{ "best_metric": 0.8695652173913043, "best_model_checkpoint": "CTMAE-P2-V4-S1/checkpoint-2489", "epoch": 49.01246153846154, "eval_steps": 500, "global_step": 6500, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0015384615384615385, "grad_norm": 6.0519328117370605, "learning_rate": 1.5384615384615387e-07, "loss": 0.7005, "step": 10 }, { "epoch": 0.003076923076923077, "grad_norm": 1.9966832399368286, "learning_rate": 3.0769230769230774e-07, "loss": 0.6944, "step": 20 }, { "epoch": 0.004615384615384616, "grad_norm": 2.0280706882476807, "learning_rate": 4.615384615384616e-07, "loss": 0.6865, "step": 30 }, { "epoch": 0.006153846153846154, "grad_norm": 3.468716621398926, "learning_rate": 6.153846153846155e-07, "loss": 0.677, "step": 40 }, { "epoch": 0.007692307692307693, "grad_norm": 6.667259693145752, "learning_rate": 7.692307692307694e-07, "loss": 0.6879, "step": 50 }, { "epoch": 0.009230769230769232, "grad_norm": 2.7107460498809814, "learning_rate": 9.230769230769232e-07, "loss": 0.6369, "step": 60 }, { "epoch": 0.010769230769230769, "grad_norm": 5.00808048248291, "learning_rate": 1.076923076923077e-06, "loss": 0.6337, "step": 70 }, { "epoch": 0.012307692307692308, "grad_norm": 7.976367950439453, "learning_rate": 1.230769230769231e-06, "loss": 0.5953, "step": 80 }, { "epoch": 0.013846153846153847, "grad_norm": 5.392799377441406, "learning_rate": 1.3846153846153848e-06, "loss": 0.5653, "step": 90 }, { "epoch": 0.015384615384615385, "grad_norm": 5.476315498352051, "learning_rate": 1.5384615384615387e-06, "loss": 0.6058, "step": 100 }, { "epoch": 0.016923076923076923, "grad_norm": 6.422147750854492, "learning_rate": 1.6923076923076926e-06, "loss": 0.6694, "step": 110 }, { "epoch": 0.018461538461538463, "grad_norm": 4.2546281814575195, "learning_rate": 1.8461538461538465e-06, "loss": 0.3622, "step": 120 }, { "epoch": 0.02, "grad_norm": 6.692765235900879, "learning_rate": 2.0000000000000003e-06, "loss": 0.6432, "step": 130 }, { "epoch": 0.020153846153846154, "eval_accuracy": 0.5434782608695652, "eval_loss": 0.8218211531639099, "eval_runtime": 15.023, "eval_samples_per_second": 3.062, "eval_steps_per_second": 1.531, "step": 131 }, { "epoch": 1.0013846153846153, "grad_norm": 6.260814189910889, "learning_rate": 2.153846153846154e-06, "loss": 0.4914, "step": 140 }, { "epoch": 1.002923076923077, "grad_norm": 6.606039524078369, "learning_rate": 2.307692307692308e-06, "loss": 0.5981, "step": 150 }, { "epoch": 1.0044615384615385, "grad_norm": 5.284712314605713, "learning_rate": 2.461538461538462e-06, "loss": 0.9034, "step": 160 }, { "epoch": 1.006, "grad_norm": 3.669586658477783, "learning_rate": 2.615384615384616e-06, "loss": 0.6945, "step": 170 }, { "epoch": 1.0075384615384615, "grad_norm": 7.1405510902404785, "learning_rate": 2.7692307692307697e-06, "loss": 0.6841, "step": 180 }, { "epoch": 1.009076923076923, "grad_norm": 16.444698333740234, "learning_rate": 2.9230769230769236e-06, "loss": 0.6695, "step": 190 }, { "epoch": 1.0106153846153847, "grad_norm": 3.4212985038757324, "learning_rate": 3.0769230769230774e-06, "loss": 0.7139, "step": 200 }, { "epoch": 1.0121538461538462, "grad_norm": 2.3612794876098633, "learning_rate": 3.2307692307692313e-06, "loss": 0.6926, "step": 210 }, { "epoch": 1.0136923076923077, "grad_norm": 25.119522094726562, "learning_rate": 3.384615384615385e-06, "loss": 0.868, "step": 220 }, { "epoch": 1.0152307692307692, "grad_norm": 19.717683792114258, "learning_rate": 3.538461538461539e-06, "loss": 0.3797, "step": 230 }, { "epoch": 1.0167692307692309, "grad_norm": 20.694677352905273, "learning_rate": 3.692307692307693e-06, "loss": 1.3772, "step": 240 }, { "epoch": 1.0183076923076924, "grad_norm": 1.1371208429336548, "learning_rate": 3.846153846153847e-06, "loss": 0.9887, "step": 250 }, { "epoch": 1.0198461538461538, "grad_norm": 0.3055202066898346, "learning_rate": 4.000000000000001e-06, "loss": 0.4243, "step": 260 }, { "epoch": 1.0201538461538462, "eval_accuracy": 0.5434782608695652, "eval_loss": 2.131279706954956, "eval_runtime": 13.9504, "eval_samples_per_second": 3.297, "eval_steps_per_second": 1.649, "step": 262 }, { "epoch": 2.0012307692307694, "grad_norm": 19.514455795288086, "learning_rate": 4.1538461538461545e-06, "loss": 1.8913, "step": 270 }, { "epoch": 2.0027692307692306, "grad_norm": 1.755078673362732, "learning_rate": 4.307692307692308e-06, "loss": 0.9929, "step": 280 }, { "epoch": 2.0043076923076923, "grad_norm": 16.61043930053711, "learning_rate": 4.461538461538462e-06, "loss": 1.1395, "step": 290 }, { "epoch": 2.005846153846154, "grad_norm": 7.139761447906494, "learning_rate": 4.615384615384616e-06, "loss": 0.6417, "step": 300 }, { "epoch": 2.0073846153846153, "grad_norm": 7.5372395515441895, "learning_rate": 4.76923076923077e-06, "loss": 0.5172, "step": 310 }, { "epoch": 2.008923076923077, "grad_norm": 13.364325523376465, "learning_rate": 4.923076923076924e-06, "loss": 0.571, "step": 320 }, { "epoch": 2.0104615384615383, "grad_norm": 16.03362464904785, "learning_rate": 5.076923076923077e-06, "loss": 0.7023, "step": 330 }, { "epoch": 2.012, "grad_norm": 1.2387570142745972, "learning_rate": 5.230769230769232e-06, "loss": 1.0144, "step": 340 }, { "epoch": 2.0135384615384617, "grad_norm": 0.9126087427139282, "learning_rate": 5.384615384615385e-06, "loss": 0.7244, "step": 350 }, { "epoch": 2.015076923076923, "grad_norm": 15.709758758544922, "learning_rate": 5.538461538461539e-06, "loss": 1.1351, "step": 360 }, { "epoch": 2.0166153846153847, "grad_norm": 1.8802133798599243, "learning_rate": 5.692307692307692e-06, "loss": 1.0503, "step": 370 }, { "epoch": 2.018153846153846, "grad_norm": 0.583645224571228, "learning_rate": 5.846153846153847e-06, "loss": 0.4027, "step": 380 }, { "epoch": 2.0196923076923077, "grad_norm": 30.528959274291992, "learning_rate": 6e-06, "loss": 1.0867, "step": 390 }, { "epoch": 2.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.9686189889907837, "eval_runtime": 13.9707, "eval_samples_per_second": 3.293, "eval_steps_per_second": 1.646, "step": 393 }, { "epoch": 3.001076923076923, "grad_norm": 0.5618914365768433, "learning_rate": 6.153846153846155e-06, "loss": 1.1478, "step": 400 }, { "epoch": 3.0026153846153845, "grad_norm": 14.351670265197754, "learning_rate": 6.307692307692308e-06, "loss": 0.6091, "step": 410 }, { "epoch": 3.004153846153846, "grad_norm": 1.4651223421096802, "learning_rate": 6.461538461538463e-06, "loss": 0.9063, "step": 420 }, { "epoch": 3.005692307692308, "grad_norm": 1.2253414392471313, "learning_rate": 6.615384615384616e-06, "loss": 1.2102, "step": 430 }, { "epoch": 3.007230769230769, "grad_norm": 1.232196569442749, "learning_rate": 6.76923076923077e-06, "loss": 0.5107, "step": 440 }, { "epoch": 3.008769230769231, "grad_norm": 15.324777603149414, "learning_rate": 6.923076923076923e-06, "loss": 1.1337, "step": 450 }, { "epoch": 3.010307692307692, "grad_norm": 42.23984909057617, "learning_rate": 7.076923076923078e-06, "loss": 1.1615, "step": 460 }, { "epoch": 3.011846153846154, "grad_norm": 18.432376861572266, "learning_rate": 7.230769230769231e-06, "loss": 0.691, "step": 470 }, { "epoch": 3.0133846153846156, "grad_norm": 0.8665348887443542, "learning_rate": 7.384615384615386e-06, "loss": 0.8319, "step": 480 }, { "epoch": 3.014923076923077, "grad_norm": 12.59626293182373, "learning_rate": 7.538461538461539e-06, "loss": 1.3428, "step": 490 }, { "epoch": 3.0164615384615385, "grad_norm": 1.335548996925354, "learning_rate": 7.692307692307694e-06, "loss": 0.6631, "step": 500 }, { "epoch": 3.018, "grad_norm": 12.23017692565918, "learning_rate": 7.846153846153847e-06, "loss": 1.4626, "step": 510 }, { "epoch": 3.0195384615384615, "grad_norm": 4.052307605743408, "learning_rate": 8.000000000000001e-06, "loss": 0.7155, "step": 520 }, { "epoch": 3.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 0.8454405665397644, "eval_runtime": 13.1526, "eval_samples_per_second": 3.497, "eval_steps_per_second": 1.749, "step": 524 }, { "epoch": 4.000923076923077, "grad_norm": 2.5005366802215576, "learning_rate": 8.153846153846154e-06, "loss": 0.3124, "step": 530 }, { "epoch": 4.002461538461539, "grad_norm": 17.99958610534668, "learning_rate": 8.307692307692309e-06, "loss": 0.88, "step": 540 }, { "epoch": 4.004, "grad_norm": 21.894472122192383, "learning_rate": 8.461538461538462e-06, "loss": 0.7552, "step": 550 }, { "epoch": 4.005538461538461, "grad_norm": 4.645376205444336, "learning_rate": 8.615384615384617e-06, "loss": 1.2199, "step": 560 }, { "epoch": 4.007076923076923, "grad_norm": 4.8306145668029785, "learning_rate": 8.76923076923077e-06, "loss": 0.6159, "step": 570 }, { "epoch": 4.008615384615385, "grad_norm": 7.024281978607178, "learning_rate": 8.923076923076925e-06, "loss": 0.5809, "step": 580 }, { "epoch": 4.010153846153846, "grad_norm": 1.274326205253601, "learning_rate": 9.076923076923078e-06, "loss": 0.9819, "step": 590 }, { "epoch": 4.011692307692308, "grad_norm": 11.132002830505371, "learning_rate": 9.230769230769232e-06, "loss": 0.7859, "step": 600 }, { "epoch": 4.013230769230769, "grad_norm": 12.536384582519531, "learning_rate": 9.384615384615385e-06, "loss": 0.6264, "step": 610 }, { "epoch": 4.014769230769231, "grad_norm": 12.76213550567627, "learning_rate": 9.53846153846154e-06, "loss": 1.412, "step": 620 }, { "epoch": 4.016307692307692, "grad_norm": 3.126918315887451, "learning_rate": 9.692307692307693e-06, "loss": 0.6571, "step": 630 }, { "epoch": 4.017846153846154, "grad_norm": 11.6869478225708, "learning_rate": 9.846153846153848e-06, "loss": 0.3807, "step": 640 }, { "epoch": 4.019384615384616, "grad_norm": 0.48865261673927307, "learning_rate": 1e-05, "loss": 1.2743, "step": 650 }, { "epoch": 4.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.4546117782592773, "eval_runtime": 13.1792, "eval_samples_per_second": 3.49, "eval_steps_per_second": 1.745, "step": 655 }, { "epoch": 5.00076923076923, "grad_norm": 0.90949547290802, "learning_rate": 9.982905982905984e-06, "loss": 0.726, "step": 660 }, { "epoch": 5.002307692307692, "grad_norm": 0.4137256145477295, "learning_rate": 9.965811965811966e-06, "loss": 0.9188, "step": 670 }, { "epoch": 5.003846153846154, "grad_norm": 9.0480318069458, "learning_rate": 9.94871794871795e-06, "loss": 1.2085, "step": 680 }, { "epoch": 5.0053846153846155, "grad_norm": 16.014402389526367, "learning_rate": 9.931623931623933e-06, "loss": 0.7161, "step": 690 }, { "epoch": 5.006923076923077, "grad_norm": 1.0252034664154053, "learning_rate": 9.914529914529915e-06, "loss": 0.9142, "step": 700 }, { "epoch": 5.008461538461538, "grad_norm": 1.1601945161819458, "learning_rate": 9.897435897435899e-06, "loss": 0.7552, "step": 710 }, { "epoch": 5.01, "grad_norm": 8.401622772216797, "learning_rate": 9.880341880341882e-06, "loss": 0.8651, "step": 720 }, { "epoch": 5.0115384615384615, "grad_norm": 1.329608678817749, "learning_rate": 9.863247863247864e-06, "loss": 0.7054, "step": 730 }, { "epoch": 5.013076923076923, "grad_norm": 12.047654151916504, "learning_rate": 9.846153846153848e-06, "loss": 0.7337, "step": 740 }, { "epoch": 5.014615384615385, "grad_norm": 8.120343208312988, "learning_rate": 9.82905982905983e-06, "loss": 0.818, "step": 750 }, { "epoch": 5.016153846153846, "grad_norm": 1.3821797370910645, "learning_rate": 9.811965811965812e-06, "loss": 0.8734, "step": 760 }, { "epoch": 5.0176923076923075, "grad_norm": 0.37048640847206116, "learning_rate": 9.794871794871795e-06, "loss": 0.7733, "step": 770 }, { "epoch": 5.019230769230769, "grad_norm": 0.5471999645233154, "learning_rate": 9.777777777777779e-06, "loss": 0.7952, "step": 780 }, { "epoch": 5.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.4257819652557373, "eval_runtime": 13.1427, "eval_samples_per_second": 3.5, "eval_steps_per_second": 1.75, "step": 786 }, { "epoch": 6.000615384615385, "grad_norm": 8.841777801513672, "learning_rate": 9.76068376068376e-06, "loss": 0.8073, "step": 790 }, { "epoch": 6.002153846153846, "grad_norm": 0.8512263298034668, "learning_rate": 9.743589743589744e-06, "loss": 0.954, "step": 800 }, { "epoch": 6.003692307692308, "grad_norm": 0.698432207107544, "learning_rate": 9.726495726495728e-06, "loss": 0.918, "step": 810 }, { "epoch": 6.005230769230769, "grad_norm": 1.1482590436935425, "learning_rate": 9.70940170940171e-06, "loss": 1.1294, "step": 820 }, { "epoch": 6.006769230769231, "grad_norm": 0.6176318526268005, "learning_rate": 9.692307692307693e-06, "loss": 0.538, "step": 830 }, { "epoch": 6.008307692307692, "grad_norm": 16.801511764526367, "learning_rate": 9.675213675213677e-06, "loss": 1.0966, "step": 840 }, { "epoch": 6.009846153846154, "grad_norm": 0.6106650829315186, "learning_rate": 9.658119658119659e-06, "loss": 0.6375, "step": 850 }, { "epoch": 6.011384615384616, "grad_norm": 1.5337104797363281, "learning_rate": 9.641025641025642e-06, "loss": 1.0615, "step": 860 }, { "epoch": 6.012923076923077, "grad_norm": 1.1498308181762695, "learning_rate": 9.623931623931626e-06, "loss": 0.5056, "step": 870 }, { "epoch": 6.014461538461538, "grad_norm": 1.3140219449996948, "learning_rate": 9.606837606837607e-06, "loss": 0.8228, "step": 880 }, { "epoch": 6.016, "grad_norm": 0.12818345427513123, "learning_rate": 9.589743589743591e-06, "loss": 0.2663, "step": 890 }, { "epoch": 6.017538461538462, "grad_norm": 1.8986399173736572, "learning_rate": 9.572649572649575e-06, "loss": 1.5698, "step": 900 }, { "epoch": 6.019076923076923, "grad_norm": 3.083142042160034, "learning_rate": 9.555555555555556e-06, "loss": 0.7364, "step": 910 }, { "epoch": 6.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.2200851440429688, "eval_runtime": 11.8352, "eval_samples_per_second": 3.887, "eval_steps_per_second": 1.943, "step": 917 }, { "epoch": 7.000461538461538, "grad_norm": 0.997331976890564, "learning_rate": 9.53846153846154e-06, "loss": 0.5659, "step": 920 }, { "epoch": 7.002, "grad_norm": 1.0925484895706177, "learning_rate": 9.521367521367522e-06, "loss": 0.8216, "step": 930 }, { "epoch": 7.0035384615384615, "grad_norm": 0.649455726146698, "learning_rate": 9.504273504273504e-06, "loss": 0.9357, "step": 940 }, { "epoch": 7.005076923076923, "grad_norm": 10.865209579467773, "learning_rate": 9.487179487179487e-06, "loss": 0.786, "step": 950 }, { "epoch": 7.006615384615385, "grad_norm": 0.565341055393219, "learning_rate": 9.470085470085471e-06, "loss": 1.0838, "step": 960 }, { "epoch": 7.008153846153846, "grad_norm": 11.902983665466309, "learning_rate": 9.452991452991453e-06, "loss": 1.0737, "step": 970 }, { "epoch": 7.0096923076923074, "grad_norm": 0.8235186338424683, "learning_rate": 9.435897435897436e-06, "loss": 0.705, "step": 980 }, { "epoch": 7.011230769230769, "grad_norm": 0.6383252739906311, "learning_rate": 9.41880341880342e-06, "loss": 1.0167, "step": 990 }, { "epoch": 7.012769230769231, "grad_norm": 0.4198904037475586, "learning_rate": 9.401709401709402e-06, "loss": 0.4029, "step": 1000 }, { "epoch": 7.014307692307693, "grad_norm": 0.6520665287971497, "learning_rate": 9.384615384615385e-06, "loss": 0.8384, "step": 1010 }, { "epoch": 7.015846153846154, "grad_norm": 0.3294767737388611, "learning_rate": 9.367521367521369e-06, "loss": 1.1558, "step": 1020 }, { "epoch": 7.017384615384615, "grad_norm": 12.820337295532227, "learning_rate": 9.35042735042735e-06, "loss": 0.6545, "step": 1030 }, { "epoch": 7.018923076923077, "grad_norm": 1.148991584777832, "learning_rate": 9.333333333333334e-06, "loss": 1.4976, "step": 1040 }, { "epoch": 7.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 0.9824587106704712, "eval_runtime": 11.7676, "eval_samples_per_second": 3.909, "eval_steps_per_second": 1.955, "step": 1048 }, { "epoch": 8.000307692307691, "grad_norm": 2.197751045227051, "learning_rate": 9.316239316239318e-06, "loss": 0.6061, "step": 1050 }, { "epoch": 8.001846153846154, "grad_norm": 0.49791476130485535, "learning_rate": 9.2991452991453e-06, "loss": 0.6114, "step": 1060 }, { "epoch": 8.003384615384615, "grad_norm": 0.7892986536026001, "learning_rate": 9.282051282051283e-06, "loss": 1.132, "step": 1070 }, { "epoch": 8.004923076923077, "grad_norm": 0.17233861982822418, "learning_rate": 9.264957264957267e-06, "loss": 0.2331, "step": 1080 }, { "epoch": 8.006461538461538, "grad_norm": 27.1361083984375, "learning_rate": 9.247863247863249e-06, "loss": 1.9122, "step": 1090 }, { "epoch": 8.008, "grad_norm": 3.272629976272583, "learning_rate": 9.230769230769232e-06, "loss": 0.9478, "step": 1100 }, { "epoch": 8.009538461538462, "grad_norm": 2.749512195587158, "learning_rate": 9.213675213675214e-06, "loss": 0.8206, "step": 1110 }, { "epoch": 8.011076923076923, "grad_norm": 2.485612392425537, "learning_rate": 9.196581196581196e-06, "loss": 0.4048, "step": 1120 }, { "epoch": 8.012615384615385, "grad_norm": 0.7969468235969543, "learning_rate": 9.17948717948718e-06, "loss": 0.5164, "step": 1130 }, { "epoch": 8.014153846153846, "grad_norm": 19.155437469482422, "learning_rate": 9.162393162393163e-06, "loss": 1.4568, "step": 1140 }, { "epoch": 8.015692307692309, "grad_norm": 9.772501945495605, "learning_rate": 9.145299145299145e-06, "loss": 0.5541, "step": 1150 }, { "epoch": 8.01723076923077, "grad_norm": 0.7671765685081482, "learning_rate": 9.128205128205129e-06, "loss": 0.7012, "step": 1160 }, { "epoch": 8.01876923076923, "grad_norm": 0.7729051113128662, "learning_rate": 9.111111111111112e-06, "loss": 0.7766, "step": 1170 }, { "epoch": 8.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.4507434368133545, "eval_runtime": 12.5475, "eval_samples_per_second": 3.666, "eval_steps_per_second": 1.833, "step": 1179 }, { "epoch": 9.000153846153847, "grad_norm": 0.7271968126296997, "learning_rate": 9.094017094017094e-06, "loss": 0.8353, "step": 1180 }, { "epoch": 9.001692307692307, "grad_norm": 10.993083000183105, "learning_rate": 9.076923076923078e-06, "loss": 0.8165, "step": 1190 }, { "epoch": 9.00323076923077, "grad_norm": 0.34806424379348755, "learning_rate": 9.059829059829061e-06, "loss": 0.8572, "step": 1200 }, { "epoch": 9.00476923076923, "grad_norm": 7.650801181793213, "learning_rate": 9.042735042735043e-06, "loss": 1.6574, "step": 1210 }, { "epoch": 9.006307692307692, "grad_norm": 1.8591915369033813, "learning_rate": 9.025641025641027e-06, "loss": 0.7499, "step": 1220 }, { "epoch": 9.007846153846154, "grad_norm": 5.587526798248291, "learning_rate": 9.00854700854701e-06, "loss": 0.6896, "step": 1230 }, { "epoch": 9.009384615384615, "grad_norm": 1.4475104808807373, "learning_rate": 8.991452991452992e-06, "loss": 0.4126, "step": 1240 }, { "epoch": 9.010923076923078, "grad_norm": 9.723190307617188, "learning_rate": 8.974358974358976e-06, "loss": 0.2476, "step": 1250 }, { "epoch": 9.012461538461539, "grad_norm": 0.20038339495658875, "learning_rate": 8.957264957264959e-06, "loss": 0.6484, "step": 1260 }, { "epoch": 9.014, "grad_norm": 0.46997225284576416, "learning_rate": 8.940170940170941e-06, "loss": 1.1973, "step": 1270 }, { "epoch": 9.015538461538462, "grad_norm": 7.039483547210693, "learning_rate": 8.923076923076925e-06, "loss": 1.7299, "step": 1280 }, { "epoch": 9.017076923076923, "grad_norm": 8.236711502075195, "learning_rate": 8.905982905982906e-06, "loss": 0.5039, "step": 1290 }, { "epoch": 9.018615384615385, "grad_norm": 8.963602066040039, "learning_rate": 8.888888888888888e-06, "loss": 0.5795, "step": 1300 }, { "epoch": 9.020153846153846, "grad_norm": 2.075558662414551, "learning_rate": 8.871794871794872e-06, "loss": 0.8254, "step": 1310 }, { "epoch": 9.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.0844687223434448, "eval_runtime": 12.4975, "eval_samples_per_second": 3.681, "eval_steps_per_second": 1.84, "step": 1310 }, { "epoch": 10.00153846153846, "grad_norm": 1.5464600324630737, "learning_rate": 8.854700854700855e-06, "loss": 0.6458, "step": 1320 }, { "epoch": 10.003076923076923, "grad_norm": 9.952284812927246, "learning_rate": 8.837606837606837e-06, "loss": 0.7211, "step": 1330 }, { "epoch": 10.004615384615384, "grad_norm": 0.5603816509246826, "learning_rate": 8.820512820512821e-06, "loss": 0.5634, "step": 1340 }, { "epoch": 10.006153846153847, "grad_norm": 17.749595642089844, "learning_rate": 8.803418803418804e-06, "loss": 1.0856, "step": 1350 }, { "epoch": 10.007692307692308, "grad_norm": 16.027507781982422, "learning_rate": 8.786324786324786e-06, "loss": 0.7795, "step": 1360 }, { "epoch": 10.009230769230768, "grad_norm": 0.3579050600528717, "learning_rate": 8.76923076923077e-06, "loss": 0.3561, "step": 1370 }, { "epoch": 10.010769230769231, "grad_norm": 25.084381103515625, "learning_rate": 8.752136752136753e-06, "loss": 1.1388, "step": 1380 }, { "epoch": 10.012307692307692, "grad_norm": 1.1872502565383911, "learning_rate": 8.735042735042735e-06, "loss": 0.8342, "step": 1390 }, { "epoch": 10.013846153846155, "grad_norm": 11.217093467712402, "learning_rate": 8.717948717948719e-06, "loss": 0.6937, "step": 1400 }, { "epoch": 10.015384615384615, "grad_norm": 8.533183097839355, "learning_rate": 8.700854700854702e-06, "loss": 1.1324, "step": 1410 }, { "epoch": 10.016923076923076, "grad_norm": 8.81098747253418, "learning_rate": 8.683760683760684e-06, "loss": 1.4037, "step": 1420 }, { "epoch": 10.018461538461539, "grad_norm": 1.5235704183578491, "learning_rate": 8.666666666666668e-06, "loss": 0.3098, "step": 1430 }, { "epoch": 10.02, "grad_norm": 0.6626542210578918, "learning_rate": 8.649572649572651e-06, "loss": 0.7267, "step": 1440 }, { "epoch": 10.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.270902395248413, "eval_runtime": 12.4316, "eval_samples_per_second": 3.7, "eval_steps_per_second": 1.85, "step": 1441 }, { "epoch": 11.001384615384616, "grad_norm": 16.215713500976562, "learning_rate": 8.632478632478633e-06, "loss": 1.2159, "step": 1450 }, { "epoch": 11.002923076923077, "grad_norm": 0.3360421061515808, "learning_rate": 8.615384615384617e-06, "loss": 0.7116, "step": 1460 }, { "epoch": 11.004461538461538, "grad_norm": 8.118979454040527, "learning_rate": 8.598290598290599e-06, "loss": 0.6784, "step": 1470 }, { "epoch": 11.006, "grad_norm": 1.1598076820373535, "learning_rate": 8.58119658119658e-06, "loss": 1.2064, "step": 1480 }, { "epoch": 11.007538461538461, "grad_norm": 8.073029518127441, "learning_rate": 8.564102564102564e-06, "loss": 1.0324, "step": 1490 }, { "epoch": 11.009076923076924, "grad_norm": 2.737980604171753, "learning_rate": 8.547008547008548e-06, "loss": 0.6106, "step": 1500 }, { "epoch": 11.010615384615384, "grad_norm": 1.8243392705917358, "learning_rate": 8.52991452991453e-06, "loss": 0.4851, "step": 1510 }, { "epoch": 11.012153846153845, "grad_norm": 2.195685863494873, "learning_rate": 8.512820512820513e-06, "loss": 0.2933, "step": 1520 }, { "epoch": 11.013692307692308, "grad_norm": 11.78964900970459, "learning_rate": 8.495726495726497e-06, "loss": 1.4347, "step": 1530 }, { "epoch": 11.015230769230769, "grad_norm": 7.746746063232422, "learning_rate": 8.478632478632479e-06, "loss": 0.6918, "step": 1540 }, { "epoch": 11.016769230769231, "grad_norm": 0.4824214577674866, "learning_rate": 8.461538461538462e-06, "loss": 0.3756, "step": 1550 }, { "epoch": 11.018307692307692, "grad_norm": 7.721612453460693, "learning_rate": 8.444444444444446e-06, "loss": 0.9131, "step": 1560 }, { "epoch": 11.019846153846153, "grad_norm": 1.2157827615737915, "learning_rate": 8.427350427350428e-06, "loss": 0.6193, "step": 1570 }, { "epoch": 11.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.2886906862258911, "eval_runtime": 12.4564, "eval_samples_per_second": 3.693, "eval_steps_per_second": 1.846, "step": 1572 }, { "epoch": 12.00123076923077, "grad_norm": 15.795467376708984, "learning_rate": 8.410256410256411e-06, "loss": 0.5223, "step": 1580 }, { "epoch": 12.00276923076923, "grad_norm": 0.2842244803905487, "learning_rate": 8.393162393162395e-06, "loss": 0.5231, "step": 1590 }, { "epoch": 12.004307692307693, "grad_norm": 18.358034133911133, "learning_rate": 8.376068376068377e-06, "loss": 0.8485, "step": 1600 }, { "epoch": 12.005846153846154, "grad_norm": 18.451539993286133, "learning_rate": 8.35897435897436e-06, "loss": 1.1096, "step": 1610 }, { "epoch": 12.007384615384616, "grad_norm": 14.058935165405273, "learning_rate": 8.341880341880344e-06, "loss": 0.4075, "step": 1620 }, { "epoch": 12.008923076923077, "grad_norm": 3.4443199634552, "learning_rate": 8.324786324786326e-06, "loss": 0.831, "step": 1630 }, { "epoch": 12.010461538461538, "grad_norm": 0.17183153331279755, "learning_rate": 8.307692307692309e-06, "loss": 1.2585, "step": 1640 }, { "epoch": 12.012, "grad_norm": 0.2953018546104431, "learning_rate": 8.290598290598293e-06, "loss": 0.4976, "step": 1650 }, { "epoch": 12.013538461538461, "grad_norm": 14.40168571472168, "learning_rate": 8.273504273504273e-06, "loss": 1.2874, "step": 1660 }, { "epoch": 12.015076923076924, "grad_norm": 1.2708631753921509, "learning_rate": 8.256410256410256e-06, "loss": 0.9277, "step": 1670 }, { "epoch": 12.016615384615385, "grad_norm": 7.030950546264648, "learning_rate": 8.23931623931624e-06, "loss": 0.8064, "step": 1680 }, { "epoch": 12.018153846153846, "grad_norm": 5.090956687927246, "learning_rate": 8.222222222222222e-06, "loss": 0.8262, "step": 1690 }, { "epoch": 12.019692307692308, "grad_norm": 6.321172714233398, "learning_rate": 8.205128205128205e-06, "loss": 0.637, "step": 1700 }, { "epoch": 12.020153846153846, "eval_accuracy": 0.5869565217391305, "eval_loss": 0.6063858270645142, "eval_runtime": 11.7832, "eval_samples_per_second": 3.904, "eval_steps_per_second": 1.952, "step": 1703 }, { "epoch": 13.001076923076923, "grad_norm": 6.917983531951904, "learning_rate": 8.188034188034189e-06, "loss": 0.371, "step": 1710 }, { "epoch": 13.002615384615385, "grad_norm": 2.6089928150177, "learning_rate": 8.17094017094017e-06, "loss": 0.3077, "step": 1720 }, { "epoch": 13.004153846153846, "grad_norm": 0.46078240871429443, "learning_rate": 8.153846153846154e-06, "loss": 0.7638, "step": 1730 }, { "epoch": 13.005692307692307, "grad_norm": 11.816447257995605, "learning_rate": 8.136752136752138e-06, "loss": 0.9692, "step": 1740 }, { "epoch": 13.00723076923077, "grad_norm": 3.5472028255462646, "learning_rate": 8.11965811965812e-06, "loss": 0.4919, "step": 1750 }, { "epoch": 13.00876923076923, "grad_norm": 2.8847992420196533, "learning_rate": 8.102564102564103e-06, "loss": 1.0981, "step": 1760 }, { "epoch": 13.010307692307693, "grad_norm": 7.087960720062256, "learning_rate": 8.085470085470087e-06, "loss": 0.7239, "step": 1770 }, { "epoch": 13.011846153846154, "grad_norm": 2.2250142097473145, "learning_rate": 8.068376068376069e-06, "loss": 0.5984, "step": 1780 }, { "epoch": 13.013384615384615, "grad_norm": 12.928844451904297, "learning_rate": 8.051282051282052e-06, "loss": 0.6177, "step": 1790 }, { "epoch": 13.014923076923077, "grad_norm": 39.16981887817383, "learning_rate": 8.034188034188036e-06, "loss": 0.9743, "step": 1800 }, { "epoch": 13.016461538461538, "grad_norm": 1.7411730289459229, "learning_rate": 8.017094017094018e-06, "loss": 0.9101, "step": 1810 }, { "epoch": 13.018, "grad_norm": 0.1714581549167633, "learning_rate": 8.000000000000001e-06, "loss": 0.5956, "step": 1820 }, { "epoch": 13.019538461538462, "grad_norm": 18.013525009155273, "learning_rate": 7.982905982905985e-06, "loss": 0.7321, "step": 1830 }, { "epoch": 13.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 0.9141055941581726, "eval_runtime": 11.8108, "eval_samples_per_second": 3.895, "eval_steps_per_second": 1.947, "step": 1834 }, { "epoch": 14.000923076923076, "grad_norm": 43.33615493774414, "learning_rate": 7.965811965811967e-06, "loss": 1.1067, "step": 1840 }, { "epoch": 14.002461538461539, "grad_norm": 13.135071754455566, "learning_rate": 7.948717948717949e-06, "loss": 0.5216, "step": 1850 }, { "epoch": 14.004, "grad_norm": 2.0259616374969482, "learning_rate": 7.931623931623932e-06, "loss": 0.4612, "step": 1860 }, { "epoch": 14.005538461538462, "grad_norm": 8.404983520507812, "learning_rate": 7.914529914529914e-06, "loss": 0.6618, "step": 1870 }, { "epoch": 14.007076923076923, "grad_norm": 10.403939247131348, "learning_rate": 7.897435897435898e-06, "loss": 0.4837, "step": 1880 }, { "epoch": 14.008615384615384, "grad_norm": 3.110192060470581, "learning_rate": 7.880341880341881e-06, "loss": 0.2619, "step": 1890 }, { "epoch": 14.010153846153846, "grad_norm": 61.4323844909668, "learning_rate": 7.863247863247863e-06, "loss": 0.5425, "step": 1900 }, { "epoch": 14.011692307692307, "grad_norm": 37.54103469848633, "learning_rate": 7.846153846153847e-06, "loss": 0.7557, "step": 1910 }, { "epoch": 14.01323076923077, "grad_norm": 17.662887573242188, "learning_rate": 7.82905982905983e-06, "loss": 1.6872, "step": 1920 }, { "epoch": 14.01476923076923, "grad_norm": 6.54232120513916, "learning_rate": 7.811965811965812e-06, "loss": 0.9641, "step": 1930 }, { "epoch": 14.016307692307691, "grad_norm": 2.722644090652466, "learning_rate": 7.794871794871796e-06, "loss": 0.0621, "step": 1940 }, { "epoch": 14.017846153846154, "grad_norm": 0.08367615938186646, "learning_rate": 7.77777777777778e-06, "loss": 1.3665, "step": 1950 }, { "epoch": 14.019384615384615, "grad_norm": 0.41343918442726135, "learning_rate": 7.760683760683761e-06, "loss": 1.0134, "step": 1960 }, { "epoch": 14.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.6379895210266113, "eval_runtime": 12.2392, "eval_samples_per_second": 3.758, "eval_steps_per_second": 1.879, "step": 1965 }, { "epoch": 15.000769230769231, "grad_norm": 0.2228805422782898, "learning_rate": 7.743589743589745e-06, "loss": 0.8081, "step": 1970 }, { "epoch": 15.002307692307692, "grad_norm": 0.21026729047298431, "learning_rate": 7.726495726495728e-06, "loss": 0.4037, "step": 1980 }, { "epoch": 15.003846153846155, "grad_norm": 26.223190307617188, "learning_rate": 7.70940170940171e-06, "loss": 1.2921, "step": 1990 }, { "epoch": 15.005384615384616, "grad_norm": 15.445429801940918, "learning_rate": 7.692307692307694e-06, "loss": 1.4975, "step": 2000 }, { "epoch": 15.006923076923076, "grad_norm": 4.351694583892822, "learning_rate": 7.675213675213677e-06, "loss": 0.7678, "step": 2010 }, { "epoch": 15.008461538461539, "grad_norm": 16.3181095123291, "learning_rate": 7.658119658119659e-06, "loss": 0.3639, "step": 2020 }, { "epoch": 15.01, "grad_norm": 79.23714447021484, "learning_rate": 7.641025641025641e-06, "loss": 0.7732, "step": 2030 }, { "epoch": 15.011538461538462, "grad_norm": 0.3189145028591156, "learning_rate": 7.6239316239316244e-06, "loss": 0.285, "step": 2040 }, { "epoch": 15.013076923076923, "grad_norm": 0.07052574306726456, "learning_rate": 7.606837606837607e-06, "loss": 0.5103, "step": 2050 }, { "epoch": 15.014615384615384, "grad_norm": 8.32340145111084, "learning_rate": 7.58974358974359e-06, "loss": 0.232, "step": 2060 }, { "epoch": 15.016153846153847, "grad_norm": 30.055326461791992, "learning_rate": 7.572649572649573e-06, "loss": 0.603, "step": 2070 }, { "epoch": 15.017692307692307, "grad_norm": 21.265649795532227, "learning_rate": 7.555555555555556e-06, "loss": 1.3278, "step": 2080 }, { "epoch": 15.01923076923077, "grad_norm": 13.572966575622559, "learning_rate": 7.538461538461539e-06, "loss": 0.2139, "step": 2090 }, { "epoch": 15.020153846153846, "eval_accuracy": 0.6956521739130435, "eval_loss": 0.6075901985168457, "eval_runtime": 12.5118, "eval_samples_per_second": 3.677, "eval_steps_per_second": 1.838, "step": 2096 }, { "epoch": 16.000615384615383, "grad_norm": 6.703625679016113, "learning_rate": 7.521367521367522e-06, "loss": 0.5604, "step": 2100 }, { "epoch": 16.002153846153846, "grad_norm": 0.617182731628418, "learning_rate": 7.504273504273505e-06, "loss": 0.561, "step": 2110 }, { "epoch": 16.003692307692308, "grad_norm": 41.750701904296875, "learning_rate": 7.487179487179488e-06, "loss": 0.8217, "step": 2120 }, { "epoch": 16.00523076923077, "grad_norm": 0.29348522424697876, "learning_rate": 7.4700854700854706e-06, "loss": 0.5181, "step": 2130 }, { "epoch": 16.00676923076923, "grad_norm": 33.394954681396484, "learning_rate": 7.452991452991454e-06, "loss": 0.8665, "step": 2140 }, { "epoch": 16.008307692307692, "grad_norm": 18.40253448486328, "learning_rate": 7.435897435897437e-06, "loss": 0.4002, "step": 2150 }, { "epoch": 16.009846153846155, "grad_norm": 61.75335693359375, "learning_rate": 7.4188034188034196e-06, "loss": 0.8184, "step": 2160 }, { "epoch": 16.011384615384614, "grad_norm": 30.64858627319336, "learning_rate": 7.401709401709402e-06, "loss": 0.9719, "step": 2170 }, { "epoch": 16.012923076923077, "grad_norm": 20.503198623657227, "learning_rate": 7.384615384615386e-06, "loss": 0.6894, "step": 2180 }, { "epoch": 16.01446153846154, "grad_norm": 0.41924965381622314, "learning_rate": 7.3675213675213686e-06, "loss": 0.5238, "step": 2190 }, { "epoch": 16.016, "grad_norm": 0.09503482282161713, "learning_rate": 7.350427350427351e-06, "loss": 0.3226, "step": 2200 }, { "epoch": 16.01753846153846, "grad_norm": 15.260120391845703, "learning_rate": 7.333333333333333e-06, "loss": 1.7077, "step": 2210 }, { "epoch": 16.019076923076923, "grad_norm": 40.093467712402344, "learning_rate": 7.316239316239317e-06, "loss": 0.687, "step": 2220 }, { "epoch": 16.020153846153846, "eval_accuracy": 0.5434782608695652, "eval_loss": 1.1890361309051514, "eval_runtime": 12.5109, "eval_samples_per_second": 3.677, "eval_steps_per_second": 1.838, "step": 2227 }, { "epoch": 17.00046153846154, "grad_norm": 22.440990447998047, "learning_rate": 7.299145299145299e-06, "loss": 0.7403, "step": 2230 }, { "epoch": 17.002, "grad_norm": 30.42022705078125, "learning_rate": 7.282051282051282e-06, "loss": 0.6101, "step": 2240 }, { "epoch": 17.00353846153846, "grad_norm": 28.41487693786621, "learning_rate": 7.264957264957266e-06, "loss": 0.6806, "step": 2250 }, { "epoch": 17.005076923076924, "grad_norm": 0.5429576635360718, "learning_rate": 7.247863247863248e-06, "loss": 0.7311, "step": 2260 }, { "epoch": 17.006615384615383, "grad_norm": 0.37554216384887695, "learning_rate": 7.230769230769231e-06, "loss": 0.6523, "step": 2270 }, { "epoch": 17.008153846153846, "grad_norm": 12.936208724975586, "learning_rate": 7.213675213675214e-06, "loss": 0.7911, "step": 2280 }, { "epoch": 17.00969230769231, "grad_norm": 20.977012634277344, "learning_rate": 7.196581196581197e-06, "loss": 0.2984, "step": 2290 }, { "epoch": 17.01123076923077, "grad_norm": 0.6222103834152222, "learning_rate": 7.17948717948718e-06, "loss": 0.3332, "step": 2300 }, { "epoch": 17.01276923076923, "grad_norm": 16.443161010742188, "learning_rate": 7.162393162393163e-06, "loss": 0.8002, "step": 2310 }, { "epoch": 17.014307692307693, "grad_norm": 0.1513749361038208, "learning_rate": 7.145299145299146e-06, "loss": 0.7041, "step": 2320 }, { "epoch": 17.015846153846155, "grad_norm": 0.09506427496671677, "learning_rate": 7.128205128205129e-06, "loss": 0.6438, "step": 2330 }, { "epoch": 17.017384615384614, "grad_norm": 14.386686325073242, "learning_rate": 7.111111111111112e-06, "loss": 0.3637, "step": 2340 }, { "epoch": 17.018923076923077, "grad_norm": 18.201074600219727, "learning_rate": 7.0940170940170945e-06, "loss": 0.8722, "step": 2350 }, { "epoch": 17.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 0.5483578443527222, "eval_runtime": 12.6125, "eval_samples_per_second": 3.647, "eval_steps_per_second": 1.824, "step": 2358 }, { "epoch": 18.000307692307693, "grad_norm": 28.085227966308594, "learning_rate": 7.076923076923078e-06, "loss": 0.4885, "step": 2360 }, { "epoch": 18.001846153846152, "grad_norm": 46.327796936035156, "learning_rate": 7.059829059829061e-06, "loss": 0.896, "step": 2370 }, { "epoch": 18.003384615384615, "grad_norm": 8.360269546508789, "learning_rate": 7.0427350427350435e-06, "loss": 0.3218, "step": 2380 }, { "epoch": 18.004923076923077, "grad_norm": 16.091289520263672, "learning_rate": 7.025641025641025e-06, "loss": 0.4428, "step": 2390 }, { "epoch": 18.00646153846154, "grad_norm": 0.039979055523872375, "learning_rate": 7.008547008547009e-06, "loss": 0.1822, "step": 2400 }, { "epoch": 18.008, "grad_norm": 0.1585582047700882, "learning_rate": 6.991452991452992e-06, "loss": 0.4568, "step": 2410 }, { "epoch": 18.00953846153846, "grad_norm": 25.942386627197266, "learning_rate": 6.974358974358974e-06, "loss": 0.5834, "step": 2420 }, { "epoch": 18.011076923076924, "grad_norm": 22.262611389160156, "learning_rate": 6.957264957264958e-06, "loss": 0.6944, "step": 2430 }, { "epoch": 18.012615384615383, "grad_norm": 39.91915512084961, "learning_rate": 6.940170940170941e-06, "loss": 0.3754, "step": 2440 }, { "epoch": 18.014153846153846, "grad_norm": 12.724471092224121, "learning_rate": 6.923076923076923e-06, "loss": 0.4949, "step": 2450 }, { "epoch": 18.01569230769231, "grad_norm": 0.6947053670883179, "learning_rate": 6.905982905982906e-06, "loss": 0.5062, "step": 2460 }, { "epoch": 18.017230769230768, "grad_norm": 12.842805862426758, "learning_rate": 6.88888888888889e-06, "loss": 1.3376, "step": 2470 }, { "epoch": 18.01876923076923, "grad_norm": 65.57331085205078, "learning_rate": 6.871794871794872e-06, "loss": 1.176, "step": 2480 }, { "epoch": 18.020153846153846, "eval_accuracy": 0.8695652173913043, "eval_loss": 0.40178215503692627, "eval_runtime": 11.8576, "eval_samples_per_second": 3.879, "eval_steps_per_second": 1.94, "step": 2489 }, { "epoch": 19.000153846153847, "grad_norm": 23.745304107666016, "learning_rate": 6.854700854700855e-06, "loss": 0.5627, "step": 2490 }, { "epoch": 19.00169230769231, "grad_norm": 7.5470075607299805, "learning_rate": 6.837606837606839e-06, "loss": 0.5329, "step": 2500 }, { "epoch": 19.00323076923077, "grad_norm": 0.15952041745185852, "learning_rate": 6.820512820512821e-06, "loss": 0.5293, "step": 2510 }, { "epoch": 19.00476923076923, "grad_norm": 22.13117790222168, "learning_rate": 6.803418803418804e-06, "loss": 1.0728, "step": 2520 }, { "epoch": 19.006307692307693, "grad_norm": 3.7636966705322266, "learning_rate": 6.786324786324787e-06, "loss": 0.6426, "step": 2530 }, { "epoch": 19.007846153846153, "grad_norm": 41.60896682739258, "learning_rate": 6.76923076923077e-06, "loss": 0.4748, "step": 2540 }, { "epoch": 19.009384615384615, "grad_norm": 2.7363779544830322, "learning_rate": 6.752136752136753e-06, "loss": 0.682, "step": 2550 }, { "epoch": 19.010923076923078, "grad_norm": 49.60470199584961, "learning_rate": 6.735042735042736e-06, "loss": 0.4322, "step": 2560 }, { "epoch": 19.012461538461537, "grad_norm": 0.17257747054100037, "learning_rate": 6.717948717948718e-06, "loss": 0.298, "step": 2570 }, { "epoch": 19.014, "grad_norm": 45.82680892944336, "learning_rate": 6.700854700854701e-06, "loss": 0.77, "step": 2580 }, { "epoch": 19.015538461538462, "grad_norm": 68.6201171875, "learning_rate": 6.683760683760684e-06, "loss": 0.6183, "step": 2590 }, { "epoch": 19.017076923076925, "grad_norm": 7.198949337005615, "learning_rate": 6.666666666666667e-06, "loss": 0.3093, "step": 2600 }, { "epoch": 19.018615384615384, "grad_norm": 0.10329917073249817, "learning_rate": 6.64957264957265e-06, "loss": 0.5356, "step": 2610 }, { "epoch": 19.020153846153846, "grad_norm": 90.10623168945312, "learning_rate": 6.632478632478633e-06, "loss": 0.8042, "step": 2620 }, { "epoch": 19.020153846153846, "eval_accuracy": 0.717391304347826, "eval_loss": 0.9282347559928894, "eval_runtime": 11.7547, "eval_samples_per_second": 3.913, "eval_steps_per_second": 1.957, "step": 2620 }, { "epoch": 20.001538461538463, "grad_norm": 58.79783248901367, "learning_rate": 6.615384615384616e-06, "loss": 0.4324, "step": 2630 }, { "epoch": 20.00307692307692, "grad_norm": 0.12335860729217529, "learning_rate": 6.598290598290598e-06, "loss": 0.6378, "step": 2640 }, { "epoch": 20.004615384615384, "grad_norm": 1.2578246593475342, "learning_rate": 6.581196581196582e-06, "loss": 0.5711, "step": 2650 }, { "epoch": 20.006153846153847, "grad_norm": 9.80817699432373, "learning_rate": 6.564102564102565e-06, "loss": 0.2199, "step": 2660 }, { "epoch": 20.00769230769231, "grad_norm": 0.04995802044868469, "learning_rate": 6.547008547008547e-06, "loss": 0.5418, "step": 2670 }, { "epoch": 20.00923076923077, "grad_norm": 76.49468231201172, "learning_rate": 6.529914529914531e-06, "loss": 0.4855, "step": 2680 }, { "epoch": 20.01076923076923, "grad_norm": 1.1559330224990845, "learning_rate": 6.512820512820514e-06, "loss": 0.9054, "step": 2690 }, { "epoch": 20.012307692307694, "grad_norm": 91.41716003417969, "learning_rate": 6.495726495726496e-06, "loss": 0.3235, "step": 2700 }, { "epoch": 20.013846153846153, "grad_norm": 70.10804748535156, "learning_rate": 6.478632478632479e-06, "loss": 1.0521, "step": 2710 }, { "epoch": 20.015384615384615, "grad_norm": 0.5130220055580139, "learning_rate": 6.461538461538463e-06, "loss": 0.4354, "step": 2720 }, { "epoch": 20.016923076923078, "grad_norm": 26.976499557495117, "learning_rate": 6.444444444444445e-06, "loss": 1.1985, "step": 2730 }, { "epoch": 20.018461538461537, "grad_norm": 49.21026611328125, "learning_rate": 6.427350427350428e-06, "loss": 1.6716, "step": 2740 }, { "epoch": 20.02, "grad_norm": 0.3829849660396576, "learning_rate": 6.410256410256412e-06, "loss": 0.3013, "step": 2750 }, { "epoch": 20.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 0.5082508325576782, "eval_runtime": 13.0921, "eval_samples_per_second": 3.514, "eval_steps_per_second": 1.757, "step": 2751 }, { "epoch": 21.001384615384616, "grad_norm": 10.179632186889648, "learning_rate": 6.3931623931623935e-06, "loss": 0.6051, "step": 2760 }, { "epoch": 21.00292307692308, "grad_norm": 20.752038955688477, "learning_rate": 6.376068376068376e-06, "loss": 0.8644, "step": 2770 }, { "epoch": 21.004461538461538, "grad_norm": 0.5752691030502319, "learning_rate": 6.358974358974359e-06, "loss": 0.4362, "step": 2780 }, { "epoch": 21.006, "grad_norm": 1.0075253248214722, "learning_rate": 6.3418803418803425e-06, "loss": 0.9587, "step": 2790 }, { "epoch": 21.007538461538463, "grad_norm": 1.7478026151657104, "learning_rate": 6.324786324786325e-06, "loss": 0.9172, "step": 2800 }, { "epoch": 21.009076923076922, "grad_norm": 59.94158935546875, "learning_rate": 6.307692307692308e-06, "loss": 0.3976, "step": 2810 }, { "epoch": 21.010615384615384, "grad_norm": 0.6232649683952332, "learning_rate": 6.290598290598291e-06, "loss": 0.2702, "step": 2820 }, { "epoch": 21.012153846153847, "grad_norm": 0.06733821332454681, "learning_rate": 6.273504273504274e-06, "loss": 0.7284, "step": 2830 }, { "epoch": 21.013692307692306, "grad_norm": 0.04839285835623741, "learning_rate": 6.256410256410257e-06, "loss": 0.7976, "step": 2840 }, { "epoch": 21.01523076923077, "grad_norm": 0.7099334597587585, "learning_rate": 6.23931623931624e-06, "loss": 0.0915, "step": 2850 }, { "epoch": 21.01676923076923, "grad_norm": 37.17452621459961, "learning_rate": 6.222222222222223e-06, "loss": 0.8673, "step": 2860 }, { "epoch": 21.018307692307694, "grad_norm": 4.46577262878418, "learning_rate": 6.205128205128206e-06, "loss": 0.9571, "step": 2870 }, { "epoch": 21.019846153846153, "grad_norm": 3.844195604324341, "learning_rate": 6.188034188034189e-06, "loss": 0.646, "step": 2880 }, { "epoch": 21.020153846153846, "eval_accuracy": 0.8260869565217391, "eval_loss": 0.5288403630256653, "eval_runtime": 11.7673, "eval_samples_per_second": 3.909, "eval_steps_per_second": 1.955, "step": 2882 }, { "epoch": 22.00123076923077, "grad_norm": 0.1899593472480774, "learning_rate": 6.170940170940171e-06, "loss": 0.9999, "step": 2890 }, { "epoch": 22.002769230769232, "grad_norm": 12.806218147277832, "learning_rate": 6.153846153846155e-06, "loss": 0.7985, "step": 2900 }, { "epoch": 22.00430769230769, "grad_norm": 20.41181182861328, "learning_rate": 6.136752136752138e-06, "loss": 0.5117, "step": 2910 }, { "epoch": 22.005846153846154, "grad_norm": 0.4226987957954407, "learning_rate": 6.11965811965812e-06, "loss": 0.2177, "step": 2920 }, { "epoch": 22.007384615384616, "grad_norm": 39.437103271484375, "learning_rate": 6.102564102564104e-06, "loss": 0.2983, "step": 2930 }, { "epoch": 22.008923076923075, "grad_norm": 0.6932996511459351, "learning_rate": 6.085470085470086e-06, "loss": 0.3892, "step": 2940 }, { "epoch": 22.010461538461538, "grad_norm": 0.663569986820221, "learning_rate": 6.0683760683760684e-06, "loss": 0.5625, "step": 2950 }, { "epoch": 22.012, "grad_norm": 0.21277102828025818, "learning_rate": 6.051282051282051e-06, "loss": 0.4287, "step": 2960 }, { "epoch": 22.013538461538463, "grad_norm": 85.67404174804688, "learning_rate": 6.034188034188035e-06, "loss": 1.0995, "step": 2970 }, { "epoch": 22.015076923076922, "grad_norm": 103.44427490234375, "learning_rate": 6.0170940170940174e-06, "loss": 0.4695, "step": 2980 }, { "epoch": 22.016615384615385, "grad_norm": 0.0456969179213047, "learning_rate": 6e-06, "loss": 0.4399, "step": 2990 }, { "epoch": 22.018153846153847, "grad_norm": 39.62498474121094, "learning_rate": 5.982905982905983e-06, "loss": 0.6504, "step": 3000 }, { "epoch": 22.019692307692306, "grad_norm": 18.204975128173828, "learning_rate": 5.9658119658119664e-06, "loss": 1.3053, "step": 3010 }, { "epoch": 22.020153846153846, "eval_accuracy": 0.6521739130434783, "eval_loss": 1.1223881244659424, "eval_runtime": 11.8128, "eval_samples_per_second": 3.894, "eval_steps_per_second": 1.947, "step": 3013 }, { "epoch": 23.001076923076923, "grad_norm": 0.4275420904159546, "learning_rate": 5.948717948717949e-06, "loss": 0.2397, "step": 3020 }, { "epoch": 23.002615384615385, "grad_norm": 4.155840873718262, "learning_rate": 5.931623931623932e-06, "loss": 0.545, "step": 3030 }, { "epoch": 23.004153846153844, "grad_norm": 0.14375780522823334, "learning_rate": 5.914529914529915e-06, "loss": 0.6312, "step": 3040 }, { "epoch": 23.005692307692307, "grad_norm": 0.049200814217329025, "learning_rate": 5.897435897435898e-06, "loss": 0.6866, "step": 3050 }, { "epoch": 23.00723076923077, "grad_norm": 15.059101104736328, "learning_rate": 5.880341880341881e-06, "loss": 0.7464, "step": 3060 }, { "epoch": 23.008769230769232, "grad_norm": 37.120506286621094, "learning_rate": 5.863247863247864e-06, "loss": 0.2967, "step": 3070 }, { "epoch": 23.01030769230769, "grad_norm": 0.6163169741630554, "learning_rate": 5.846153846153847e-06, "loss": 0.4502, "step": 3080 }, { "epoch": 23.011846153846154, "grad_norm": 79.16121673583984, "learning_rate": 5.82905982905983e-06, "loss": 0.9185, "step": 3090 }, { "epoch": 23.013384615384616, "grad_norm": 3.2994115352630615, "learning_rate": 5.8119658119658126e-06, "loss": 0.4721, "step": 3100 }, { "epoch": 23.014923076923075, "grad_norm": 189.37454223632812, "learning_rate": 5.794871794871796e-06, "loss": 0.7812, "step": 3110 }, { "epoch": 23.016461538461538, "grad_norm": 0.10399169474840164, "learning_rate": 5.777777777777778e-06, "loss": 0.7956, "step": 3120 }, { "epoch": 23.018, "grad_norm": 0.03775927051901817, "learning_rate": 5.760683760683761e-06, "loss": 0.292, "step": 3130 }, { "epoch": 23.019538461538463, "grad_norm": 55.82505798339844, "learning_rate": 5.743589743589743e-06, "loss": 0.9903, "step": 3140 }, { "epoch": 23.020153846153846, "eval_accuracy": 0.8043478260869565, "eval_loss": 0.47060108184814453, "eval_runtime": 11.9647, "eval_samples_per_second": 3.845, "eval_steps_per_second": 1.922, "step": 3144 }, { "epoch": 24.000923076923076, "grad_norm": 1.2134654521942139, "learning_rate": 5.726495726495727e-06, "loss": 0.5882, "step": 3150 }, { "epoch": 24.00246153846154, "grad_norm": 9.802472114562988, "learning_rate": 5.70940170940171e-06, "loss": 0.1785, "step": 3160 }, { "epoch": 24.004, "grad_norm": 0.06365148723125458, "learning_rate": 5.692307692307692e-06, "loss": 0.3565, "step": 3170 }, { "epoch": 24.00553846153846, "grad_norm": 19.367048263549805, "learning_rate": 5.675213675213675e-06, "loss": 1.3728, "step": 3180 }, { "epoch": 24.007076923076923, "grad_norm": 0.3147422671318054, "learning_rate": 5.658119658119659e-06, "loss": 0.7831, "step": 3190 }, { "epoch": 24.008615384615386, "grad_norm": 5.211735248565674, "learning_rate": 5.641025641025641e-06, "loss": 0.4236, "step": 3200 }, { "epoch": 24.010153846153845, "grad_norm": 0.2542964816093445, "learning_rate": 5.623931623931624e-06, "loss": 0.1983, "step": 3210 }, { "epoch": 24.011692307692307, "grad_norm": 34.508331298828125, "learning_rate": 5.606837606837608e-06, "loss": 0.3243, "step": 3220 }, { "epoch": 24.01323076923077, "grad_norm": 0.13946391642093658, "learning_rate": 5.58974358974359e-06, "loss": 0.3546, "step": 3230 }, { "epoch": 24.014769230769232, "grad_norm": 78.96417999267578, "learning_rate": 5.572649572649573e-06, "loss": 0.8219, "step": 3240 }, { "epoch": 24.01630769230769, "grad_norm": 0.11380944401025772, "learning_rate": 5.555555555555557e-06, "loss": 0.669, "step": 3250 }, { "epoch": 24.017846153846154, "grad_norm": 47.72481918334961, "learning_rate": 5.538461538461539e-06, "loss": 1.383, "step": 3260 }, { "epoch": 24.019384615384617, "grad_norm": 30.822315216064453, "learning_rate": 5.521367521367522e-06, "loss": 0.7157, "step": 3270 }, { "epoch": 24.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 0.8162667751312256, "eval_runtime": 11.7412, "eval_samples_per_second": 3.918, "eval_steps_per_second": 1.959, "step": 3275 }, { "epoch": 25.00076923076923, "grad_norm": 0.24460583925247192, "learning_rate": 5.504273504273505e-06, "loss": 0.9906, "step": 3280 }, { "epoch": 25.002307692307692, "grad_norm": 81.8243179321289, "learning_rate": 5.487179487179488e-06, "loss": 0.4271, "step": 3290 }, { "epoch": 25.003846153846155, "grad_norm": 49.65717315673828, "learning_rate": 5.470085470085471e-06, "loss": 0.2202, "step": 3300 }, { "epoch": 25.005384615384614, "grad_norm": 0.29510897397994995, "learning_rate": 5.452991452991453e-06, "loss": 0.5606, "step": 3310 }, { "epoch": 25.006923076923076, "grad_norm": 0.05217023566365242, "learning_rate": 5.435897435897436e-06, "loss": 0.6757, "step": 3320 }, { "epoch": 25.00846153846154, "grad_norm": 0.026966163888573647, "learning_rate": 5.418803418803419e-06, "loss": 1.164, "step": 3330 }, { "epoch": 25.01, "grad_norm": 0.2099008560180664, "learning_rate": 5.401709401709402e-06, "loss": 0.7998, "step": 3340 }, { "epoch": 25.01153846153846, "grad_norm": 3.2398674488067627, "learning_rate": 5.384615384615385e-06, "loss": 0.7801, "step": 3350 }, { "epoch": 25.013076923076923, "grad_norm": 8.356966018676758, "learning_rate": 5.367521367521367e-06, "loss": 0.4305, "step": 3360 }, { "epoch": 25.014615384615386, "grad_norm": 84.3459701538086, "learning_rate": 5.350427350427351e-06, "loss": 0.391, "step": 3370 }, { "epoch": 25.016153846153845, "grad_norm": 2.4751970767974854, "learning_rate": 5.333333333333334e-06, "loss": 0.3874, "step": 3380 }, { "epoch": 25.017692307692307, "grad_norm": 51.4213981628418, "learning_rate": 5.316239316239316e-06, "loss": 0.4588, "step": 3390 }, { "epoch": 25.01923076923077, "grad_norm": 0.3578292429447174, "learning_rate": 5.2991452991453e-06, "loss": 0.3073, "step": 3400 }, { "epoch": 25.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.1408771276474, "eval_runtime": 13.2155, "eval_samples_per_second": 3.481, "eval_steps_per_second": 1.74, "step": 3406 }, { "epoch": 26.000615384615383, "grad_norm": 0.9141789674758911, "learning_rate": 5.282051282051283e-06, "loss": 0.3622, "step": 3410 }, { "epoch": 26.002153846153846, "grad_norm": 318.7442932128906, "learning_rate": 5.264957264957265e-06, "loss": 0.4683, "step": 3420 }, { "epoch": 26.003692307692308, "grad_norm": 65.43302917480469, "learning_rate": 5.247863247863249e-06, "loss": 0.9523, "step": 3430 }, { "epoch": 26.00523076923077, "grad_norm": 0.06784581393003464, "learning_rate": 5.230769230769232e-06, "loss": 0.3191, "step": 3440 }, { "epoch": 26.00676923076923, "grad_norm": 35.10286331176758, "learning_rate": 5.213675213675214e-06, "loss": 1.0472, "step": 3450 }, { "epoch": 26.008307692307692, "grad_norm": 78.8796157836914, "learning_rate": 5.196581196581197e-06, "loss": 0.5854, "step": 3460 }, { "epoch": 26.009846153846155, "grad_norm": 104.45633697509766, "learning_rate": 5.179487179487181e-06, "loss": 0.4423, "step": 3470 }, { "epoch": 26.011384615384614, "grad_norm": 0.09899577498435974, "learning_rate": 5.162393162393163e-06, "loss": 0.6182, "step": 3480 }, { "epoch": 26.012923076923077, "grad_norm": 85.23323059082031, "learning_rate": 5.145299145299145e-06, "loss": 0.4479, "step": 3490 }, { "epoch": 26.01446153846154, "grad_norm": 84.03840637207031, "learning_rate": 5.128205128205128e-06, "loss": 0.4672, "step": 3500 }, { "epoch": 26.016, "grad_norm": 60.73435974121094, "learning_rate": 5.1111111111111115e-06, "loss": 0.5186, "step": 3510 }, { "epoch": 26.01753846153846, "grad_norm": 0.2027343213558197, "learning_rate": 5.094017094017094e-06, "loss": 0.2607, "step": 3520 }, { "epoch": 26.019076923076923, "grad_norm": 0.8834146857261658, "learning_rate": 5.076923076923077e-06, "loss": 0.5069, "step": 3530 }, { "epoch": 26.020153846153846, "eval_accuracy": 0.8043478260869565, "eval_loss": 0.5148446559906006, "eval_runtime": 13.9275, "eval_samples_per_second": 3.303, "eval_steps_per_second": 1.651, "step": 3537 }, { "epoch": 27.00046153846154, "grad_norm": 0.4461596608161926, "learning_rate": 5.05982905982906e-06, "loss": 0.1949, "step": 3540 }, { "epoch": 27.002, "grad_norm": 12.315749168395996, "learning_rate": 5.042735042735043e-06, "loss": 0.139, "step": 3550 }, { "epoch": 27.00353846153846, "grad_norm": 0.21825408935546875, "learning_rate": 5.025641025641026e-06, "loss": 0.5893, "step": 3560 }, { "epoch": 27.005076923076924, "grad_norm": 33.19759750366211, "learning_rate": 5.008547008547009e-06, "loss": 0.8881, "step": 3570 }, { "epoch": 27.006615384615383, "grad_norm": 30.013351440429688, "learning_rate": 4.991452991452992e-06, "loss": 0.4414, "step": 3580 }, { "epoch": 27.008153846153846, "grad_norm": 286.8537292480469, "learning_rate": 4.974358974358975e-06, "loss": 0.6297, "step": 3590 }, { "epoch": 27.00969230769231, "grad_norm": 65.54366302490234, "learning_rate": 4.957264957264958e-06, "loss": 0.8037, "step": 3600 }, { "epoch": 27.01123076923077, "grad_norm": 57.52277755737305, "learning_rate": 4.940170940170941e-06, "loss": 0.4425, "step": 3610 }, { "epoch": 27.01276923076923, "grad_norm": 0.022357992827892303, "learning_rate": 4.923076923076924e-06, "loss": 0.3324, "step": 3620 }, { "epoch": 27.014307692307693, "grad_norm": 2.10927677154541, "learning_rate": 4.905982905982906e-06, "loss": 0.1722, "step": 3630 }, { "epoch": 27.015846153846155, "grad_norm": 214.25904846191406, "learning_rate": 4.888888888888889e-06, "loss": 0.5133, "step": 3640 }, { "epoch": 27.017384615384614, "grad_norm": 26.58594512939453, "learning_rate": 4.871794871794872e-06, "loss": 0.8377, "step": 3650 }, { "epoch": 27.018923076923077, "grad_norm": 0.19922487437725067, "learning_rate": 4.854700854700855e-06, "loss": 0.2683, "step": 3660 }, { "epoch": 27.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.2583907842636108, "eval_runtime": 13.2151, "eval_samples_per_second": 3.481, "eval_steps_per_second": 1.74, "step": 3668 }, { "epoch": 28.000307692307693, "grad_norm": 150.6471710205078, "learning_rate": 4.837606837606838e-06, "loss": 0.6211, "step": 3670 }, { "epoch": 28.001846153846152, "grad_norm": 56.26972579956055, "learning_rate": 4.820512820512821e-06, "loss": 0.7831, "step": 3680 }, { "epoch": 28.003384615384615, "grad_norm": 101.69254302978516, "learning_rate": 4.803418803418804e-06, "loss": 0.8847, "step": 3690 }, { "epoch": 28.004923076923077, "grad_norm": 55.50261306762695, "learning_rate": 4.786324786324787e-06, "loss": 0.5746, "step": 3700 }, { "epoch": 28.00646153846154, "grad_norm": 0.04360581934452057, "learning_rate": 4.76923076923077e-06, "loss": 0.8075, "step": 3710 }, { "epoch": 28.008, "grad_norm": 46.10295104980469, "learning_rate": 4.752136752136752e-06, "loss": 0.2659, "step": 3720 }, { "epoch": 28.00953846153846, "grad_norm": 0.35408440232276917, "learning_rate": 4.7350427350427355e-06, "loss": 0.4686, "step": 3730 }, { "epoch": 28.011076923076924, "grad_norm": 0.19690901041030884, "learning_rate": 4.717948717948718e-06, "loss": 0.3786, "step": 3740 }, { "epoch": 28.012615384615383, "grad_norm": 0.33714887499809265, "learning_rate": 4.700854700854701e-06, "loss": 0.0679, "step": 3750 }, { "epoch": 28.014153846153846, "grad_norm": 646.0299072265625, "learning_rate": 4.6837606837606844e-06, "loss": 0.2962, "step": 3760 }, { "epoch": 28.01569230769231, "grad_norm": 227.01092529296875, "learning_rate": 4.666666666666667e-06, "loss": 0.996, "step": 3770 }, { "epoch": 28.017230769230768, "grad_norm": 132.82162475585938, "learning_rate": 4.64957264957265e-06, "loss": 0.4102, "step": 3780 }, { "epoch": 28.01876923076923, "grad_norm": 0.0445413812994957, "learning_rate": 4.6324786324786334e-06, "loss": 0.4506, "step": 3790 }, { "epoch": 28.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 0.8241485953330994, "eval_runtime": 13.2068, "eval_samples_per_second": 3.483, "eval_steps_per_second": 1.742, "step": 3799 }, { "epoch": 29.000153846153847, "grad_norm": 0.13771362602710724, "learning_rate": 4.615384615384616e-06, "loss": 0.4563, "step": 3800 }, { "epoch": 29.00169230769231, "grad_norm": 0.031042367219924927, "learning_rate": 4.598290598290598e-06, "loss": 0.8861, "step": 3810 }, { "epoch": 29.00323076923077, "grad_norm": 20.32213592529297, "learning_rate": 4.581196581196582e-06, "loss": 0.5048, "step": 3820 }, { "epoch": 29.00476923076923, "grad_norm": 135.0621337890625, "learning_rate": 4.564102564102564e-06, "loss": 0.5914, "step": 3830 }, { "epoch": 29.006307692307693, "grad_norm": 0.17822052538394928, "learning_rate": 4.547008547008547e-06, "loss": 1.1719, "step": 3840 }, { "epoch": 29.007846153846153, "grad_norm": 49.83673858642578, "learning_rate": 4.5299145299145306e-06, "loss": 0.5184, "step": 3850 }, { "epoch": 29.009384615384615, "grad_norm": 0.018014727160334587, "learning_rate": 4.512820512820513e-06, "loss": 0.4954, "step": 3860 }, { "epoch": 29.010923076923078, "grad_norm": 0.09828851372003555, "learning_rate": 4.495726495726496e-06, "loss": 0.9552, "step": 3870 }, { "epoch": 29.012461538461537, "grad_norm": 5.368714809417725, "learning_rate": 4.4786324786324796e-06, "loss": 0.5076, "step": 3880 }, { "epoch": 29.014, "grad_norm": 34.52876663208008, "learning_rate": 4.461538461538462e-06, "loss": 0.3437, "step": 3890 }, { "epoch": 29.015538461538462, "grad_norm": 0.25485557317733765, "learning_rate": 4.444444444444444e-06, "loss": 0.6708, "step": 3900 }, { "epoch": 29.017076923076925, "grad_norm": 0.6703222990036011, "learning_rate": 4.427350427350428e-06, "loss": 0.0826, "step": 3910 }, { "epoch": 29.018615384615384, "grad_norm": 25.28285026550293, "learning_rate": 4.4102564102564104e-06, "loss": 0.1152, "step": 3920 }, { "epoch": 29.020153846153846, "grad_norm": 0.022476214915513992, "learning_rate": 4.393162393162393e-06, "loss": 0.6501, "step": 3930 }, { "epoch": 29.020153846153846, "eval_accuracy": 0.8260869565217391, "eval_loss": 0.9973244071006775, "eval_runtime": 11.7165, "eval_samples_per_second": 3.926, "eval_steps_per_second": 1.963, "step": 3930 }, { "epoch": 30.001538461538463, "grad_norm": 1.4469164609909058, "learning_rate": 4.376068376068377e-06, "loss": 0.7975, "step": 3940 }, { "epoch": 30.00307692307692, "grad_norm": 0.03349291905760765, "learning_rate": 4.358974358974359e-06, "loss": 0.9181, "step": 3950 }, { "epoch": 30.004615384615384, "grad_norm": 1.3015611171722412, "learning_rate": 4.341880341880342e-06, "loss": 0.3424, "step": 3960 }, { "epoch": 30.006153846153847, "grad_norm": 0.5967833399772644, "learning_rate": 4.324786324786326e-06, "loss": 0.2202, "step": 3970 }, { "epoch": 30.00769230769231, "grad_norm": 0.056970831006765366, "learning_rate": 4.307692307692308e-06, "loss": 0.215, "step": 3980 }, { "epoch": 30.00923076923077, "grad_norm": 0.009819090366363525, "learning_rate": 4.29059829059829e-06, "loss": 0.287, "step": 3990 }, { "epoch": 30.01076923076923, "grad_norm": 3.7894606590270996, "learning_rate": 4.273504273504274e-06, "loss": 0.2059, "step": 4000 }, { "epoch": 30.012307692307694, "grad_norm": 0.008765391074120998, "learning_rate": 4.2564102564102566e-06, "loss": 0.1785, "step": 4010 }, { "epoch": 30.013846153846153, "grad_norm": 0.07702428847551346, "learning_rate": 4.239316239316239e-06, "loss": 0.0549, "step": 4020 }, { "epoch": 30.015384615384615, "grad_norm": 0.10615424811840057, "learning_rate": 4.222222222222223e-06, "loss": 0.3937, "step": 4030 }, { "epoch": 30.016923076923078, "grad_norm": 7.000075340270996, "learning_rate": 4.2051282051282055e-06, "loss": 0.8288, "step": 4040 }, { "epoch": 30.018461538461537, "grad_norm": 0.042215779423713684, "learning_rate": 4.188034188034188e-06, "loss": 0.7491, "step": 4050 }, { "epoch": 30.02, "grad_norm": 0.014717997051775455, "learning_rate": 4.170940170940172e-06, "loss": 0.6552, "step": 4060 }, { "epoch": 30.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.0375455617904663, "eval_runtime": 11.779, "eval_samples_per_second": 3.905, "eval_steps_per_second": 1.953, "step": 4061 }, { "epoch": 31.001384615384616, "grad_norm": 2.11873722076416, "learning_rate": 4.1538461538461545e-06, "loss": 0.0935, "step": 4070 }, { "epoch": 31.00292307692308, "grad_norm": 3.1624631881713867, "learning_rate": 4.136752136752136e-06, "loss": 0.2214, "step": 4080 }, { "epoch": 31.004461538461538, "grad_norm": 4.239480018615723, "learning_rate": 4.11965811965812e-06, "loss": 0.4094, "step": 4090 }, { "epoch": 31.006, "grad_norm": 0.017483239993453026, "learning_rate": 4.102564102564103e-06, "loss": 0.2046, "step": 4100 }, { "epoch": 31.007538461538463, "grad_norm": 104.08116149902344, "learning_rate": 4.085470085470085e-06, "loss": 0.4756, "step": 4110 }, { "epoch": 31.009076923076922, "grad_norm": 175.4596710205078, "learning_rate": 4.068376068376069e-06, "loss": 1.1802, "step": 4120 }, { "epoch": 31.010615384615384, "grad_norm": 47.34334182739258, "learning_rate": 4.051282051282052e-06, "loss": 0.0158, "step": 4130 }, { "epoch": 31.012153846153847, "grad_norm": 0.1495905965566635, "learning_rate": 4.034188034188034e-06, "loss": 0.5666, "step": 4140 }, { "epoch": 31.013692307692306, "grad_norm": 102.30220794677734, "learning_rate": 4.017094017094018e-06, "loss": 0.2828, "step": 4150 }, { "epoch": 31.01523076923077, "grad_norm": 0.18496392667293549, "learning_rate": 4.000000000000001e-06, "loss": 0.1033, "step": 4160 }, { "epoch": 31.01676923076923, "grad_norm": 24.492019653320312, "learning_rate": 3.982905982905983e-06, "loss": 0.4495, "step": 4170 }, { "epoch": 31.018307692307694, "grad_norm": 0.6317444443702698, "learning_rate": 3.965811965811966e-06, "loss": 0.2928, "step": 4180 }, { "epoch": 31.019846153846153, "grad_norm": 26.888805389404297, "learning_rate": 3.948717948717949e-06, "loss": 0.0676, "step": 4190 }, { "epoch": 31.020153846153846, "eval_accuracy": 0.8043478260869565, "eval_loss": 0.915073573589325, "eval_runtime": 12.5845, "eval_samples_per_second": 3.655, "eval_steps_per_second": 1.828, "step": 4192 }, { "epoch": 32.001230769230766, "grad_norm": 52.63165283203125, "learning_rate": 3.9316239316239315e-06, "loss": 0.9533, "step": 4200 }, { "epoch": 32.00276923076923, "grad_norm": 0.004477484151721001, "learning_rate": 3.914529914529915e-06, "loss": 0.0349, "step": 4210 }, { "epoch": 32.00430769230769, "grad_norm": 2.399238348007202, "learning_rate": 3.897435897435898e-06, "loss": 0.5123, "step": 4220 }, { "epoch": 32.00584615384616, "grad_norm": 0.004892773926258087, "learning_rate": 3.8803418803418805e-06, "loss": 0.1441, "step": 4230 }, { "epoch": 32.007384615384616, "grad_norm": 328.5547790527344, "learning_rate": 3.863247863247864e-06, "loss": 0.1297, "step": 4240 }, { "epoch": 32.008923076923075, "grad_norm": 0.249193474650383, "learning_rate": 3.846153846153847e-06, "loss": 0.0022, "step": 4250 }, { "epoch": 32.01046153846154, "grad_norm": 0.024616535753011703, "learning_rate": 3.8290598290598295e-06, "loss": 0.4523, "step": 4260 }, { "epoch": 32.012, "grad_norm": 0.03703987970948219, "learning_rate": 3.8119658119658122e-06, "loss": 0.4191, "step": 4270 }, { "epoch": 32.01353846153846, "grad_norm": 0.008937480859458447, "learning_rate": 3.794871794871795e-06, "loss": 0.0589, "step": 4280 }, { "epoch": 32.015076923076926, "grad_norm": 0.0255438182502985, "learning_rate": 3.777777777777778e-06, "loss": 0.2685, "step": 4290 }, { "epoch": 32.016615384615385, "grad_norm": 0.5818451046943665, "learning_rate": 3.760683760683761e-06, "loss": 0.2815, "step": 4300 }, { "epoch": 32.018153846153844, "grad_norm": 3.7883238792419434, "learning_rate": 3.743589743589744e-06, "loss": 0.5135, "step": 4310 }, { "epoch": 32.01969230769231, "grad_norm": 0.006536950822919607, "learning_rate": 3.726495726495727e-06, "loss": 0.0514, "step": 4320 }, { "epoch": 32.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.3664484024047852, "eval_runtime": 14.0178, "eval_samples_per_second": 3.282, "eval_steps_per_second": 1.641, "step": 4323 }, { "epoch": 33.00107692307692, "grad_norm": 0.07465940713882446, "learning_rate": 3.7094017094017098e-06, "loss": 0.0032, "step": 4330 }, { "epoch": 33.00261538461538, "grad_norm": 0.08228836953639984, "learning_rate": 3.692307692307693e-06, "loss": 0.3491, "step": 4340 }, { "epoch": 33.00415384615385, "grad_norm": 33.137882232666016, "learning_rate": 3.6752136752136756e-06, "loss": 0.8265, "step": 4350 }, { "epoch": 33.00569230769231, "grad_norm": 0.9730855822563171, "learning_rate": 3.6581196581196584e-06, "loss": 0.3718, "step": 4360 }, { "epoch": 33.007230769230766, "grad_norm": 94.9317626953125, "learning_rate": 3.641025641025641e-06, "loss": 1.0373, "step": 4370 }, { "epoch": 33.00876923076923, "grad_norm": 0.011030171066522598, "learning_rate": 3.623931623931624e-06, "loss": 0.1513, "step": 4380 }, { "epoch": 33.01030769230769, "grad_norm": 2.059166431427002, "learning_rate": 3.606837606837607e-06, "loss": 0.3012, "step": 4390 }, { "epoch": 33.01184615384616, "grad_norm": 0.2883024215698242, "learning_rate": 3.58974358974359e-06, "loss": 0.3411, "step": 4400 }, { "epoch": 33.013384615384616, "grad_norm": 47.48052215576172, "learning_rate": 3.572649572649573e-06, "loss": 0.9255, "step": 4410 }, { "epoch": 33.014923076923075, "grad_norm": 0.042310360819101334, "learning_rate": 3.555555555555556e-06, "loss": 0.0015, "step": 4420 }, { "epoch": 33.01646153846154, "grad_norm": 0.2150733917951584, "learning_rate": 3.538461538461539e-06, "loss": 0.2367, "step": 4430 }, { "epoch": 33.018, "grad_norm": 0.3859119117259979, "learning_rate": 3.5213675213675218e-06, "loss": 0.1022, "step": 4440 }, { "epoch": 33.01953846153846, "grad_norm": 0.15183915197849274, "learning_rate": 3.5042735042735045e-06, "loss": 0.5447, "step": 4450 }, { "epoch": 33.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.4551424980163574, "eval_runtime": 13.2027, "eval_samples_per_second": 3.484, "eval_steps_per_second": 1.742, "step": 4454 }, { "epoch": 34.00092307692308, "grad_norm": 0.4627760350704193, "learning_rate": 3.487179487179487e-06, "loss": 0.9165, "step": 4460 }, { "epoch": 34.00246153846154, "grad_norm": 110.15799713134766, "learning_rate": 3.4700854700854703e-06, "loss": 0.3152, "step": 4470 }, { "epoch": 34.004, "grad_norm": 199.7411346435547, "learning_rate": 3.452991452991453e-06, "loss": 0.4348, "step": 4480 }, { "epoch": 34.005538461538464, "grad_norm": 0.009803686290979385, "learning_rate": 3.435897435897436e-06, "loss": 0.0009, "step": 4490 }, { "epoch": 34.00707692307692, "grad_norm": 269.10552978515625, "learning_rate": 3.4188034188034193e-06, "loss": 0.4589, "step": 4500 }, { "epoch": 34.00861538461538, "grad_norm": 0.06494735926389694, "learning_rate": 3.401709401709402e-06, "loss": 0.3366, "step": 4510 }, { "epoch": 34.01015384615385, "grad_norm": 214.33226013183594, "learning_rate": 3.384615384615385e-06, "loss": 0.2165, "step": 4520 }, { "epoch": 34.01169230769231, "grad_norm": 43.572837829589844, "learning_rate": 3.367521367521368e-06, "loss": 0.7485, "step": 4530 }, { "epoch": 34.013230769230766, "grad_norm": 0.27556538581848145, "learning_rate": 3.3504273504273506e-06, "loss": 0.5172, "step": 4540 }, { "epoch": 34.01476923076923, "grad_norm": 0.007643167860805988, "learning_rate": 3.3333333333333333e-06, "loss": 0.4093, "step": 4550 }, { "epoch": 34.01630769230769, "grad_norm": 63.7769660949707, "learning_rate": 3.3162393162393165e-06, "loss": 0.3615, "step": 4560 }, { "epoch": 34.01784615384615, "grad_norm": 0.06752733886241913, "learning_rate": 3.299145299145299e-06, "loss": 0.357, "step": 4570 }, { "epoch": 34.01938461538462, "grad_norm": 0.05651543661952019, "learning_rate": 3.2820512820512823e-06, "loss": 0.3701, "step": 4580 }, { "epoch": 34.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.3387607336044312, "eval_runtime": 13.1827, "eval_samples_per_second": 3.489, "eval_steps_per_second": 1.745, "step": 4585 }, { "epoch": 35.00076923076923, "grad_norm": 0.06349892169237137, "learning_rate": 3.2649572649572655e-06, "loss": 0.0061, "step": 4590 }, { "epoch": 35.002307692307696, "grad_norm": 0.1903676837682724, "learning_rate": 3.247863247863248e-06, "loss": 0.0449, "step": 4600 }, { "epoch": 35.003846153846155, "grad_norm": 13.700048446655273, "learning_rate": 3.2307692307692313e-06, "loss": 0.3592, "step": 4610 }, { "epoch": 35.005384615384614, "grad_norm": 0.04581848159432411, "learning_rate": 3.213675213675214e-06, "loss": 0.4179, "step": 4620 }, { "epoch": 35.00692307692308, "grad_norm": 165.4931640625, "learning_rate": 3.1965811965811967e-06, "loss": 0.3106, "step": 4630 }, { "epoch": 35.00846153846154, "grad_norm": 0.022043738514184952, "learning_rate": 3.1794871794871795e-06, "loss": 0.4508, "step": 4640 }, { "epoch": 35.01, "grad_norm": 0.03153197467327118, "learning_rate": 3.1623931623931626e-06, "loss": 0.0166, "step": 4650 }, { "epoch": 35.011538461538464, "grad_norm": 3.3061561584472656, "learning_rate": 3.1452991452991453e-06, "loss": 0.7242, "step": 4660 }, { "epoch": 35.01307692307692, "grad_norm": 0.6018900871276855, "learning_rate": 3.1282051282051284e-06, "loss": 0.1013, "step": 4670 }, { "epoch": 35.01461538461538, "grad_norm": 18.714059829711914, "learning_rate": 3.1111111111111116e-06, "loss": 0.0358, "step": 4680 }, { "epoch": 35.01615384615385, "grad_norm": 290.2351989746094, "learning_rate": 3.0940170940170943e-06, "loss": 0.1013, "step": 4690 }, { "epoch": 35.01769230769231, "grad_norm": 0.00717718992382288, "learning_rate": 3.0769230769230774e-06, "loss": 0.3149, "step": 4700 }, { "epoch": 35.01923076923077, "grad_norm": 0.06576789915561676, "learning_rate": 3.05982905982906e-06, "loss": 0.6869, "step": 4710 }, { "epoch": 35.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.3302055597305298, "eval_runtime": 13.1813, "eval_samples_per_second": 3.49, "eval_steps_per_second": 1.745, "step": 4716 }, { "epoch": 36.00061538461539, "grad_norm": 0.015848729759454727, "learning_rate": 3.042735042735043e-06, "loss": 0.0065, "step": 4720 }, { "epoch": 36.002153846153846, "grad_norm": 0.10322362929582596, "learning_rate": 3.0256410256410256e-06, "loss": 0.2323, "step": 4730 }, { "epoch": 36.003692307692305, "grad_norm": 0.03065655007958412, "learning_rate": 3.0085470085470087e-06, "loss": 0.1782, "step": 4740 }, { "epoch": 36.00523076923077, "grad_norm": 0.0023964950814843178, "learning_rate": 2.9914529914529914e-06, "loss": 0.3799, "step": 4750 }, { "epoch": 36.00676923076923, "grad_norm": 203.9849090576172, "learning_rate": 2.9743589743589746e-06, "loss": 0.2273, "step": 4760 }, { "epoch": 36.00830769230769, "grad_norm": 0.5617519617080688, "learning_rate": 2.9572649572649577e-06, "loss": 0.2777, "step": 4770 }, { "epoch": 36.009846153846155, "grad_norm": 0.004091217648237944, "learning_rate": 2.9401709401709404e-06, "loss": 0.1771, "step": 4780 }, { "epoch": 36.011384615384614, "grad_norm": 0.03037184663116932, "learning_rate": 2.9230769230769236e-06, "loss": 0.1418, "step": 4790 }, { "epoch": 36.01292307692308, "grad_norm": 0.013294171541929245, "learning_rate": 2.9059829059829063e-06, "loss": 0.4339, "step": 4800 }, { "epoch": 36.01446153846154, "grad_norm": 0.0023209303617477417, "learning_rate": 2.888888888888889e-06, "loss": 0.2652, "step": 4810 }, { "epoch": 36.016, "grad_norm": 63.16322708129883, "learning_rate": 2.8717948717948717e-06, "loss": 0.5442, "step": 4820 }, { "epoch": 36.017538461538464, "grad_norm": 436.7990417480469, "learning_rate": 2.854700854700855e-06, "loss": 0.5163, "step": 4830 }, { "epoch": 36.01907692307692, "grad_norm": 0.8330660462379456, "learning_rate": 2.8376068376068376e-06, "loss": 0.1371, "step": 4840 }, { "epoch": 36.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.4859415292739868, "eval_runtime": 12.7482, "eval_samples_per_second": 3.608, "eval_steps_per_second": 1.804, "step": 4847 }, { "epoch": 37.000461538461536, "grad_norm": 64.5765151977539, "learning_rate": 2.8205128205128207e-06, "loss": 0.3055, "step": 4850 }, { "epoch": 37.002, "grad_norm": 0.06066625565290451, "learning_rate": 2.803418803418804e-06, "loss": 0.0016, "step": 4860 }, { "epoch": 37.00353846153846, "grad_norm": 0.007019346114248037, "learning_rate": 2.7863247863247866e-06, "loss": 0.5366, "step": 4870 }, { "epoch": 37.00507692307692, "grad_norm": 0.008301621302962303, "learning_rate": 2.7692307692307697e-06, "loss": 0.4092, "step": 4880 }, { "epoch": 37.00661538461539, "grad_norm": 0.020703766494989395, "learning_rate": 2.7521367521367524e-06, "loss": 0.2475, "step": 4890 }, { "epoch": 37.008153846153846, "grad_norm": 2.0971076488494873, "learning_rate": 2.7350427350427355e-06, "loss": 0.0706, "step": 4900 }, { "epoch": 37.009692307692305, "grad_norm": 9.441088676452637, "learning_rate": 2.717948717948718e-06, "loss": 0.0394, "step": 4910 }, { "epoch": 37.01123076923077, "grad_norm": 385.3316345214844, "learning_rate": 2.700854700854701e-06, "loss": 0.2511, "step": 4920 }, { "epoch": 37.01276923076923, "grad_norm": 0.006842486094683409, "learning_rate": 2.6837606837606837e-06, "loss": 0.7641, "step": 4930 }, { "epoch": 37.01430769230769, "grad_norm": 0.004666823893785477, "learning_rate": 2.666666666666667e-06, "loss": 0.5452, "step": 4940 }, { "epoch": 37.015846153846155, "grad_norm": 0.49463582038879395, "learning_rate": 2.64957264957265e-06, "loss": 0.2126, "step": 4950 }, { "epoch": 37.017384615384614, "grad_norm": 5.189666271209717, "learning_rate": 2.6324786324786327e-06, "loss": 0.2353, "step": 4960 }, { "epoch": 37.01892307692308, "grad_norm": 0.033904653042554855, "learning_rate": 2.615384615384616e-06, "loss": 0.5565, "step": 4970 }, { "epoch": 37.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.410302996635437, "eval_runtime": 12.6149, "eval_samples_per_second": 3.646, "eval_steps_per_second": 1.823, "step": 4978 }, { "epoch": 38.00030769230769, "grad_norm": 0.08889954537153244, "learning_rate": 2.5982905982905985e-06, "loss": 0.5995, "step": 4980 }, { "epoch": 38.00184615384615, "grad_norm": 0.17473474144935608, "learning_rate": 2.5811965811965817e-06, "loss": 0.0306, "step": 4990 }, { "epoch": 38.00338461538462, "grad_norm": 209.5628204345703, "learning_rate": 2.564102564102564e-06, "loss": 0.3034, "step": 5000 }, { "epoch": 38.00492307692308, "grad_norm": 0.004574312828481197, "learning_rate": 2.547008547008547e-06, "loss": 0.3718, "step": 5010 }, { "epoch": 38.00646153846154, "grad_norm": 0.0036003710702061653, "learning_rate": 2.52991452991453e-06, "loss": 0.3068, "step": 5020 }, { "epoch": 38.008, "grad_norm": 0.2461969256401062, "learning_rate": 2.512820512820513e-06, "loss": 0.6456, "step": 5030 }, { "epoch": 38.00953846153846, "grad_norm": 0.10105012357234955, "learning_rate": 2.495726495726496e-06, "loss": 0.2028, "step": 5040 }, { "epoch": 38.01107692307692, "grad_norm": 35.23940658569336, "learning_rate": 2.478632478632479e-06, "loss": 0.3564, "step": 5050 }, { "epoch": 38.01261538461539, "grad_norm": 2.9968295097351074, "learning_rate": 2.461538461538462e-06, "loss": 0.3022, "step": 5060 }, { "epoch": 38.014153846153846, "grad_norm": 0.0421447791159153, "learning_rate": 2.4444444444444447e-06, "loss": 0.001, "step": 5070 }, { "epoch": 38.015692307692305, "grad_norm": 0.06525139510631561, "learning_rate": 2.4273504273504274e-06, "loss": 0.4745, "step": 5080 }, { "epoch": 38.01723076923077, "grad_norm": 0.09138324856758118, "learning_rate": 2.4102564102564105e-06, "loss": 0.0038, "step": 5090 }, { "epoch": 38.01876923076923, "grad_norm": 0.06561964750289917, "learning_rate": 2.3931623931623937e-06, "loss": 0.0022, "step": 5100 }, { "epoch": 38.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.3428059816360474, "eval_runtime": 13.1706, "eval_samples_per_second": 3.493, "eval_steps_per_second": 1.746, "step": 5109 }, { "epoch": 39.00015384615384, "grad_norm": 0.08478695154190063, "learning_rate": 2.376068376068376e-06, "loss": 0.2608, "step": 5110 }, { "epoch": 39.00169230769231, "grad_norm": 0.8815059065818787, "learning_rate": 2.358974358974359e-06, "loss": 0.0008, "step": 5120 }, { "epoch": 39.00323076923077, "grad_norm": 0.08619926869869232, "learning_rate": 2.3418803418803422e-06, "loss": 0.0012, "step": 5130 }, { "epoch": 39.00476923076923, "grad_norm": 0.015669770538806915, "learning_rate": 2.324786324786325e-06, "loss": 0.0009, "step": 5140 }, { "epoch": 39.00630769230769, "grad_norm": 0.07468448579311371, "learning_rate": 2.307692307692308e-06, "loss": 0.003, "step": 5150 }, { "epoch": 39.00784615384615, "grad_norm": 0.09687216579914093, "learning_rate": 2.290598290598291e-06, "loss": 0.0039, "step": 5160 }, { "epoch": 39.00938461538462, "grad_norm": 0.6945201754570007, "learning_rate": 2.2735042735042735e-06, "loss": 0.355, "step": 5170 }, { "epoch": 39.01092307692308, "grad_norm": 0.0034933302085846663, "learning_rate": 2.2564102564102566e-06, "loss": 0.9715, "step": 5180 }, { "epoch": 39.01246153846154, "grad_norm": 0.003874710062518716, "learning_rate": 2.2393162393162398e-06, "loss": 0.0534, "step": 5190 }, { "epoch": 39.014, "grad_norm": 60.87498474121094, "learning_rate": 2.222222222222222e-06, "loss": 0.3278, "step": 5200 }, { "epoch": 39.01553846153846, "grad_norm": 0.1837165504693985, "learning_rate": 2.2051282051282052e-06, "loss": 0.5495, "step": 5210 }, { "epoch": 39.01707692307692, "grad_norm": 0.1290096938610077, "learning_rate": 2.1880341880341884e-06, "loss": 0.0861, "step": 5220 }, { "epoch": 39.01861538461539, "grad_norm": 0.02415439300239086, "learning_rate": 2.170940170940171e-06, "loss": 0.1622, "step": 5230 }, { "epoch": 39.020153846153846, "grad_norm": 65.6265640258789, "learning_rate": 2.153846153846154e-06, "loss": 1.5897, "step": 5240 }, { "epoch": 39.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.4124759435653687, "eval_runtime": 11.8096, "eval_samples_per_second": 3.895, "eval_steps_per_second": 1.948, "step": 5240 }, { "epoch": 40.00153846153846, "grad_norm": 0.01301716361194849, "learning_rate": 2.136752136752137e-06, "loss": 0.429, "step": 5250 }, { "epoch": 40.003076923076925, "grad_norm": 0.023441065102815628, "learning_rate": 2.1196581196581196e-06, "loss": 0.4187, "step": 5260 }, { "epoch": 40.004615384615384, "grad_norm": 0.004829849116504192, "learning_rate": 2.1025641025641028e-06, "loss": 0.0711, "step": 5270 }, { "epoch": 40.00615384615384, "grad_norm": 0.045608654618263245, "learning_rate": 2.085470085470086e-06, "loss": 0.0636, "step": 5280 }, { "epoch": 40.00769230769231, "grad_norm": 322.0433044433594, "learning_rate": 2.068376068376068e-06, "loss": 0.2149, "step": 5290 }, { "epoch": 40.00923076923077, "grad_norm": 42.855224609375, "learning_rate": 2.0512820512820513e-06, "loss": 0.4352, "step": 5300 }, { "epoch": 40.01076923076923, "grad_norm": 0.002952966606244445, "learning_rate": 2.0341880341880345e-06, "loss": 0.6437, "step": 5310 }, { "epoch": 40.012307692307694, "grad_norm": 0.003094248939305544, "learning_rate": 2.017094017094017e-06, "loss": 0.0081, "step": 5320 }, { "epoch": 40.01384615384615, "grad_norm": 0.003211932722479105, "learning_rate": 2.0000000000000003e-06, "loss": 0.586, "step": 5330 }, { "epoch": 40.01538461538462, "grad_norm": 0.005071121733635664, "learning_rate": 1.982905982905983e-06, "loss": 0.0073, "step": 5340 }, { "epoch": 40.01692307692308, "grad_norm": 0.08057103306055069, "learning_rate": 1.9658119658119658e-06, "loss": 0.0025, "step": 5350 }, { "epoch": 40.01846153846154, "grad_norm": 0.005270438268780708, "learning_rate": 1.948717948717949e-06, "loss": 0.0004, "step": 5360 }, { "epoch": 40.02, "grad_norm": 0.008296512998640537, "learning_rate": 1.931623931623932e-06, "loss": 0.1302, "step": 5370 }, { "epoch": 40.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.3892539739608765, "eval_runtime": 12.4205, "eval_samples_per_second": 3.704, "eval_steps_per_second": 1.852, "step": 5371 }, { "epoch": 41.001384615384616, "grad_norm": 0.0030074138194322586, "learning_rate": 1.9145299145299148e-06, "loss": 0.1594, "step": 5380 }, { "epoch": 41.002923076923075, "grad_norm": 0.04759305715560913, "learning_rate": 1.8974358974358975e-06, "loss": 0.0777, "step": 5390 }, { "epoch": 41.00446153846154, "grad_norm": 0.020955096930265427, "learning_rate": 1.8803418803418804e-06, "loss": 0.0888, "step": 5400 }, { "epoch": 41.006, "grad_norm": 0.020802758634090424, "learning_rate": 1.8632478632478635e-06, "loss": 0.0009, "step": 5410 }, { "epoch": 41.00753846153846, "grad_norm": 196.2333221435547, "learning_rate": 1.8461538461538465e-06, "loss": 0.4222, "step": 5420 }, { "epoch": 41.009076923076925, "grad_norm": 0.2782272398471832, "learning_rate": 1.8290598290598292e-06, "loss": 0.3757, "step": 5430 }, { "epoch": 41.010615384615384, "grad_norm": 0.020732909440994263, "learning_rate": 1.811965811965812e-06, "loss": 0.0009, "step": 5440 }, { "epoch": 41.01215384615384, "grad_norm": 0.012637576088309288, "learning_rate": 1.794871794871795e-06, "loss": 0.4355, "step": 5450 }, { "epoch": 41.01369230769231, "grad_norm": 0.011620227247476578, "learning_rate": 1.777777777777778e-06, "loss": 0.399, "step": 5460 }, { "epoch": 41.01523076923077, "grad_norm": 0.025496402755379677, "learning_rate": 1.7606837606837609e-06, "loss": 0.1318, "step": 5470 }, { "epoch": 41.01676923076923, "grad_norm": 94.95100402832031, "learning_rate": 1.7435897435897436e-06, "loss": 0.3159, "step": 5480 }, { "epoch": 41.018307692307694, "grad_norm": 0.048655539751052856, "learning_rate": 1.7264957264957265e-06, "loss": 0.2638, "step": 5490 }, { "epoch": 41.01984615384615, "grad_norm": 0.004611465148627758, "learning_rate": 1.7094017094017097e-06, "loss": 0.2463, "step": 5500 }, { "epoch": 41.020153846153846, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.559918761253357, "eval_runtime": 12.4494, "eval_samples_per_second": 3.695, "eval_steps_per_second": 1.847, "step": 5502 }, { "epoch": 42.001230769230766, "grad_norm": 0.023069551214575768, "learning_rate": 1.6923076923076926e-06, "loss": 0.8881, "step": 5510 }, { "epoch": 42.00276923076923, "grad_norm": 0.006624761037528515, "learning_rate": 1.6752136752136753e-06, "loss": 0.2337, "step": 5520 }, { "epoch": 42.00430769230769, "grad_norm": 96.46836853027344, "learning_rate": 1.6581196581196582e-06, "loss": 0.3171, "step": 5530 }, { "epoch": 42.00584615384616, "grad_norm": 317.9783935546875, "learning_rate": 1.6410256410256412e-06, "loss": 0.4246, "step": 5540 }, { "epoch": 42.007384615384616, "grad_norm": 0.0062882499769330025, "learning_rate": 1.623931623931624e-06, "loss": 0.2339, "step": 5550 }, { "epoch": 42.008923076923075, "grad_norm": 0.005598544143140316, "learning_rate": 1.606837606837607e-06, "loss": 0.4042, "step": 5560 }, { "epoch": 42.01046153846154, "grad_norm": 0.030421020463109016, "learning_rate": 1.5897435897435897e-06, "loss": 0.0007, "step": 5570 }, { "epoch": 42.012, "grad_norm": 0.006722313351929188, "learning_rate": 1.5726495726495727e-06, "loss": 0.6109, "step": 5580 }, { "epoch": 42.01353846153846, "grad_norm": 0.05044174566864967, "learning_rate": 1.5555555555555558e-06, "loss": 0.7525, "step": 5590 }, { "epoch": 42.015076923076926, "grad_norm": 0.08396083116531372, "learning_rate": 1.5384615384615387e-06, "loss": 0.0008, "step": 5600 }, { "epoch": 42.016615384615385, "grad_norm": 0.0033397823572158813, "learning_rate": 1.5213675213675214e-06, "loss": 0.1934, "step": 5610 }, { "epoch": 42.018153846153844, "grad_norm": 0.5943049192428589, "learning_rate": 1.5042735042735044e-06, "loss": 0.0032, "step": 5620 }, { "epoch": 42.01969230769231, "grad_norm": 0.5589593648910522, "learning_rate": 1.4871794871794873e-06, "loss": 0.0007, "step": 5630 }, { "epoch": 42.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.3541553020477295, "eval_runtime": 11.8006, "eval_samples_per_second": 3.898, "eval_steps_per_second": 1.949, "step": 5633 }, { "epoch": 43.00107692307692, "grad_norm": 0.016454918310046196, "learning_rate": 1.4700854700854702e-06, "loss": 0.0767, "step": 5640 }, { "epoch": 43.00261538461538, "grad_norm": 0.4291391968727112, "learning_rate": 1.4529914529914531e-06, "loss": 0.1525, "step": 5650 }, { "epoch": 43.00415384615385, "grad_norm": 2.03802490234375, "learning_rate": 1.4358974358974359e-06, "loss": 0.0424, "step": 5660 }, { "epoch": 43.00569230769231, "grad_norm": 0.0391116738319397, "learning_rate": 1.4188034188034188e-06, "loss": 0.0007, "step": 5670 }, { "epoch": 43.007230769230766, "grad_norm": 0.04126261919736862, "learning_rate": 1.401709401709402e-06, "loss": 0.0006, "step": 5680 }, { "epoch": 43.00876923076923, "grad_norm": 0.15058164298534393, "learning_rate": 1.3846153846153848e-06, "loss": 1.0264, "step": 5690 }, { "epoch": 43.01030769230769, "grad_norm": 25.521728515625, "learning_rate": 1.3675213675213678e-06, "loss": 0.0045, "step": 5700 }, { "epoch": 43.01184615384616, "grad_norm": 0.009670182131230831, "learning_rate": 1.3504273504273505e-06, "loss": 0.0003, "step": 5710 }, { "epoch": 43.013384615384616, "grad_norm": 0.0412202887237072, "learning_rate": 1.3333333333333334e-06, "loss": 0.0057, "step": 5720 }, { "epoch": 43.014923076923075, "grad_norm": 108.24889373779297, "learning_rate": 1.3162393162393163e-06, "loss": 0.4326, "step": 5730 }, { "epoch": 43.01646153846154, "grad_norm": 0.012433959171175957, "learning_rate": 1.2991452991452993e-06, "loss": 0.2995, "step": 5740 }, { "epoch": 43.018, "grad_norm": 0.015085268765687943, "learning_rate": 1.282051282051282e-06, "loss": 0.0008, "step": 5750 }, { "epoch": 43.01953846153846, "grad_norm": 0.08252725005149841, "learning_rate": 1.264957264957265e-06, "loss": 0.0004, "step": 5760 }, { "epoch": 43.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.8062047958374023, "eval_runtime": 11.7555, "eval_samples_per_second": 3.913, "eval_steps_per_second": 1.957, "step": 5764 }, { "epoch": 44.00092307692308, "grad_norm": 325.7167053222656, "learning_rate": 1.247863247863248e-06, "loss": 0.8863, "step": 5770 }, { "epoch": 44.00246153846154, "grad_norm": 0.023112991824746132, "learning_rate": 1.230769230769231e-06, "loss": 0.0018, "step": 5780 }, { "epoch": 44.004, "grad_norm": 0.03161883354187012, "learning_rate": 1.2136752136752137e-06, "loss": 0.5382, "step": 5790 }, { "epoch": 44.005538461538464, "grad_norm": 0.226994588971138, "learning_rate": 1.1965811965811968e-06, "loss": 0.0069, "step": 5800 }, { "epoch": 44.00707692307692, "grad_norm": 0.38113123178482056, "learning_rate": 1.1794871794871795e-06, "loss": 0.0085, "step": 5810 }, { "epoch": 44.00861538461538, "grad_norm": 0.032291267067193985, "learning_rate": 1.1623931623931625e-06, "loss": 0.0003, "step": 5820 }, { "epoch": 44.01015384615385, "grad_norm": 0.1154903769493103, "learning_rate": 1.1452991452991454e-06, "loss": 0.0081, "step": 5830 }, { "epoch": 44.01169230769231, "grad_norm": 0.05736929178237915, "learning_rate": 1.1282051282051283e-06, "loss": 1.1695, "step": 5840 }, { "epoch": 44.013230769230766, "grad_norm": 0.016687337309122086, "learning_rate": 1.111111111111111e-06, "loss": 0.0583, "step": 5850 }, { "epoch": 44.01476923076923, "grad_norm": 0.0028132584411650896, "learning_rate": 1.0940170940170942e-06, "loss": 0.129, "step": 5860 }, { "epoch": 44.01630769230769, "grad_norm": 0.007681978866457939, "learning_rate": 1.076923076923077e-06, "loss": 0.0005, "step": 5870 }, { "epoch": 44.01784615384615, "grad_norm": 4.105467796325684, "learning_rate": 1.0598290598290598e-06, "loss": 0.0324, "step": 5880 }, { "epoch": 44.01938461538462, "grad_norm": 0.0026431982405483723, "learning_rate": 1.042735042735043e-06, "loss": 0.0803, "step": 5890 }, { "epoch": 44.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.5013920068740845, "eval_runtime": 11.8321, "eval_samples_per_second": 3.888, "eval_steps_per_second": 1.944, "step": 5895 }, { "epoch": 45.00076923076923, "grad_norm": 0.15906000137329102, "learning_rate": 1.0256410256410257e-06, "loss": 0.0844, "step": 5900 }, { "epoch": 45.002307692307696, "grad_norm": 0.07680743932723999, "learning_rate": 1.0085470085470086e-06, "loss": 0.0004, "step": 5910 }, { "epoch": 45.003846153846155, "grad_norm": 0.0016978027997538447, "learning_rate": 9.914529914529915e-07, "loss": 0.0031, "step": 5920 }, { "epoch": 45.005384615384614, "grad_norm": 0.04445165395736694, "learning_rate": 9.743589743589745e-07, "loss": 0.319, "step": 5930 }, { "epoch": 45.00692307692308, "grad_norm": 0.005968225654214621, "learning_rate": 9.572649572649574e-07, "loss": 0.3625, "step": 5940 }, { "epoch": 45.00846153846154, "grad_norm": 0.010701755993068218, "learning_rate": 9.401709401709402e-07, "loss": 0.0003, "step": 5950 }, { "epoch": 45.01, "grad_norm": 0.0020827685948461294, "learning_rate": 9.230769230769232e-07, "loss": 0.0168, "step": 5960 }, { "epoch": 45.011538461538464, "grad_norm": 0.03637014701962471, "learning_rate": 9.05982905982906e-07, "loss": 0.0005, "step": 5970 }, { "epoch": 45.01307692307692, "grad_norm": 0.002841139445081353, "learning_rate": 8.88888888888889e-07, "loss": 0.0003, "step": 5980 }, { "epoch": 45.01461538461538, "grad_norm": 0.05124405771493912, "learning_rate": 8.717948717948718e-07, "loss": 0.0692, "step": 5990 }, { "epoch": 45.01615384615385, "grad_norm": 0.0027790500316768885, "learning_rate": 8.547008547008548e-07, "loss": 0.0007, "step": 6000 }, { "epoch": 45.01769230769231, "grad_norm": 9.085582733154297, "learning_rate": 8.376068376068377e-07, "loss": 0.1762, "step": 6010 }, { "epoch": 45.01923076923077, "grad_norm": 0.016285724937915802, "learning_rate": 8.205128205128206e-07, "loss": 0.3332, "step": 6020 }, { "epoch": 45.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.5840011835098267, "eval_runtime": 11.9274, "eval_samples_per_second": 3.857, "eval_steps_per_second": 1.928, "step": 6026 }, { "epoch": 46.00061538461539, "grad_norm": 0.02442973293364048, "learning_rate": 8.034188034188035e-07, "loss": 0.5798, "step": 6030 }, { "epoch": 46.002153846153846, "grad_norm": 0.011202111840248108, "learning_rate": 7.863247863247863e-07, "loss": 0.0004, "step": 6040 }, { "epoch": 46.003692307692305, "grad_norm": 0.04835433140397072, "learning_rate": 7.692307692307694e-07, "loss": 0.0371, "step": 6050 }, { "epoch": 46.00523076923077, "grad_norm": 0.036281924694776535, "learning_rate": 7.521367521367522e-07, "loss": 0.0007, "step": 6060 }, { "epoch": 46.00676923076923, "grad_norm": 0.00819159671664238, "learning_rate": 7.350427350427351e-07, "loss": 0.0002, "step": 6070 }, { "epoch": 46.00830769230769, "grad_norm": 0.058177947998046875, "learning_rate": 7.179487179487179e-07, "loss": 0.2703, "step": 6080 }, { "epoch": 46.009846153846155, "grad_norm": 0.005668714642524719, "learning_rate": 7.00854700854701e-07, "loss": 0.2731, "step": 6090 }, { "epoch": 46.011384615384614, "grad_norm": 0.002219714689999819, "learning_rate": 6.837606837606839e-07, "loss": 0.0002, "step": 6100 }, { "epoch": 46.01292307692308, "grad_norm": 0.028290949761867523, "learning_rate": 6.666666666666667e-07, "loss": 0.0008, "step": 6110 }, { "epoch": 46.01446153846154, "grad_norm": 0.06725189834833145, "learning_rate": 6.495726495726496e-07, "loss": 0.4295, "step": 6120 }, { "epoch": 46.016, "grad_norm": 0.0370306521654129, "learning_rate": 6.324786324786325e-07, "loss": 0.2355, "step": 6130 }, { "epoch": 46.017538461538464, "grad_norm": 0.005414789542555809, "learning_rate": 6.153846153846155e-07, "loss": 0.2553, "step": 6140 }, { "epoch": 46.01907692307692, "grad_norm": 0.6177451610565186, "learning_rate": 5.982905982905984e-07, "loss": 0.0004, "step": 6150 }, { "epoch": 46.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.5470941066741943, "eval_runtime": 11.9254, "eval_samples_per_second": 3.857, "eval_steps_per_second": 1.929, "step": 6157 }, { "epoch": 47.000461538461536, "grad_norm": 0.36279433965682983, "learning_rate": 5.811965811965812e-07, "loss": 0.1052, "step": 6160 }, { "epoch": 47.002, "grad_norm": 0.010881070047616959, "learning_rate": 5.641025641025642e-07, "loss": 0.0509, "step": 6170 }, { "epoch": 47.00353846153846, "grad_norm": 0.0020965421572327614, "learning_rate": 5.470085470085471e-07, "loss": 0.4284, "step": 6180 }, { "epoch": 47.00507692307692, "grad_norm": 0.03328271955251694, "learning_rate": 5.299145299145299e-07, "loss": 0.0003, "step": 6190 }, { "epoch": 47.00661538461539, "grad_norm": 0.014189296402037144, "learning_rate": 5.128205128205128e-07, "loss": 0.0007, "step": 6200 }, { "epoch": 47.008153846153846, "grad_norm": 0.009917080402374268, "learning_rate": 4.957264957264958e-07, "loss": 0.343, "step": 6210 }, { "epoch": 47.009692307692305, "grad_norm": 0.016622619703412056, "learning_rate": 4.786324786324787e-07, "loss": 0.0019, "step": 6220 }, { "epoch": 47.01123076923077, "grad_norm": 0.01013423502445221, "learning_rate": 4.615384615384616e-07, "loss": 0.4385, "step": 6230 }, { "epoch": 47.01276923076923, "grad_norm": 0.0019633290357887745, "learning_rate": 4.444444444444445e-07, "loss": 0.2154, "step": 6240 }, { "epoch": 47.01430769230769, "grad_norm": 119.6874771118164, "learning_rate": 4.273504273504274e-07, "loss": 0.0332, "step": 6250 }, { "epoch": 47.015846153846155, "grad_norm": 0.03719891607761383, "learning_rate": 4.102564102564103e-07, "loss": 0.0003, "step": 6260 }, { "epoch": 47.017384615384614, "grad_norm": 1.4264271259307861, "learning_rate": 3.9316239316239316e-07, "loss": 0.2852, "step": 6270 }, { "epoch": 47.01892307692308, "grad_norm": 0.01973658986389637, "learning_rate": 3.760683760683761e-07, "loss": 0.0004, "step": 6280 }, { "epoch": 47.020153846153846, "eval_accuracy": 0.782608695652174, "eval_loss": 1.398191213607788, "eval_runtime": 13.3455, "eval_samples_per_second": 3.447, "eval_steps_per_second": 1.723, "step": 6288 }, { "epoch": 48.00030769230769, "grad_norm": 0.0016242864076048136, "learning_rate": 3.5897435897435896e-07, "loss": 0.7016, "step": 6290 }, { "epoch": 48.00184615384615, "grad_norm": 0.015563036315143108, "learning_rate": 3.4188034188034194e-07, "loss": 0.5297, "step": 6300 }, { "epoch": 48.00338461538462, "grad_norm": 0.20225787162780762, "learning_rate": 3.247863247863248e-07, "loss": 0.0009, "step": 6310 }, { "epoch": 48.00492307692308, "grad_norm": 0.008095812052488327, "learning_rate": 3.0769230769230774e-07, "loss": 0.0005, "step": 6320 }, { "epoch": 48.00646153846154, "grad_norm": 0.0020545681472867727, "learning_rate": 2.905982905982906e-07, "loss": 0.1397, "step": 6330 }, { "epoch": 48.008, "grad_norm": 0.03401735797524452, "learning_rate": 2.7350427350427354e-07, "loss": 0.0047, "step": 6340 }, { "epoch": 48.00953846153846, "grad_norm": 1.6771128177642822, "learning_rate": 2.564102564102564e-07, "loss": 0.2786, "step": 6350 }, { "epoch": 48.01107692307692, "grad_norm": 0.0300904531031847, "learning_rate": 2.3931623931623934e-07, "loss": 0.1159, "step": 6360 }, { "epoch": 48.01261538461539, "grad_norm": 0.002180424053221941, "learning_rate": 2.2222222222222224e-07, "loss": 0.2689, "step": 6370 }, { "epoch": 48.014153846153846, "grad_norm": 0.004255858715623617, "learning_rate": 2.0512820512820514e-07, "loss": 0.0019, "step": 6380 }, { "epoch": 48.015692307692305, "grad_norm": 158.4202117919922, "learning_rate": 1.8803418803418804e-07, "loss": 0.0496, "step": 6390 }, { "epoch": 48.01723076923077, "grad_norm": 0.05164070799946785, "learning_rate": 1.7094017094017097e-07, "loss": 0.0574, "step": 6400 }, { "epoch": 48.01876923076923, "grad_norm": 0.09816153347492218, "learning_rate": 1.5384615384615387e-07, "loss": 0.0019, "step": 6410 }, { "epoch": 48.020153846153846, "eval_accuracy": 0.7391304347826086, "eval_loss": 1.488898515701294, "eval_runtime": 13.5505, "eval_samples_per_second": 3.395, "eval_steps_per_second": 1.697, "step": 6419 }, { "epoch": 49.00015384615384, "grad_norm": 0.028583820909261703, "learning_rate": 1.3675213675213677e-07, "loss": 0.0017, "step": 6420 }, { "epoch": 49.00169230769231, "grad_norm": 0.4499860107898712, "learning_rate": 1.1965811965811967e-07, "loss": 0.0003, "step": 6430 }, { "epoch": 49.00323076923077, "grad_norm": 6.271653652191162, "learning_rate": 1.0256410256410257e-07, "loss": 0.2985, "step": 6440 }, { "epoch": 49.00476923076923, "grad_norm": 0.0023325886577367783, "learning_rate": 8.547008547008549e-08, "loss": 0.0946, "step": 6450 }, { "epoch": 49.00630769230769, "grad_norm": 476.3946228027344, "learning_rate": 6.837606837606839e-08, "loss": 0.1531, "step": 6460 }, { "epoch": 49.00784615384615, "grad_norm": 0.0066766985692083836, "learning_rate": 5.1282051282051286e-08, "loss": 0.0096, "step": 6470 }, { "epoch": 49.00938461538462, "grad_norm": 14.471455574035645, "learning_rate": 3.418803418803419e-08, "loss": 0.0033, "step": 6480 }, { "epoch": 49.01092307692308, "grad_norm": 0.01039290800690651, "learning_rate": 1.7094017094017096e-08, "loss": 0.0098, "step": 6490 }, { "epoch": 49.01246153846154, "grad_norm": 0.005664715077728033, "learning_rate": 0.0, "loss": 0.0006, "step": 6500 }, { "epoch": 49.01246153846154, "eval_accuracy": 0.7608695652173914, "eval_loss": 1.4876021146774292, "eval_runtime": 14.8808, "eval_samples_per_second": 3.091, "eval_steps_per_second": 1.546, "step": 6500 }, { "epoch": 49.01246153846154, "step": 6500, "total_flos": 5.686818180940844e+19, "train_loss": 0.5316385031611157, "train_runtime": 7434.7076, "train_samples_per_second": 1.749, "train_steps_per_second": 0.874 }, { "epoch": 49.01246153846154, "eval_accuracy": 0.8695652173913043, "eval_loss": 0.40178221464157104, "eval_runtime": 13.1957, "eval_samples_per_second": 3.486, "eval_steps_per_second": 1.743, "step": 6500 }, { "epoch": 49.01246153846154, "eval_accuracy": 0.8695652173913043, "eval_loss": 0.40178215503692627, "eval_runtime": 13.2103, "eval_samples_per_second": 3.482, "eval_steps_per_second": 1.741, "step": 6500 } ], "logging_steps": 10, "max_steps": 6500, "num_input_tokens_seen": 0, "num_train_epochs": 9223372036854775807, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 5.686818180940844e+19, "train_batch_size": 2, "trial_name": null, "trial_params": null }