| { | |
| "best_global_step": 4000, | |
| "best_metric": 0.17307986319065094, | |
| "best_model_checkpoint": "/data/alamparan/mattext_ckpt/results/2026-02-05/01-13-34/pretrain/checkpoints/robocrys_rep_test-pretrain/checkpoint-4000", | |
| "epoch": 3.875968992248062, | |
| "eval_steps": 50, | |
| "global_step": 4000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.04844961240310078, | |
| "grad_norm": 1.166297197341919, | |
| "learning_rate": 0.00019981007751937986, | |
| "loss": 5.9313, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.04844961240310078, | |
| "eval_loss": 4.05467414855957, | |
| "eval_runtime": 86.2208, | |
| "eval_samples_per_second": 220.411, | |
| "eval_steps_per_second": 4.593, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.09689922480620156, | |
| "grad_norm": 0.7852576375007629, | |
| "learning_rate": 0.00019961627906976747, | |
| "loss": 3.8511, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.09689922480620156, | |
| "eval_loss": 3.614811658859253, | |
| "eval_runtime": 105.419, | |
| "eval_samples_per_second": 180.271, | |
| "eval_steps_per_second": 3.756, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.14534883720930233, | |
| "grad_norm": 0.9356492161750793, | |
| "learning_rate": 0.00019942248062015505, | |
| "loss": 3.5665, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.14534883720930233, | |
| "eval_loss": 3.42499041557312, | |
| "eval_runtime": 119.6728, | |
| "eval_samples_per_second": 158.8, | |
| "eval_steps_per_second": 3.309, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1937984496124031, | |
| "grad_norm": 0.8641079068183899, | |
| "learning_rate": 0.00019922868217054265, | |
| "loss": 3.4112, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1937984496124031, | |
| "eval_loss": 3.2944159507751465, | |
| "eval_runtime": 107.5464, | |
| "eval_samples_per_second": 176.705, | |
| "eval_steps_per_second": 3.682, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24224806201550386, | |
| "grad_norm": 0.8657866716384888, | |
| "learning_rate": 0.00019903488372093023, | |
| "loss": 3.2915, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.24224806201550386, | |
| "eval_loss": 3.1811130046844482, | |
| "eval_runtime": 111.4126, | |
| "eval_samples_per_second": 170.573, | |
| "eval_steps_per_second": 3.554, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.29069767441860467, | |
| "grad_norm": 1.086035132408142, | |
| "learning_rate": 0.00019884108527131784, | |
| "loss": 3.2011, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.29069767441860467, | |
| "eval_loss": 3.0944509506225586, | |
| "eval_runtime": 109.0889, | |
| "eval_samples_per_second": 174.207, | |
| "eval_steps_per_second": 3.63, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3391472868217054, | |
| "grad_norm": 1.14714515209198, | |
| "learning_rate": 0.00019864728682170545, | |
| "loss": 3.1197, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3391472868217054, | |
| "eval_loss": 3.013719081878662, | |
| "eval_runtime": 113.2854, | |
| "eval_samples_per_second": 167.753, | |
| "eval_steps_per_second": 3.496, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.3875968992248062, | |
| "grad_norm": 1.3692086935043335, | |
| "learning_rate": 0.00019845348837209303, | |
| "loss": 3.0293, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.3875968992248062, | |
| "eval_loss": 2.9276790618896484, | |
| "eval_runtime": 109.3164, | |
| "eval_samples_per_second": 173.844, | |
| "eval_steps_per_second": 3.623, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.436046511627907, | |
| "grad_norm": 1.0482667684555054, | |
| "learning_rate": 0.00019825968992248064, | |
| "loss": 2.937, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.436046511627907, | |
| "eval_loss": 2.824265718460083, | |
| "eval_runtime": 113.3929, | |
| "eval_samples_per_second": 167.594, | |
| "eval_steps_per_second": 3.492, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.4844961240310077, | |
| "grad_norm": 1.4893537759780884, | |
| "learning_rate": 0.00019806589147286822, | |
| "loss": 2.8401, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4844961240310077, | |
| "eval_loss": 2.681960344314575, | |
| "eval_runtime": 112.2413, | |
| "eval_samples_per_second": 169.314, | |
| "eval_steps_per_second": 3.528, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5329457364341085, | |
| "grad_norm": 1.3668580055236816, | |
| "learning_rate": 0.00019787209302325582, | |
| "loss": 2.6953, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5329457364341085, | |
| "eval_loss": 2.482776165008545, | |
| "eval_runtime": 109.465, | |
| "eval_samples_per_second": 173.608, | |
| "eval_steps_per_second": 3.618, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5813953488372093, | |
| "grad_norm": 1.7991057634353638, | |
| "learning_rate": 0.00019767829457364343, | |
| "loss": 2.4792, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5813953488372093, | |
| "eval_loss": 2.0755560398101807, | |
| "eval_runtime": 111.9949, | |
| "eval_samples_per_second": 169.686, | |
| "eval_steps_per_second": 3.536, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6298449612403101, | |
| "grad_norm": 1.6921905279159546, | |
| "learning_rate": 0.000197484496124031, | |
| "loss": 1.9556, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6298449612403101, | |
| "eval_loss": 1.431490421295166, | |
| "eval_runtime": 111.5482, | |
| "eval_samples_per_second": 170.366, | |
| "eval_steps_per_second": 3.55, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6782945736434108, | |
| "grad_norm": 1.6035951375961304, | |
| "learning_rate": 0.00019729069767441862, | |
| "loss": 1.4683, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.6782945736434108, | |
| "eval_loss": 1.0505075454711914, | |
| "eval_runtime": 110.2108, | |
| "eval_samples_per_second": 172.433, | |
| "eval_steps_per_second": 3.593, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7267441860465116, | |
| "grad_norm": 1.1903802156448364, | |
| "learning_rate": 0.0001970968992248062, | |
| "loss": 1.1292, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7267441860465116, | |
| "eval_loss": 0.881108820438385, | |
| "eval_runtime": 112.8095, | |
| "eval_samples_per_second": 168.461, | |
| "eval_steps_per_second": 3.51, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7751937984496124, | |
| "grad_norm": 1.0895042419433594, | |
| "learning_rate": 0.0001969031007751938, | |
| "loss": 0.9955, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.7751937984496124, | |
| "eval_loss": 0.7912827730178833, | |
| "eval_runtime": 113.3802, | |
| "eval_samples_per_second": 167.613, | |
| "eval_steps_per_second": 3.493, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8236434108527132, | |
| "grad_norm": 1.1862001419067383, | |
| "learning_rate": 0.0001967093023255814, | |
| "loss": 0.9121, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.8236434108527132, | |
| "eval_loss": 0.7358818650245667, | |
| "eval_runtime": 110.3147, | |
| "eval_samples_per_second": 172.271, | |
| "eval_steps_per_second": 3.59, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.872093023255814, | |
| "grad_norm": 0.9009504318237305, | |
| "learning_rate": 0.000196515503875969, | |
| "loss": 0.8213, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.872093023255814, | |
| "eval_loss": 0.673936665058136, | |
| "eval_runtime": 110.4965, | |
| "eval_samples_per_second": 171.987, | |
| "eval_steps_per_second": 3.584, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9205426356589147, | |
| "grad_norm": 0.8950819373130798, | |
| "learning_rate": 0.0001963217054263566, | |
| "loss": 0.7789, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9205426356589147, | |
| "eval_loss": 0.6290402412414551, | |
| "eval_runtime": 126.6737, | |
| "eval_samples_per_second": 150.023, | |
| "eval_steps_per_second": 3.126, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9689922480620154, | |
| "grad_norm": 0.8964680433273315, | |
| "learning_rate": 0.00019612790697674418, | |
| "loss": 0.7162, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.9689922480620154, | |
| "eval_loss": 0.6034494042396545, | |
| "eval_runtime": 107.8336, | |
| "eval_samples_per_second": 176.235, | |
| "eval_steps_per_second": 3.672, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0174418604651163, | |
| "grad_norm": 0.8578282594680786, | |
| "learning_rate": 0.0001959341085271318, | |
| "loss": 0.6918, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.0174418604651163, | |
| "eval_loss": 0.571689784526825, | |
| "eval_runtime": 109.0331, | |
| "eval_samples_per_second": 174.296, | |
| "eval_steps_per_second": 3.632, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.0658914728682172, | |
| "grad_norm": 0.8138054609298706, | |
| "learning_rate": 0.00019574031007751937, | |
| "loss": 0.6908, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.0658914728682172, | |
| "eval_loss": 0.5466533899307251, | |
| "eval_runtime": 107.3453, | |
| "eval_samples_per_second": 177.036, | |
| "eval_steps_per_second": 3.689, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.1143410852713178, | |
| "grad_norm": 0.8871294856071472, | |
| "learning_rate": 0.00019554651162790698, | |
| "loss": 0.649, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.1143410852713178, | |
| "eval_loss": 0.523186206817627, | |
| "eval_runtime": 110.6186, | |
| "eval_samples_per_second": 171.798, | |
| "eval_steps_per_second": 3.58, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.1627906976744187, | |
| "grad_norm": 0.7663435339927673, | |
| "learning_rate": 0.0001953527131782946, | |
| "loss": 0.5908, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.1627906976744187, | |
| "eval_loss": 0.5085064172744751, | |
| "eval_runtime": 108.5509, | |
| "eval_samples_per_second": 175.07, | |
| "eval_steps_per_second": 3.648, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.2112403100775193, | |
| "grad_norm": 0.7912157773971558, | |
| "learning_rate": 0.00019515891472868217, | |
| "loss": 0.5644, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.2112403100775193, | |
| "eval_loss": 0.4843023419380188, | |
| "eval_runtime": 107.8466, | |
| "eval_samples_per_second": 176.213, | |
| "eval_steps_per_second": 3.672, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.2596899224806202, | |
| "grad_norm": 0.7282177805900574, | |
| "learning_rate": 0.00019496511627906978, | |
| "loss": 0.5577, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.2596899224806202, | |
| "eval_loss": 0.46449020504951477, | |
| "eval_runtime": 106.8549, | |
| "eval_samples_per_second": 177.849, | |
| "eval_steps_per_second": 3.706, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.308139534883721, | |
| "grad_norm": 0.6605934500694275, | |
| "learning_rate": 0.00019477131782945736, | |
| "loss": 0.5335, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.308139534883721, | |
| "eval_loss": 0.4484124481678009, | |
| "eval_runtime": 126.0012, | |
| "eval_samples_per_second": 150.824, | |
| "eval_steps_per_second": 3.143, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.3565891472868217, | |
| "grad_norm": 0.6995412707328796, | |
| "learning_rate": 0.00019457751937984496, | |
| "loss": 0.5396, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.3565891472868217, | |
| "eval_loss": 0.4377157688140869, | |
| "eval_runtime": 118.8732, | |
| "eval_samples_per_second": 159.868, | |
| "eval_steps_per_second": 3.331, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.4050387596899225, | |
| "grad_norm": 0.6287787556648254, | |
| "learning_rate": 0.00019438372093023257, | |
| "loss": 0.4988, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.4050387596899225, | |
| "eval_loss": 0.41537100076675415, | |
| "eval_runtime": 117.3214, | |
| "eval_samples_per_second": 161.982, | |
| "eval_steps_per_second": 3.375, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.4534883720930232, | |
| "grad_norm": 0.7416621446609497, | |
| "learning_rate": 0.00019418992248062015, | |
| "loss": 0.4888, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.4534883720930232, | |
| "eval_loss": 0.4060279130935669, | |
| "eval_runtime": 118.3486, | |
| "eval_samples_per_second": 160.577, | |
| "eval_steps_per_second": 3.346, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.501937984496124, | |
| "grad_norm": 0.70711749792099, | |
| "learning_rate": 0.00019399612403100776, | |
| "loss": 0.4664, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.501937984496124, | |
| "eval_loss": 0.38495033979415894, | |
| "eval_runtime": 116.7883, | |
| "eval_samples_per_second": 162.722, | |
| "eval_steps_per_second": 3.391, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.550387596899225, | |
| "grad_norm": 0.6704577207565308, | |
| "learning_rate": 0.00019380232558139534, | |
| "loss": 0.4331, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.550387596899225, | |
| "eval_loss": 0.36785683035850525, | |
| "eval_runtime": 134.2231, | |
| "eval_samples_per_second": 141.585, | |
| "eval_steps_per_second": 2.95, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.5988372093023255, | |
| "grad_norm": 0.7335864305496216, | |
| "learning_rate": 0.00019360852713178295, | |
| "loss": 0.4298, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.5988372093023255, | |
| "eval_loss": 0.34367096424102783, | |
| "eval_runtime": 193.7622, | |
| "eval_samples_per_second": 98.079, | |
| "eval_steps_per_second": 2.044, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.6472868217054264, | |
| "grad_norm": 0.6677731871604919, | |
| "learning_rate": 0.00019341472868217055, | |
| "loss": 0.3978, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.6472868217054264, | |
| "eval_loss": 0.32630470395088196, | |
| "eval_runtime": 242.3944, | |
| "eval_samples_per_second": 78.401, | |
| "eval_steps_per_second": 1.634, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.695736434108527, | |
| "grad_norm": 0.8195075392723083, | |
| "learning_rate": 0.00019322093023255813, | |
| "loss": 0.3778, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.695736434108527, | |
| "eval_loss": 0.31104475259780884, | |
| "eval_runtime": 231.8205, | |
| "eval_samples_per_second": 81.977, | |
| "eval_steps_per_second": 1.708, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.744186046511628, | |
| "grad_norm": 0.735170841217041, | |
| "learning_rate": 0.00019302713178294574, | |
| "loss": 0.35, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.744186046511628, | |
| "eval_loss": 0.2956381142139435, | |
| "eval_runtime": 236.8085, | |
| "eval_samples_per_second": 80.25, | |
| "eval_steps_per_second": 1.672, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.7926356589147288, | |
| "grad_norm": 0.5745160579681396, | |
| "learning_rate": 0.00019283333333333332, | |
| "loss": 0.3407, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.7926356589147288, | |
| "eval_loss": 0.2869073152542114, | |
| "eval_runtime": 232.6087, | |
| "eval_samples_per_second": 81.699, | |
| "eval_steps_per_second": 1.702, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.8410852713178296, | |
| "grad_norm": 0.6333633661270142, | |
| "learning_rate": 0.00019263953488372093, | |
| "loss": 0.3317, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.8410852713178296, | |
| "eval_loss": 0.2736239731311798, | |
| "eval_runtime": 235.3635, | |
| "eval_samples_per_second": 80.743, | |
| "eval_steps_per_second": 1.683, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.8895348837209303, | |
| "grad_norm": 0.6406286358833313, | |
| "learning_rate": 0.00019244573643410854, | |
| "loss": 0.3247, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.8895348837209303, | |
| "eval_loss": 0.26805564761161804, | |
| "eval_runtime": 235.5356, | |
| "eval_samples_per_second": 80.684, | |
| "eval_steps_per_second": 1.681, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.937984496124031, | |
| "grad_norm": 0.4822098910808563, | |
| "learning_rate": 0.00019225193798449612, | |
| "loss": 0.3057, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.937984496124031, | |
| "eval_loss": 0.26311546564102173, | |
| "eval_runtime": 235.1083, | |
| "eval_samples_per_second": 80.831, | |
| "eval_steps_per_second": 1.684, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.9864341085271318, | |
| "grad_norm": 0.6586690545082092, | |
| "learning_rate": 0.00019205813953488375, | |
| "loss": 0.2954, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 1.9864341085271318, | |
| "eval_loss": 0.2547176778316498, | |
| "eval_runtime": 235.1319, | |
| "eval_samples_per_second": 80.823, | |
| "eval_steps_per_second": 1.684, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 2.0348837209302326, | |
| "grad_norm": 0.648765504360199, | |
| "learning_rate": 0.00019186434108527133, | |
| "loss": 0.3008, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.0348837209302326, | |
| "eval_loss": 0.2547691762447357, | |
| "eval_runtime": 227.135, | |
| "eval_samples_per_second": 83.668, | |
| "eval_steps_per_second": 1.743, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "grad_norm": 0.6317530870437622, | |
| "learning_rate": 0.00019167054263565894, | |
| "loss": 0.2911, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.0833333333333335, | |
| "eval_loss": 0.24894124269485474, | |
| "eval_runtime": 225.2054, | |
| "eval_samples_per_second": 84.385, | |
| "eval_steps_per_second": 1.758, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.1317829457364343, | |
| "grad_norm": 0.5900191068649292, | |
| "learning_rate": 0.00019147674418604652, | |
| "loss": 0.2873, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.1317829457364343, | |
| "eval_loss": 0.24607273936271667, | |
| "eval_runtime": 231.2435, | |
| "eval_samples_per_second": 82.182, | |
| "eval_steps_per_second": 1.712, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.1802325581395348, | |
| "grad_norm": 0.5394904017448425, | |
| "learning_rate": 0.00019128294573643413, | |
| "loss": 0.2716, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.1802325581395348, | |
| "eval_loss": 0.2443438619375229, | |
| "eval_runtime": 225.1673, | |
| "eval_samples_per_second": 84.399, | |
| "eval_steps_per_second": 1.759, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.2286821705426356, | |
| "grad_norm": 0.5458412766456604, | |
| "learning_rate": 0.00019108914728682174, | |
| "loss": 0.2727, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.2286821705426356, | |
| "eval_loss": 0.23843063414096832, | |
| "eval_runtime": 233.0169, | |
| "eval_samples_per_second": 81.556, | |
| "eval_steps_per_second": 1.699, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.2771317829457365, | |
| "grad_norm": 0.6243239641189575, | |
| "learning_rate": 0.00019089534883720932, | |
| "loss": 0.2741, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.2771317829457365, | |
| "eval_loss": 0.23107607662677765, | |
| "eval_runtime": 223.8881, | |
| "eval_samples_per_second": 84.882, | |
| "eval_steps_per_second": 1.769, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.3255813953488373, | |
| "grad_norm": 0.6094734072685242, | |
| "learning_rate": 0.00019070155038759692, | |
| "loss": 0.2719, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.3255813953488373, | |
| "eval_loss": 0.23176899552345276, | |
| "eval_runtime": 219.8841, | |
| "eval_samples_per_second": 86.427, | |
| "eval_steps_per_second": 1.801, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.374031007751938, | |
| "grad_norm": 0.6149279475212097, | |
| "learning_rate": 0.0001905077519379845, | |
| "loss": 0.2641, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.374031007751938, | |
| "eval_loss": 0.22322338819503784, | |
| "eval_runtime": 224.8759, | |
| "eval_samples_per_second": 84.509, | |
| "eval_steps_per_second": 1.761, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.4224806201550386, | |
| "grad_norm": 0.5470075607299805, | |
| "learning_rate": 0.0001903139534883721, | |
| "loss": 0.2558, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.4224806201550386, | |
| "eval_loss": 0.22380074858665466, | |
| "eval_runtime": 220.6655, | |
| "eval_samples_per_second": 86.121, | |
| "eval_steps_per_second": 1.795, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.4709302325581395, | |
| "grad_norm": 0.6210835576057434, | |
| "learning_rate": 0.00019012015503875972, | |
| "loss": 0.2544, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.4709302325581395, | |
| "eval_loss": 0.22070536017417908, | |
| "eval_runtime": 223.1476, | |
| "eval_samples_per_second": 85.163, | |
| "eval_steps_per_second": 1.775, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.5193798449612403, | |
| "grad_norm": 0.43341922760009766, | |
| "learning_rate": 0.0001899263565891473, | |
| "loss": 0.2496, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.5193798449612403, | |
| "eval_loss": 0.2163931280374527, | |
| "eval_runtime": 232.0825, | |
| "eval_samples_per_second": 81.885, | |
| "eval_steps_per_second": 1.706, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.567829457364341, | |
| "grad_norm": 0.4871957302093506, | |
| "learning_rate": 0.0001897325581395349, | |
| "loss": 0.2506, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.567829457364341, | |
| "eval_loss": 0.21461744606494904, | |
| "eval_runtime": 219.2336, | |
| "eval_samples_per_second": 86.684, | |
| "eval_steps_per_second": 1.806, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.616279069767442, | |
| "grad_norm": 0.5310043692588806, | |
| "learning_rate": 0.0001895387596899225, | |
| "loss": 0.247, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.616279069767442, | |
| "eval_loss": 0.21458372473716736, | |
| "eval_runtime": 224.6247, | |
| "eval_samples_per_second": 84.603, | |
| "eval_steps_per_second": 1.763, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.6647286821705425, | |
| "grad_norm": 0.6074191927909851, | |
| "learning_rate": 0.0001893449612403101, | |
| "loss": 0.2455, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.6647286821705425, | |
| "eval_loss": 0.21306496858596802, | |
| "eval_runtime": 223.479, | |
| "eval_samples_per_second": 85.037, | |
| "eval_steps_per_second": 1.772, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.7131782945736433, | |
| "grad_norm": 0.5458905100822449, | |
| "learning_rate": 0.00018915116279069768, | |
| "loss": 0.243, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.7131782945736433, | |
| "eval_loss": 0.2097616195678711, | |
| "eval_runtime": 220.5231, | |
| "eval_samples_per_second": 86.177, | |
| "eval_steps_per_second": 1.796, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.761627906976744, | |
| "grad_norm": 0.4153260588645935, | |
| "learning_rate": 0.00018895736434108528, | |
| "loss": 0.2356, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 2.761627906976744, | |
| "eval_loss": 0.20619072020053864, | |
| "eval_runtime": 214.0517, | |
| "eval_samples_per_second": 88.782, | |
| "eval_steps_per_second": 1.85, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 2.810077519379845, | |
| "grad_norm": 0.4979017674922943, | |
| "learning_rate": 0.0001887635658914729, | |
| "loss": 0.2361, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.810077519379845, | |
| "eval_loss": 0.20226989686489105, | |
| "eval_runtime": 216.8896, | |
| "eval_samples_per_second": 87.621, | |
| "eval_steps_per_second": 1.826, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.858527131782946, | |
| "grad_norm": 0.5476083755493164, | |
| "learning_rate": 0.00018856976744186047, | |
| "loss": 0.2347, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 2.858527131782946, | |
| "eval_loss": 0.2028292417526245, | |
| "eval_runtime": 214.8481, | |
| "eval_samples_per_second": 88.453, | |
| "eval_steps_per_second": 1.843, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 2.9069767441860463, | |
| "grad_norm": 0.5912747979164124, | |
| "learning_rate": 0.00018837596899224808, | |
| "loss": 0.2367, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.9069767441860463, | |
| "eval_loss": 0.1990610510110855, | |
| "eval_runtime": 219.8884, | |
| "eval_samples_per_second": 86.426, | |
| "eval_steps_per_second": 1.801, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.955426356589147, | |
| "grad_norm": 0.4696279466152191, | |
| "learning_rate": 0.00018818217054263566, | |
| "loss": 0.2292, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 2.955426356589147, | |
| "eval_loss": 0.20148637890815735, | |
| "eval_runtime": 225.0994, | |
| "eval_samples_per_second": 84.425, | |
| "eval_steps_per_second": 1.759, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 3.003875968992248, | |
| "grad_norm": 0.47198325395584106, | |
| "learning_rate": 0.00018798837209302327, | |
| "loss": 0.2286, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 3.003875968992248, | |
| "eval_loss": 0.1965794712305069, | |
| "eval_runtime": 216.7977, | |
| "eval_samples_per_second": 87.658, | |
| "eval_steps_per_second": 1.827, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 3.052325581395349, | |
| "grad_norm": 0.49365678429603577, | |
| "learning_rate": 0.00018779457364341087, | |
| "loss": 0.2239, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 3.052325581395349, | |
| "eval_loss": 0.1976451575756073, | |
| "eval_runtime": 215.2246, | |
| "eval_samples_per_second": 88.298, | |
| "eval_steps_per_second": 1.84, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 3.10077519379845, | |
| "grad_norm": 0.5356510281562805, | |
| "learning_rate": 0.00018760077519379845, | |
| "loss": 0.224, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 3.10077519379845, | |
| "eval_loss": 0.1925686150789261, | |
| "eval_runtime": 218.8392, | |
| "eval_samples_per_second": 86.84, | |
| "eval_steps_per_second": 1.81, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 3.14922480620155, | |
| "grad_norm": 0.46862003207206726, | |
| "learning_rate": 0.00018740697674418606, | |
| "loss": 0.212, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 3.14922480620155, | |
| "eval_loss": 0.19242693483829498, | |
| "eval_runtime": 216.1261, | |
| "eval_samples_per_second": 87.93, | |
| "eval_steps_per_second": 1.832, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 3.197674418604651, | |
| "grad_norm": 0.40870046615600586, | |
| "learning_rate": 0.00018721317829457364, | |
| "loss": 0.2188, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 3.197674418604651, | |
| "eval_loss": 0.18868175148963928, | |
| "eval_runtime": 214.9811, | |
| "eval_samples_per_second": 88.398, | |
| "eval_steps_per_second": 1.842, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 3.246124031007752, | |
| "grad_norm": 0.5424318909645081, | |
| "learning_rate": 0.00018701937984496125, | |
| "loss": 0.2157, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 3.246124031007752, | |
| "eval_loss": 0.1919330656528473, | |
| "eval_runtime": 217.7522, | |
| "eval_samples_per_second": 87.274, | |
| "eval_steps_per_second": 1.819, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 3.294573643410853, | |
| "grad_norm": 0.41225236654281616, | |
| "learning_rate": 0.00018682558139534886, | |
| "loss": 0.2156, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 3.294573643410853, | |
| "eval_loss": 0.18903516232967377, | |
| "eval_runtime": 221.615, | |
| "eval_samples_per_second": 85.752, | |
| "eval_steps_per_second": 1.787, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 3.3430232558139537, | |
| "grad_norm": 0.4955701231956482, | |
| "learning_rate": 0.00018663178294573644, | |
| "loss": 0.2087, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 3.3430232558139537, | |
| "eval_loss": 0.18453241884708405, | |
| "eval_runtime": 217.9492, | |
| "eval_samples_per_second": 87.195, | |
| "eval_steps_per_second": 1.817, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 3.391472868217054, | |
| "grad_norm": 0.4107048809528351, | |
| "learning_rate": 0.00018643798449612405, | |
| "loss": 0.2128, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.391472868217054, | |
| "eval_loss": 0.18335753679275513, | |
| "eval_runtime": 222.918, | |
| "eval_samples_per_second": 85.251, | |
| "eval_steps_per_second": 1.776, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.439922480620155, | |
| "grad_norm": 0.44138869643211365, | |
| "learning_rate": 0.00018624418604651163, | |
| "loss": 0.2069, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 3.439922480620155, | |
| "eval_loss": 0.18529056012630463, | |
| "eval_runtime": 226.8247, | |
| "eval_samples_per_second": 83.783, | |
| "eval_steps_per_second": 1.746, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 3.488372093023256, | |
| "grad_norm": 0.41599947214126587, | |
| "learning_rate": 0.00018605038759689923, | |
| "loss": 0.2069, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 3.488372093023256, | |
| "eval_loss": 0.18560050427913666, | |
| "eval_runtime": 237.1428, | |
| "eval_samples_per_second": 80.137, | |
| "eval_steps_per_second": 1.67, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 3.5368217054263567, | |
| "grad_norm": 0.39731481671333313, | |
| "learning_rate": 0.00018585658914728684, | |
| "loss": 0.2074, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 3.5368217054263567, | |
| "eval_loss": 0.1789853423833847, | |
| "eval_runtime": 220.4447, | |
| "eval_samples_per_second": 86.208, | |
| "eval_steps_per_second": 1.796, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 3.5852713178294575, | |
| "grad_norm": 0.46680784225463867, | |
| "learning_rate": 0.00018566279069767442, | |
| "loss": 0.2042, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 3.5852713178294575, | |
| "eval_loss": 0.1809505820274353, | |
| "eval_runtime": 218.2028, | |
| "eval_samples_per_second": 87.093, | |
| "eval_steps_per_second": 1.815, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 3.633720930232558, | |
| "grad_norm": 0.552648663520813, | |
| "learning_rate": 0.00018546899224806203, | |
| "loss": 0.2089, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 3.633720930232558, | |
| "eval_loss": 0.17837607860565186, | |
| "eval_runtime": 221.244, | |
| "eval_samples_per_second": 85.896, | |
| "eval_steps_per_second": 1.79, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 3.682170542635659, | |
| "grad_norm": 0.5326802134513855, | |
| "learning_rate": 0.0001852751937984496, | |
| "loss": 0.1985, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 3.682170542635659, | |
| "eval_loss": 0.17835576832294464, | |
| "eval_runtime": 227.413, | |
| "eval_samples_per_second": 83.566, | |
| "eval_steps_per_second": 1.741, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 3.7306201550387597, | |
| "grad_norm": 0.5640744566917419, | |
| "learning_rate": 0.00018508139534883722, | |
| "loss": 0.201, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 3.7306201550387597, | |
| "eval_loss": 0.17722435295581818, | |
| "eval_runtime": 220.9949, | |
| "eval_samples_per_second": 85.993, | |
| "eval_steps_per_second": 1.792, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 3.7790697674418605, | |
| "grad_norm": 0.41974276304244995, | |
| "learning_rate": 0.00018488759689922482, | |
| "loss": 0.1945, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 3.7790697674418605, | |
| "eval_loss": 0.17568857967853546, | |
| "eval_runtime": 227.7226, | |
| "eval_samples_per_second": 83.452, | |
| "eval_steps_per_second": 1.739, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 3.8275193798449614, | |
| "grad_norm": 0.4548977315425873, | |
| "learning_rate": 0.0001846937984496124, | |
| "loss": 0.2035, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 3.8275193798449614, | |
| "eval_loss": 0.17479188740253448, | |
| "eval_runtime": 225.9318, | |
| "eval_samples_per_second": 84.114, | |
| "eval_steps_per_second": 1.753, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 3.875968992248062, | |
| "grad_norm": 0.44886454939842224, | |
| "learning_rate": 0.0001845, | |
| "loss": 0.1965, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.875968992248062, | |
| "eval_loss": 0.17307986319065094, | |
| "eval_runtime": 221.697, | |
| "eval_samples_per_second": 85.721, | |
| "eval_steps_per_second": 1.786, | |
| "step": 4000 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 51600, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 10, | |
| "early_stopping_threshold": 0.001 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.034795399365018e+16, | |
| "train_batch_size": 96, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |