0.8607) Training 199: 14000 / 67629: Premsel loss 0.2828, acc 0.8721 (0.8841 / 0.8600) Training 199: 21000 / 67629: Premsel loss 0.2807, acc 0.8764 (0.8857 / 0.8671) Training 199: 28000 / 67629: Premsel loss 0.2816, acc 0.8754 (0.8873 / 0.8636) Training 199: 35000 / 67629: Premsel loss 0.2874, acc 0.8728 (0.8798 / 0.8658) Training 199: 42000 / 67629: Premsel loss 0.2874, acc 0.8712 (0.8827 / 0.8596) Training 199: 49000 / 67629: Premsel loss 0.2870, acc 0.8731 (0.8876 / 0.8586) Training 199: 56000 / 67629: Premsel loss 0.2780, acc 0.8758 (0.8922 / 0.8594) Training 199: 63000 / 67629: Premsel loss 0.2870, acc 0.8724 (0.8859 / 0.8589) Evaluation 199: Premsel loss 0.2872, acc 0.8719 (0.8737 / 0.8702) Loading data... Full data reached, reshuffling... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_20-query192-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d80-l32000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop02_48-query256-ctx768-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_16.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_65-query256-ctx768-w0-coop___out1.pkl.gz Training 200: 0 / 123014: Premsel loss 0.3042, acc 0.8647 (0.8899 / 0.8394) Training 200: 7000 / 123014: Premsel loss 0.3246, acc 0.8531 (0.8794 / 0.8268) Training 200: 14000 / 123014: Premsel loss 0.3077, acc 0.8604 (0.8805 / 0.8403) Training 200: 21000 / 123014: Premsel loss 0.2986, acc 0.8652 (0.8929 / 0.8375) Training 200: 28000 / 123014: Premsel loss 0.3058, acc 0.8606 (0.8866 / 0.8345) Training 200: 35000 / 123014: Premsel loss 0.3022, acc 0.8629 (0.8917 / 0.8341) Training 200: 42000 / 123014: Premsel loss 0.2970, acc 0.8657 (0.8915 / 0.8399) Training 200: 49000 / 123014: Premsel loss 0.3026, acc 0.8626 (0.8858 / 0.8394) Training 200: 56000 / 123014: Premsel loss 0.3050, acc 0.8616 (0.8855 / 0.8377) Training 200: 63000 / 123014: Premsel loss 0.2908, acc 0.8695 (0.8922 / 0.8467) Training 200: 70000 / 123014: Premsel loss 0.2903, acc 0.8687 (0.8932 / 0.8441) Training 200: 77000 / 123014: Premsel loss 0.2880, acc 0.8703 (0.8981 / 0.8425) Training 200: 84000 / 123014: Premsel loss 0.2938, acc 0.8684 (0.8974 / 0.8394) Training 200: 91000 / 123014: Premsel loss 0.2933, acc 0.8669 (0.8930 / 0.8409) Training 200: 98000 / 123014: Premsel loss 0.2991, acc 0.8641 (0.8928 / 0.8353) Training 200: 105000 / 123014: Premsel loss 0.2926, acc 0.8684 (0.8932 / 0.8437) Training 200: 112000 / 123014: Premsel loss 0.2927, acc 0.8675 (0.8950 / 0.8400) Training 200: 119000 / 123014: Premsel loss 0.2939, acc 0.8669 (0.8936 / 0.8402) Evaluation 200: Premsel loss 0.2931, acc 0.8671 (0.8931 / 0.8410) Training 201: 0 / 123014: Premsel loss 0.3005, acc 0.8647 (0.8971 / 0.8322) Training 201: 7000 / 123014: Premsel loss 0.2954, acc 0.8683 (0.8988 / 0.8377) Training 201: 14000 / 123014: Premsel loss 0.3003, acc 0.8654 (0.8939 / 0.8368) Training 201: 21000 / 123014: Premsel loss 0.2980, acc 0.8643 (0.8922 / 0.8365) Training 201: 28000 / 123014: Premsel loss 0.2936, acc 0.8679 (0.8962 / 0.8396) Training 201: 35000 / 123014: Premsel loss 0.2923, acc 0.8684 (0.8990 / 0.8378) Training 201: 42000 / 123014: Premsel loss 0.2899, acc 0.8690 (0.8986 / 0.8395) Training 201: 49000 / 123014: Premsel loss 0.2905, acc 0.8691 (0.8973 / 0.8409) Training 201: 56000 / 123014: Premsel loss 0.2918, acc 0.8680 (0.8904 / 0.8457) Training 201: 63000 / 123014: Premsel loss 0.2900, acc 0.8688 (0.9013 / 0.8364) Training 201: 70000 / 123014: Premsel loss 0.2870, acc 0.8703 (0.8971 / 0.8435) Training 201: 77000 / 123014: Premsel loss 0.2930, acc 0.8682 (0.8932 / 0.8432) Training 201: 84000 / 123014: Premsel loss 0.2816, acc 0.8731 (0.8996 / 0.8465) Training 201: 91000 / 123014: Premsel loss 0.2918, acc 0.8679 (0.8969 / 0.8389) Training 201: 98000 / 123014: Premsel loss 0.2913, acc 0.8679 (0.9000 / 0.8359) Training 201: 105000 / 123014: Premsel loss 0.2830, acc 0.8736 (0.8945 / 0.8526) Training 201: 112000 / 123014: Premsel loss 0.2921, acc 0.8684 (0.8987 / 0.8382) Training 201: 119000 / 123014: Premsel loss 0.2940, acc 0.8669 (0.8909 / 0.8429) Evaluation 201: Premsel loss 0.2880, acc 0.8705 (0.8984 / 0.8425) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_92-query128-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_45-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_17.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+solo___out1.pkl.gz Training 202: 0 / 97746: Premsel loss 0.2926, acc 0.8665 (0.8923 / 0.8407) Training 202: 7000 / 97746: Premsel loss 0.3109, acc 0.8587 (0.8802 / 0.8373) Training 202: 14000 / 97746: Premsel loss 0.3179, acc 0.8554 (0.8705 / 0.8403) Training 202: 21000 / 97746: Premsel loss 0.3188, acc 0.8541 (0.8654 / 0.8428) Training 202: 28000 / 97746: Premsel loss 0.3189, acc 0.8539 (0.8725 / 0.8354) Training 202: 35000 / 97746: Premsel loss 0.3093, acc 0.8587 (0.8776 / 0.8398) Training 202: 42000 / 97746: Premsel loss 0.3069, acc 0.8612 (0.8805 / 0.8419) Training 202: 49000 / 97746: Premsel loss 0.3067, acc 0.8610 (0.8765 / 0.8455) Training 202: 56000 / 97746: Premsel loss 0.3019, acc 0.8632 (0.8867 / 0.8398) Training 202: 63000 / 97746: Premsel loss 0.3063, acc 0.8612 (0.8843 / 0.8380) Training 202: 70000 / 97746: Premsel loss 0.3093, acc 0.8601 (0.8820 / 0.8381) Training 202: 77000 / 97746: Premsel loss 0.3137, acc 0.8569 (0.8739 / 0.8400) Training 202: 84000 / 97746: Premsel loss 0.3113, acc 0.8587 (0.8777 / 0.8398) Training 202: 91000 / 97746: Premsel loss 0.3129, acc 0.8581 (0.8792 / 0.8371) Evaluation 202: Premsel loss 0.4666, acc 0.7883 (0.8073 / 0.7694) Training 203: 0 / 97746: Premsel loss 0.5139, acc 0.7679 (0.8260 / 0.7097) Training 203: 7000 / 97746: Premsel loss 0.4253, acc 0.8067 (0.8410 / 0.7723) Training 203: 14000 / 97746: Premsel loss 0.4121, acc 0.8125 (0.8482 / 0.7768) Training 203: 21000 / 97746: Premsel loss 0.3852, acc 0.8260 (0.8549 / 0.7972) Training 203: 28000 / 97746: Premsel loss 0.3867, acc 0.8266 (0.8563 / 0.7968) Training 203: 35000 / 97746: Premsel loss 0.3709, acc 0.8318 (0.8588 / 0.8048) Training 203: 42000 / 97746: Premsel loss 0.3783, acc 0.8273 (0.8547 / 0.7999) Training 203: 49000 / 97746: Premsel loss 0.3731, acc 0.8311 (0.8544 / 0.8078) Training 203: 56000 / 97746: Premsel loss 0.3643, acc 0.8368 (0.8607 / 0.8129) Training 203: 63000 / 97746: Premsel loss 0.3577, acc 0.8395 (0.8682 / 0.8107) Training 203: 70000 / 97746: Premsel loss 0.3541, acc 0.8392 (0.8618 / 0.8167) Training 203: 77000 / 97746: Premsel loss 0.3585, acc 0.8372 (0.8664 / 0.8079) Training 203: 84000 / 97746: Premsel loss 0.3635, acc 0.8320 (0.8603 / 0.8037) Training 203: 91000 / 97746: Premsel loss 0.3582, acc 0.8365 (0.8520 / 0.8209) Evaluation 203: Premsel loss 0.3570, acc 0.8382 (0.8529 / 0.8234) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l16000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_09.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l4800-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l8000-e0.15+coop-mzr02___out1.pkl.gz Training 204: 0 / 133331: Premsel loss 0.3525, acc 0.8401 (0.8572 / 0.8230) Training 204: 7000 / 133331: Premsel loss 0.3430, acc 0.8439 (0.8733 / 0.8145) Training 204: 14000 / 133331: Premsel loss 0.3471, acc 0.8421 (0.8742 / 0.8100) Training 204: 21000 / 133331: Premsel loss 0.3457, acc 0.8436 (0.8677 / 0.8196) Training 204: 28000 / 133331: Premsel loss 0.3412, acc 0.8447 (0.8650 / 0.8243) Training 204: 35000 / 133331: Premsel loss 0.3380, acc 0.8462 (0.8723 / 0.8201) Training 204: 42000 / 133331: Premsel loss 0.3430, acc 0.8443 (0.8696 / 0.8190) Training 204: 49000 / 133331: Premsel loss 0.3393, acc 0.8456 (0.8727 / 0.8186) Training 204: 56000 / 133331: Premsel loss 0.3407, acc 0.8458 (0.8715 / 0.8200) Training 204: 63000 / 133331: Premsel loss 0.3445, acc 0.8427 (0.8671 / 0.8184) Training 204: 70000 / 133331: Premsel loss 0.3357, acc 0.8474 (0.8718 / 0.8230) Training 204: 77000 / 133331: Premsel loss 0.3385, acc 0.8470 (0.8753 / 0.8186) Training 204: 84000 / 133331: Premsel loss 0.3397, acc 0.8475 (0.8708 / 0.8241) Training 204: 91000 / 133331: Premsel loss 0.3328, acc 0.8480 (0.8707 / 0.8253) Training 204: 98000 / 133331: Premsel loss 0.3235, acc 0.8537 (0.8798 / 0.8275) Training 204: 105000 / 133331: Premsel loss 0.3320, acc 0.8486 (0.8704 / 0.8268) Training 204: 112000 / 133331: Premsel loss 0.3335, acc 0.8484 (0.8665 / 0.8304) Training 204: 119000 / 133331: Premsel loss 0.3288, acc 0.8512 (0.8755 / 0.8270) Training 204: 126000 / 133331: Premsel loss 0.3246, acc 0.8544 (0.8807 / 0.8280) Training 204: 133000 / 133331: Premsel loss 0.3304, acc 0.8499 (0.8743 / 0.8255) Evaluation 204: Premsel loss 0.3228, acc 0.8539 (0.8713 / 0.8366) Training 205: 0 / 133331: Premsel loss 0.3340, acc 0.8475 (0.8714 / 0.8236) Training 205: 7000 / 133331: Premsel loss 0.3239, acc 0.8531 (0.8745 / 0.8318) Training 205: 14000 / 133331: Premsel loss 0.3251, acc 0.8526 (0.8790 / 0.8262) Training 205: 21000 / 133331: Premsel loss 0.3266, acc 0.8520 (0.8780 / 0.8260) Training 205: 28000 / 133331: Premsel loss 0.3198, acc 0.8564 (0.8777 / 0.8351) Training 205: 35000 / 133331: Premsel loss 0.3312, acc 0.8496 (0.8739 / 0.8254) Training 205: 42000 / 133331: Premsel loss 0.3268, acc 0.8527 (0.8698 / 0.8356) Training 205: 49000 / 133331: Premsel loss 0.3378, acc 0.8455 (0.8809 / 0.8102) Training 205: 56000 / 133331: Premsel loss 0.3252, acc 0.8533 (0.8764 / 0.8301) Training 205: 63000 / 133331: Premsel loss 0.3222, acc 0.8542 (0.8757 / 0.8327) Training 205: 70000 / 133331: Premsel loss 0.3278, acc 0.8517 (0.8762 / 0.8272) Training 205: 77000 / 133331: Premsel loss 0.3171, acc 0.8566 (0.8799 / 0.8334) Training 205: 84000 / 133331: Premsel loss 0.3240, acc 0.8531 (0.8733 / 0.8329) Training 205: 91000 / 133331: Premsel loss 0.3288, acc 0.8506 (0.8750 / 0.8262) Training 205: 98000 / 133331: Premsel loss 0.3239, acc 0.8534 (0.8723 / 0.8346) Training 205: 105000 / 133331: Premsel loss 0.3297, acc 0.8495 (0.8758 / 0.8231) Training 205: 112000 / 133331: Premsel loss 0.3230, acc 0.8550 (0.8743 / 0.8357) Training 205: 119000 / 133331: Premsel loss 0.3229, acc 0.8557 (0.8794 / 0.8320) Training 205: 126000 / 133331: Premsel loss 0.3303, acc 0.8502 (0.8743 / 0.8261) Training 205: 133000 / 133331: Premsel loss 0.3224, acc 0.8559 (0.8743 / 0.8375) Evaluation 205: Premsel loss 0.3199, acc 0.8559 (0.9049 / 0.8069) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_19.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d30-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_68-query128-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l8000-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_22.pkl.gz Training 206: 0 / 126492: Premsel loss 0.3202, acc 0.8574 (0.8867 / 0.8282) Training 206: 7000 / 126492: Premsel loss 0.3341, acc 0.8482 (0.8738 / 0.8227) Training 206: 14000 / 126492: Premsel loss 0.3271, acc 0.8507 (0.8714 / 0.8300) Training 206: 21000 / 126492: Premsel loss 0.3342, acc 0.8471 (0.8712 / 0.8229) Training 206: 28000 / 126492: Premsel loss 0.3338, acc 0.8475 (0.8729 / 0.8221) Training 206: 35000 / 126492: Premsel loss 0.3200, acc 0.8555 (0.8771 / 0.8339) Training 206: 42000 / 126492: Premsel loss 0.3259, acc 0.8520 (0.8702 / 0.8337) Training 206: 49000 / 126492: Premsel loss 0.3262, acc 0.8501 (0.8704 / 0.8299) Training 206: 56000 / 126492: Premsel loss 0.3250, acc 0.8522 (0.8766 / 0.8279) Training 206: 63000 / 126492: Premsel loss 0.3219, acc 0.8553 (0.8763 / 0.8343) Training 206: 70000 / 126492: Premsel loss 0.3231, acc 0.8529 (0.8765 / 0.8293) Training 206: 77000 / 126492: Premsel loss 0.3228, acc 0.8537 (0.8725 / 0.8349) Training 206: 84000 / 126492: Premsel loss 0.3271, acc 0.8512 (0.8811 / 0.8212) Training 206: 91000 / 126492: Premsel loss 0.3227, acc 0.8537 (0.8722 / 0.8353) Training 206: 98000 / 126492: Premsel loss 0.3229, acc 0.8541 (0.8784 / 0.8298) Training 206: 105000 / 126492: Premsel loss 0.3388, acc 0.8451 (0.8729 / 0.8172) Training 206: 112000 / 126492: Premsel loss 0.3206, acc 0.8542 (0.8758 / 0.8326) Training 206: 119000 / 126492: Premsel loss 0.3189, acc 0.8552 (0.8744 / 0.8361) Training 206: 126000 / 126492: Premsel loss 0.3179, acc 0.8551 (0.8798 / 0.8305) Evaluation 206: Premsel loss 0.3241, acc 0.8528 (0.8471 / 0.8585) Training 207: 0 / 126492: Premsel loss 0.3221, acc 0.8541 (0.8804 / 0.8278) Training 207: 7000 / 126492: Premsel loss 0.3237, acc 0.8524 (0.8826 / 0.8222) Training 207: 14000 / 126492: Premsel loss 0.3347, acc 0.8482 (0.8731 / 0.8233) Training 207: 21000 / 126492: Premsel loss 0.3320, acc 0.8474 (0.8727 / 0.8222) Training 207: 28000 / 126492: Premsel loss 0.3208, acc 0.8538 (0.8749 / 0.8326) Training 207: 35000 / 126492: Premsel loss 0.3252, acc 0.8515 (0.8786 / 0.8244) Training 207: 42000 / 126492: Premsel loss 0.3163, acc 0.8570 (0.8787 / 0.8353) Training 207: 49000 / 126492: Premsel loss 0.3236, acc 0.8523 (0.8731 / 0.8314) Training 207: 56000 / 126492: Premsel loss 0.3293, acc 0.8491 (0.8776 / 0.8206) Training 207: 63000 / 126492: Premsel loss 0.3186, acc 0.8547 (0.8777 / 0.8317) Training 207: 70000 / 126492: Premsel loss 0.3231, acc 0.8531 (0.8767 / 0.8295) Training 207: 77000 / 126492: Premsel loss 0.3261, acc 0.8500 (0.8792 / 0.8208) Training 207: 84000 / 126492: Premsel loss 0.3190, acc 0.8549 (0.8790 / 0.8308) Training 207: 91000 / 126492: Premsel loss 0.3176, acc 0.8547 (0.8773 / 0.8321) Training 207: 98000 / 126492: Premsel loss 0.3168, acc 0.8557 (0.8838 / 0.8277) Training 207: 105000 / 126492: Premsel loss 0.3323, acc 0.8471 (0.8691 / 0.8252) Training 207: 112000 / 126492: Premsel loss 0.3126, acc 0.8574 (0.8784 / 0.8363) Training 207: 119000 / 126492: Premsel loss 0.3155, acc 0.8557 (0.8787 / 0.8327) Training 207: 126000 / 126492: Premsel loss 0.3205, acc 0.8542 (0.8732 / 0.8351) Evaluation 207: Premsel loss 0.3769, acc 0.8266 (0.7304 / 0.9228) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+xgb-d12-e0.2+solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_27.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___easypredo2___bb_preds__160___out1.pkl.gz Training 208: 0 / 106998: Premsel loss 0.3364, acc 0.8426 (0.8927 / 0.7926) Training 208: 7000 / 106998: Premsel loss 0.2885, acc 0.8697 (0.8933 / 0.8461) Training 208: 14000 / 106998: Premsel loss 0.2951, acc 0.8679 (0.8874 / 0.8484) Training 208: 21000 / 106998: Premsel loss 0.2888, acc 0.8702 (0.8941 / 0.8463) Training 208: 28000 / 106998: Premsel loss 0.2994, acc 0.8641 (0.8914 / 0.8367) Training 208: 35000 / 106998: Premsel loss 0.2951, acc 0.8670 (0.8894 / 0.8445) Training 208: 42000 / 106998: Premsel loss 0.3042, acc 0.8622 (0.8901 / 0.8342) Training 208: 49000 / 106998: Premsel loss 0.2920, acc 0.8688 (0.8905 / 0.8471) Training 208: 56000 / 106998: Premsel loss 0.2888, acc 0.8707 (0.8937 / 0.8478) Training 208: 63000 / 106998: Premsel loss 0.2806, acc 0.8747 (0.9005 / 0.8489) Training 208: 70000 / 106998: Premsel loss 0.2906, acc 0.8699 (0.8941 / 0.8457) Training 208: 77000 / 106998: Premsel loss 0.2936, acc 0.8680 (0.8900 / 0.8459) Training 208: 84000 / 106998: Premsel loss 0.2849, acc 0.8719 (0.8958 / 0.8480) Training 208: 91000 / 106998: Premsel loss 0.2849, acc 0.8726 (0.8945 / 0.8506) Training 208: 98000 / 106998: Premsel loss 0.2897, acc 0.8688 (0.8954 / 0.8422) Training 208: 105000 / 106998: Premsel loss 0.2904, acc 0.8688 (0.8915 / 0.8460) Evaluation 208: Premsel loss 0.2923, acc 0.8689 (0.8767 / 0.8611) Training 209: 0 / 106998: Premsel loss 0.2951, acc 0.8674 (0.8982 / 0.8367) Training 209: 7000 / 106998: Premsel loss 0.2838, acc 0.8732 (0.8984 / 0.8481) Training 209: 14000 / 106998: Premsel loss 0.2924, acc 0.8676 (0.8929 / 0.8424) Training 209: 21000 / 106998: Premsel loss 0.2972, acc 0.8662 (0.8901 / 0.8424) Training 209: 28000 / 106998: Premsel loss 0.2894, acc 0.8699 (0.8948 / 0.8450) Training 209: 35000 / 106998: Premsel loss 0.2884, acc 0.8723 (0.8959 / 0.8486) Training 209: 42000 / 106998: Premsel loss 0.2858, acc 0.8718 (0.8997 / 0.8439) Training 209: 49000 / 106998: Premsel loss 0.2913, acc 0.8692 (0.8901 / 0.8483) Training 209: 56000 / 106998: Premsel loss 0.2909, acc 0.8684 (0.8873 / 0.8495) Training 209: 63000 / 106998: Premsel loss 0.2926, acc 0.8693 (0.8922 / 0.8465) Training 209: 70000 / 106998: Premsel loss 0.2849, acc 0.8718 (0.8944 / 0.8492) Training 209: 77000 / 106998: Premsel loss 0.2928, acc 0.8681 (0.8889 / 0.8472) Training 209: 84000 / 106998: Premsel loss 0.2880, acc 0.8711 (0.9010 / 0.8411) Training 209: 91000 / 106998: Premsel loss 0.2842, acc 0.8729 (0.8956 / 0.8501) Training 209: 98000 / 106998: Premsel loss 0.2862, acc 0.8715 (0.8941 / 0.8489) Training 209: 105000 / 106998: Premsel loss 0.2980, acc 0.8658 (0.8971 / 0.8345) Evaluation 209: Premsel loss 0.2877, acc 0.8714 (0.8936 / 0.8492) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_42-query512-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_05.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_13-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr___bb_preds__0.05___out1.pkl.gz Training 210: 0 / 100131: Premsel loss 0.2861, acc 0.8714 (0.8947 / 0.8480) Training 210: 7000 / 100131: Premsel loss 0.3203, acc 0.8549 (0.8629 / 0.8470) Training 210: 14000 / 100131: Premsel loss 0.3223, acc 0.8539 (0.8609 / 0.8469) Training 210: 21000 / 100131: Premsel loss 0.3275, acc 0.8511 (0.8562 / 0.8459) Training 210: 28000 / 100131: Premsel loss 0.3209, acc 0.8541 (0.8717 / 0.8365) Training 210: 35000 / 100131: Premsel loss 0.3261, acc 0.8517 (0.8739 / 0.8296) Training 210: 42000 / 100131: Premsel loss 0.3178, acc 0.8556 (0.8657 / 0.8454) Training 210: 49000 / 100131: Premsel loss 0.3189, acc 0.8564 (0.8662 / 0.8466) Training 210: 56000 / 100131: Premsel loss 0.3202, acc 0.8544 (0.8694 / 0.8393) Training 210: 63000 / 100131: Premsel loss 0.3148, acc 0.8580 (0.8678 / 0.8482) Training 210: 70000 / 100131: Premsel loss 0.3218, acc 0.8548 (0.8665 / 0.8431) Training 210: 77000 / 100131: Premsel loss 0.3289, acc 0.8504 (0.8544 / 0.8464) Training 210: 84000 / 100131: Premsel loss 0.3307, acc 0.8501 (0.8543 / 0.8459) Training 210: 91000 / 100131: Premsel loss 0.3163, acc 0.8569 (0.8648 / 0.8489) Training 210: 98000 / 100131: Premsel loss 0.3170, acc 0.8557 (0.8695 / 0.8420) Evaluation 210: Premsel loss 0.3178, acc 0.8560 (0.8697 / 0.8423) Training 211: 0 / 100131: Premsel loss 0.3253, acc 0.8514 (0.8660 / 0.8368) Training 211: 7000 / 100131: Premsel loss 0.3152, acc 0.8550 (0.8633 / 0.8468) Training 211: 14000 / 100131: Premsel loss 0.3187, acc 0.8535 (0.8654 / 0.8417) Training 211: 21000 / 100131: Premsel loss 0.3212, acc 0.8544 (0.8664 / 0.8423) Training 211: 28000 / 100131: Premsel loss 0.3212, acc 0.8556 (0.8643 / 0.8469) Training 211: 35000 / 100131: Premsel loss 0.3269, acc 0.8518 (0.8686 / 0.8350) Training 211: 42000 / 100131: Premsel loss 0.3202, acc 0.8553 (0.8664 / 0.8443) Training 211: 49000 / 100131: Premsel loss 0.3285, acc 0.8505 (0.8630 / 0.8380) Training 211: 56000 / 100131: Premsel loss 0.3377, acc 0.8455 (0.8572 / 0.8338) Training 211: 63000 / 100131: Premsel loss 0.3344, acc 0.8482 (0.8532 / 0.8431) Training 211: 70000 / 100131: Premsel loss 0.3415, acc 0.8446 (0.8566 / 0.8327) Training 211: 77000 / 100131: Premsel loss 0.3235, acc 0.8535 (0.8609 / 0.8462) Training 211: 84000 / 100131: Premsel loss 0.3292, acc 0.8509 (0.8683 / 0.8334) Training 211: 91000 / 100131: Premsel loss 0.3240, acc 0.8538 (0.8674 / 0.8401) Training 211: 98000 / 100131: Premsel loss 0.3298, acc 0.8500 (0.8640 / 0.8361) Evaluation 211: Premsel loss 0.3260, acc 0.8518 (0.8855 / 0.8180) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_49-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_06.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_47-query256-ctx768-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_66-query512-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l4800-e0.15+coop-mzr02___out1.pkl.gz Training 212: 0 / 127234: Premsel loss 0.3255, acc 0.8525 (0.8591 / 0.8458) Training 212: 7000 / 127234: Premsel loss 0.3179, acc 0.8549 (0.8797 / 0.8301) Training 212: 14000 / 127234: Premsel loss 0.3203, acc 0.8546 (0.8743 / 0.8350) Training 212: 21000 / 127234: Premsel loss 0.3145, acc 0.8569 (0.8755 / 0.8383) Training 212: 28000 / 127234: Premsel loss 0.3156, acc 0.8571 (0.8739 / 0.8402) Training 212: 35000 / 127234: Premsel loss 0.3136, acc 0.8592 (0.8736 / 0.8448) Training 212: 42000 / 127234: Premsel loss 0.3153, acc 0.8573 (0.8736 / 0.8410) Training 212: 49000 / 127234: Premsel loss 0.3180, acc 0.8553 (0.8813 / 0.8292) Training 212: 56000 / 127234: Premsel loss 0.3092, acc 0.8601 (0.8809 / 0.8393) Training 212: 63000 / 127234: Premsel loss 0.3093, acc 0.8604 (0.8758 / 0.8449) Training 212: 70000 / 127234: Premsel loss 0.3095, acc 0.8589 (0.8725 / 0.8454) Training 212: 77000 / 127234: Premsel loss 0.3138, acc 0.8579 (0.8773 / 0.8384) Training 212: 84000 / 127234: Premsel loss 0.3379, acc 0.8438 (0.8638 / 0.8238) Training 212: 91000 / 127234: Premsel loss 0.3322, acc 0.8473 (0.8631 / 0.8315) Training 212: 98000 / 127234: Premsel loss 0.3298, acc 0.8476 (0.8639 / 0.8313) Training 212: 105000 / 127234: Premsel loss 0.3217, acc 0.8532 (0.8718 / 0.8346) Training 212: 112000 / 127234: Premsel loss 0.3176, acc 0.8550 (0.8672 / 0.8428) Training 212: 119000 / 127234: Premsel loss 0.3224, acc 0.8539 (0.8729 / 0.8349) Training 212: 126000 / 127234: Premsel loss 0.3102, acc 0.8600 (0.8799 / 0.8402) Evaluation 212: Premsel loss 0.3124, acc 0.8577 (0.8623 / 0.8531) Training 213: 0 / 127234: Premsel loss 0.3179, acc 0.8545 (0.8718 / 0.8372) Training 213: 7000 / 127234: Premsel loss 0.3116, acc 0.8585 (0.8782 / 0.8387) Training 213: 14000 / 127234: Premsel loss 0.3214, acc 0.8517 (0.8727 / 0.8307) Training 213: 21000 / 127234: Premsel loss 0.3103, acc 0.8593 (0.8739 / 0.8448) Training 213: 28000 / 127234: Premsel loss 0.3130, acc 0.8581 (0.8805 / 0.8357) Training 213: 35000 / 127234: Premsel loss 0.3220, acc 0.8534 (0.8709 / 0.8359) Training 213: 42000 / 127234: Premsel loss 0.3111, acc 0.8592 (0.8786 / 0.8399) Training 213: 49000 / 127234: Premsel loss 0.3129, acc 0.8574 (0.8818 / 0.8329) Training 213: 56000 / 127234: Premsel loss 0.3333, acc 0.8464 (0.8712 / 0.8215) Training 213: 63000 / 127234: Premsel loss 0.3206, acc 0.8544 (0.8739 / 0.8348) Training 213: 70000 / 127234: Premsel loss 0.3157, acc 0.8557 (0.8757 / 0.8357) Training 213: 77000 / 127234: Premsel loss 0.3229, acc 0.8528 (0.8700 / 0.8356) Training 213: 84000 / 127234: Premsel loss 0.3118, acc 0.8579 (0.8779 / 0.8379) Training 213: 91000 / 127234: Premsel loss 0.3117, acc 0.8583 (0.8781 / 0.8385) Training 213: 98000 / 127234: Premsel loss 0.3160, acc 0.8572 (0.8801 / 0.8342) Training 213: 105000 / 127234: Premsel loss 0.3182, acc 0.8547 (0.8659 / 0.8436) Training 213: 112000 / 127234: Premsel loss 0.3099, acc 0.8596 (0.8832 / 0.8359) Training 213: 119000 / 127234: Premsel loss 0.3117, acc 0.8576 (0.8796 / 0.8356) Training 213: 126000 / 127234: Premsel loss 0.3238, acc 0.8514 (0.8802 / 0.8225) Evaluation 213: Premsel loss 0.3127, acc 0.8577 (0.9093 / 0.8061) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_10.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_50-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_23-query256-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l6400-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_07.pkl.gz Training 214: 0 / 122842: Premsel loss 0.3128, acc 0.8561 (0.8778 / 0.8344) Training 214: 7000 / 122842: Premsel loss 0.3176, acc 0.8554 (0.8819 / 0.8289) Training 214: 14000 / 122842: Premsel loss 0.3201, acc 0.8524 (0.8839 / 0.8210) Training 214: 21000 / 122842: Premsel loss 0.3143, acc 0.8570 (0.8815 / 0.8324) Training 214: 28000 / 122842: Premsel loss 0.3146, acc 0.8580 (0.8895 / 0.8264) Training 214: 35000 / 122842: Premsel loss 0.3104, acc 0.8579 (0.8865 / 0.8293) Training 214: 42000 / 122842: Premsel loss 0.3137, acc 0.8563 (0.8878 / 0.8248) Training 214: 49000 / 122842: Premsel loss 0.3123, acc 0.8573 (0.8839 / 0.8308) Training 214: 56000 / 122842: Premsel loss 0.3107, acc 0.8590 (0.8860 / 0.8320) Training 214: 63000 / 122842: Premsel loss 0.3066, acc 0.8615 (0.8875 / 0.8355) Training 214: 70000 / 122842: Premsel loss 0.3088, acc 0.8588 (0.8854 / 0.8323) Training 214: 77000 / 122842: Premsel loss 0.3159, acc 0.8555 (0.8851 / 0.8258) Training 214: 84000 / 122842: Premsel loss 0.3087, acc 0.8599 (0.8836 / 0.8361) Training 214: 91000 / 122842: Premsel loss 0.3039, acc 0.8621 (0.8913 / 0.8329) Training 214: 98000 / 122842: Premsel loss 0.3076, acc 0.8607 (0.8844 / 0.8369) Training 214: 105000 / 122842: Premsel loss 0.3039, acc 0.8616 (0.8863 / 0.8369) Training 214: 112000 / 122842: Premsel loss 0.3112, acc 0.8582 (0.8818 / 0.8346) Training 214: 119000 / 122842: Premsel loss 0.3078, acc 0.8605 (0.8900 / 0.8309) Evaluation 214: Premsel loss 0.3120, acc 0.8581 (0.8510 / 0.8651) Training 215: 0 / 122842: Premsel loss 0.3117, acc 0.8594 (0.8893 / 0.8295) Training 215: 7000 / 122842: Premsel loss 0.3133, acc 0.8571 (0.8870 / 0.8272) Training 215: 14000 / 122842: Premsel loss 0.3076, acc 0.8611 (0.8858 / 0.8365) Training 215: 21000 / 122842: Premsel loss 0.3118, acc 0.8587 (0.8854 / 0.8321) Training 215: 28000 / 122842: Premsel loss 0.3075, acc 0.8610 (0.8867 / 0.8353) Training 215: 35000 / 122842: Premsel loss 0.3021, acc 0.8632 (0.8884 / 0.8380) Training 215: 42000 / 122842: Premsel loss 0.3111, acc 0.8591 (0.8858 / 0.8325) Training 215: 49000 / 122842: Premsel loss 0.3126, acc 0.8578 (0.8878 / 0.8278) Training 215: 56000 / 122842: Premsel loss 0.3235, acc 0.8519 (0.8776 / 0.8262) Training 215: 63000 / 122842: Premsel loss 0.3261, acc 0.8505 (0.8744 / 0.8266) Training 215: 70000 / 122842: Premsel loss 0.3205, acc 0.8526 (0.8850 / 0.8203) Training 215: 77000 / 122842: Premsel loss 0.3155, acc 0.8565 (0.8846 / 0.8285) Training 215: 84000 / 122842: Premsel loss 0.3192, acc 0.8547 (0.8812 / 0.8283) Training 215: 91000 / 122842: Premsel loss 0.3218, acc 0.8533 (0.8869 / 0.8196) Training 215: 98000 / 122842: Premsel loss 0.3190, acc 0.8546 (0.8838 / 0.8255) Training 215: 105000 / 122842: Premsel loss 0.3352, acc 0.8465 (0.8732 / 0.8197) Training 215: 112000 / 122842: Premsel loss 0.3152, acc 0.8568 (0.8840 / 0.8296) Training 215: 119000 / 122842: Premsel loss 0.3225, acc 0.8524 (0.8800 / 0.8248) Evaluation 215: Premsel loss 0.3230, acc 0.8519 (0.8775 / 0.8263) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo2_tr_min___bb_preds__160___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_50-query512-ctx1536-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_43-query512-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_13.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.05___out1.pkl.gz Training 216: 0 / 74864: Premsel loss 0.3247, acc 0.8508 (0.8753 / 0.8263) Training 216: 7000 / 74864: Premsel loss 0.3380, acc 0.8446 (0.8623 / 0.8269) Training 216: 14000 / 74864: Premsel loss 0.3421, acc 0.8431 (0.8522 / 0.8340) Training 216: 21000 / 74864: Premsel loss 0.3298, acc 0.8495 (0.8672 / 0.8318) Training 216: 28000 / 74864: Premsel loss 0.3490, acc 0.8405 (0.8503 / 0.8307) Training 216: 35000 / 74864: Premsel loss 0.3391, acc 0.8447 (0.8613 / 0.8280) Training 216: 42000 / 74864: Premsel loss 0.3261, acc 0.8506 (0.8681 / 0.8331) Training 216: 49000 / 74864: Premsel loss 0.3232, acc 0.8510 (0.8590 / 0.8430) Training 216: 56000 / 74864: Premsel loss 0.3238, acc 0.8519 (0.8670 / 0.8368) Training 216: 63000 / 74864: Premsel loss 0.3289, acc 0.8497 (0.8670 / 0.8325) Training 216: 70000 / 74864: Premsel loss 0.3324, acc 0.8485 (0.8646 / 0.8323) Evaluation 216: Premsel loss 0.3299, acc 0.8494 (0.8547 / 0.8441) Training 217: 0 / 74864: Premsel loss 0.3322, acc 0.8472 (0.8606 / 0.8339) Training 217: 7000 / 74864: Premsel loss 0.3227, acc 0.8531 (0.8684 / 0.8378) Training 217: 14000 / 74864: Premsel loss 0.3252, acc 0.8519 (0.8649 / 0.8389) Training 217: 21000 / 74864: Premsel loss 0.3250, acc 0.8523 (0.8677 / 0.8368) Training 217: 28000 / 74864: Premsel loss 0.3328, acc 0.8479 (0.8713 / 0.8246) Training 217: 35000 / 74864: Premsel loss 0.3362, acc 0.8466 (0.8526 / 0.8405) Training 217: 42000 / 74864: Premsel loss 0.3262, acc 0.8512 (0.8673 / 0.8352) Training 217: 49000 / 74864: Premsel loss 0.3239, acc 0.8522 (0.8669 / 0.8374) Training 217: 56000 / 74864: Premsel loss 0.3177, acc 0.8560 (0.8708 / 0.8411) Training 217: 63000 / 74864: Premsel loss 0.3197, acc 0.8542 (0.8662 / 0.8422) Training 217: 70000 / 74864: Premsel loss 0.3255, acc 0.8524 (0.8725 / 0.8322) Evaluation 217: Premsel loss 0.3218, acc 0.8540 (0.8565 / 0.8515) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__0.005___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l8-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_59-query768-ctx1024-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_25.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training.min1___bb_min_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_26-query128-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_20-query128-ctx512-w0-solo___out1.pkl.gz Training 218: 0 / 91475: Premsel loss 0.3175, acc 0.8550 (0.8700 / 0.8401) Training 218: 7000 / 91475: Premsel loss 0.3144, acc 0.8574 (0.8747 / 0.8400) Training 218: 14000 / 91475: Premsel loss 0.3185, acc 0.8551 (0.8731 / 0.8371) Training 218: 21000 / 91475: Premsel loss 0.3211, acc 0.8546 (0.8714 / 0.8379) Training 218: 28000 / 91475: Premsel loss 0.3210, acc 0.8539 (0.8676 / 0.8403) Training 218: 35000 / 91475: Premsel loss 0.3257, acc 0.8505 (0.8670 / 0.8339) Training 218: 42000 / 91475: Premsel loss 0.3124, acc 0.8573 (0.8729 / 0.8418) Training 218: 49000 / 91475: Premsel loss 0.3108, acc 0.8591 (0.8816 / 0.8366) Training 218: 56000 / 91475: Premsel loss 0.3138, acc 0.8582 (0.8720 / 0.8443) Training 218: 63000 / 91475: Premsel loss 0.3098, acc 0.8593 (0.8776 / 0.8410) Training 218: 70000 / 91475: Premsel loss 0.3217, acc 0.8536 (0.8779 / 0.8293) Training 218: 77000 / 91475: Premsel loss 0.3169, acc 0.8557 (0.8733 / 0.8381) Training 218: 84000 / 91475: Premsel loss 0.3199, acc 0.8544 (0.8733 / 0.8354) Training 218: 91000 / 91475: Premsel loss 0.3084, acc 0.8609 (0.8773 / 0.8444) Evaluation 218: Premsel loss 0.3192, acc 0.8538 (0.8959 / 0.8117) Training 219: 0 / 91475: Premsel loss 0.3139, acc 0.8573 (0.8713 / 0.8434) Training 219: 7000 / 91475: Premsel loss 0.3221, acc 0.8541 (0.8803 / 0.8278) Training 219: 14000 / 91475: Premsel loss 0.3113, acc 0.8579 (0.8731 / 0.8426) Training 219: 21000 / 91475: Premsel loss 0.3149, acc 0.8564 (0.8755 / 0.8373) Training 219: 28000 / 91475: Premsel loss 0.3045, acc 0.8625 (0.8795 / 0.8456) Training 219: 35000 / 91475: Premsel loss 0.3015, acc 0.8639 (0.8789 / 0.8489) Training 219: 42000 / 91475: Premsel loss 0.3033, acc 0.8625 (0.8771 / 0.8479) Training 219: 49000 / 91475: Premsel loss 0.3062, acc 0.8610 (0.8741 / 0.8479) Training 219: 56000 / 91475: Premsel loss 0.3152, acc 0.8569 (0.8729 / 0.8410) Training 219: 63000 / 91475: Premsel loss 0.3200, acc 0.8537 (0.8642 / 0.8433) Training 219: 70000 / 91475: Premsel loss 0.3119, acc 0.8571 (0.8730 / 0.8412) Training 219: 77000 / 91475: Premsel loss 0.3150, acc 0.8558 (0.8669 / 0.8448) Training 219: 84000 / 91475: Premsel loss 0.3177, acc 0.8555 (0.8802 / 0.8309) Training 219: 91000 / 91475: Premsel loss 0.3220, acc 0.8545 (0.8672 / 0.8417) Evaluation 219: Premsel loss 0.3264, acc 0.8505 (0.8858 / 0.8151) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aamzr02-T10_loop01_10-query128-ctx512-w0-solo___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_23.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d60-l3600-e0.15+coop-mzr02___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop01-VHSLCAXPh+lgb-d50-l900-e0.15loop01+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___mizar40-all-T30___bb_Enigma+mizar40-all-T10+mega-VHSLCAXPh2e15+lgb-t150-d50-l900-e0.15+coop-mzr02___out1.pkl.gz Training 220: 0 / 120729: Premsel loss 0.3213, acc 0.8540 (0.8668 / 0.8411) Training 220: 7000 / 120729: Premsel loss 0.3159, acc 0.8561 (0.8873 / 0.8249) Training 220: 14000 / 120729: Premsel loss 0.3189, acc 0.8554 (0.8834 / 0.8273) Training 220: 21000 / 120729: Premsel loss 0.3147, acc 0.8554 (0.8774 / 0.8334) Training 220: 28000 / 120729: Premsel loss 0.3202, acc 0.8534 (0.8688 / 0.8379) Training 220: 35000 / 120729: Premsel loss 0.3202, acc 0.8526 (0.8717 / 0.8334) Training 220: 42000 / 120729: Premsel loss 0.3047, acc 0.8614 (0.8842 / 0.8386) Training 220: 49000 / 120729: Premsel loss 0.3170, acc 0.8559 (0.8771 / 0.8347) Training 220: 56000 / 120729: Premsel loss 0.3317, acc 0.8468 (0.8723 / 0.8214) Training 220: 63000 / 120729: Premsel loss 0.3297, acc 0.8471 (0.8744 / 0.8198) Training 220: 70000 / 120729: Premsel loss 0.3282, acc 0.8502 (0.8700 / 0.8304) Training 220: 77000 / 120729: Premsel loss 0.3280, acc 0.8500 (0.8761 / 0.8240) Training 220: 84000 / 120729: Premsel loss 0.3187, acc 0.8558 (0.8787 / 0.8328) Training 220: 91000 / 120729: Premsel loss 0.3203, acc 0.8534 (0.8769 / 0.8299) Training 220: 98000 / 120729: Premsel loss 0.3248, acc 0.8514 (0.8802 / 0.8227) Training 220: 105000 / 120729: Premsel loss 0.3185, acc 0.8552 (0.8774 / 0.8331) Training 220: 112000 / 120729: Premsel loss 0.3174, acc 0.8550 (0.8863 / 0.8237) Training 220: 119000 / 120729: Premsel loss 0.3242, acc 0.8512 (0.8711 / 0.8314) Evaluation 220: Premsel loss 0.3170, acc 0.8551 (0.8579 / 0.8523) Training 221: 0 / 120729: Premsel loss 0.3250, acc 0.8513 (0.8793 / 0.8233) Training 221: 7000 / 120729: Premsel loss 0.3184, acc 0.8542 (0.8747 / 0.8337) Training 221: 14000 / 120729: Premsel loss 0.3214, acc 0.8524 (0.8811 / 0.8237) Training 221: 21000 / 120729: Premsel loss 0.3179, acc 0.8549 (0.8815 / 0.8282) Training 221: 28000 / 120729: Premsel loss 0.3220, acc 0.8519 (0.8754 / 0.8285) Training 221: 35000 / 120729: Premsel loss 0.3311, acc 0.8461 (0.8660 / 0.8262) Training 221: 42000 / 120729: Premsel loss 0.3339, acc 0.8448 (0.8714 / 0.8182) Training 221: 49000 / 120729: Premsel loss 0.3240, acc 0.8513 (0.8794 / 0.8232) Training 221: 56000 / 120729: Premsel loss 0.3232, acc 0.8521 (0.8749 / 0.8293) Training 221: 63000 / 120729: Premsel loss 0.3198, acc 0.8545 (0.8742 / 0.8347) Training 221: 70000 / 120729: Premsel loss 0.3131, acc 0.8558 (0.8828 / 0.8287) Training 221: 77000 / 120729: Premsel loss 0.3137, acc 0.8565 (0.8791 / 0.8338) Training 221: 84000 / 120729: Premsel loss 0.3157, acc 0.8560 (0.8832 / 0.8287) Training 221: 91000 / 120729: Premsel loss 0.3166, acc 0.8558 (0.8800 / 0.8315) Training 221: 98000 / 120729: Premsel loss 0.3242, acc 0.8524 (0.8793 / 0.8255) Training 221: 105000 / 120729: Premsel loss 0.3248, acc 0.8500 (0.8790 / 0.8209) Training 221: 112000 / 120729: Premsel loss 0.3226, acc 0.8521 (0.8758 / 0.8284) Training 221: 119000 / 120729: Premsel loss 0.3215, acc 0.8520 (0.8795 / 0.8245) Evaluation 221: Premsel loss 0.3180, acc 0.8548 (0.9027 / 0.8069) Loading data... /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_20.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___grid1500_greed_all.training___bb_l5-mzr02-premsel_enigma_01_2020_T10_loop01_epoch_1-query256-ctx512-w0-coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___aaEnigma+mizar40-all-T10+loop02-VHSLCAXPh+lgb-d30-l1800-e0.15+coop___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/enigma___bhardpredo1_tr_min___bb_preds__192___out1.pkl.gz /home/urbanjo3/ec/mirek/enigma/pickles/addedrare_08.pkl.gz Training 222: 0 / 93378: Premsel loss 0.3167, acc 0.8566 (0.8877 / 0.8255) Training 222: 7000 / 93378: Premsel loss 0.3272, acc 0.8502 (0.8766 / 0.8239) Training 222: 14000 / 93378: Premsel loss 0.3284, acc 0.8488 (0.8732 / 0.8245) Training 222: 21000 / 93378: Premsel loss 0.3234, acc 0.8516 (0.8817 / 0.8215) Training 222: 28000 / 93378: Premsel loss 0.3180, acc 0.8559 (0.8866 / 0.8252) Training 222: 35000 / 93378: Premsel loss 0.3183, acc 0.8544 (0.8806 / 0.8282) Training 222: 42000 / 93378: Premsel loss 0.3264, acc 0.8502 (0.8828 / 0.8177) Training 222: 49000 / 93378: Premsel loss 0.3222, acc 0.8539 (0.8769 / 0.8308) Training 222: 56000 / 93378: Premsel loss 0.3143, acc 0.8570 (0.8815 / 0.8326) Training 222: 63000 / 93378: Premsel loss 0.3209, acc 0.8531 (0.8837 / 0.8224) Training 222: 70000 / 93378: Premsel loss 0.3190, acc 0.8539 (0.8822 / 0.8256) Training 222: 77000 / 93378: Premsel loss 0.3171, acc 0.8562 (0.8855 / 0.8269) Training 222: 84000 / 93378: Premsel loss 0.3216, acc 0.8536 (0.8827 / 0.8246) Training 222: 91000 / 93378: Premsel loss 0.3114, acc 0.8586 (0.8889 / 0.8283) ^C^C^C^C^Z [1]+ Stopped python premsel_network_enigma-01-2020-my1-no_symbols_tst3_10.py real 4981m37.806s user 0m0.001s sys 0m0.000s