SLM_vs_LLM_experiments
/
max_seq_length_128_experiments
/microsoft
/phi_2_patent
/trainer_state.json
{ | |
"best_metric": 0.9403125047683716, | |
"best_model_checkpoint": "../../experiments_checkpoints/MAdAiLab/microsoft/phi_2_patent/checkpoint-1150", | |
"epoch": 3.0, | |
"eval_steps": 50, | |
"global_step": 1173, | |
"is_hyper_param_search": false, | |
"is_local_process_zero": true, | |
"is_world_process_zero": true, | |
"log_history": [ | |
{ | |
"epoch": 0.03, | |
"grad_norm": 129.0223388671875, | |
"learning_rate": 4.9573742540494466e-06, | |
"loss": 2.7902, | |
"step": 10 | |
}, | |
{ | |
"epoch": 0.05, | |
"grad_norm": 88.70726776123047, | |
"learning_rate": 4.914748508098892e-06, | |
"loss": 2.2437, | |
"step": 20 | |
}, | |
{ | |
"epoch": 0.08, | |
"grad_norm": 90.6304702758789, | |
"learning_rate": 4.872122762148338e-06, | |
"loss": 2.0977, | |
"step": 30 | |
}, | |
{ | |
"epoch": 0.1, | |
"grad_norm": 104.15412902832031, | |
"learning_rate": 4.829497016197784e-06, | |
"loss": 1.9312, | |
"step": 40 | |
}, | |
{ | |
"epoch": 0.13, | |
"grad_norm": 88.25494384765625, | |
"learning_rate": 4.7868712702472295e-06, | |
"loss": 1.9105, | |
"step": 50 | |
}, | |
{ | |
"epoch": 0.13, | |
"eval_accuracy": 0.3248, | |
"eval_f1_macro": 0.22072310582114402, | |
"eval_f1_micro": 0.3248, | |
"eval_loss": 1.8337750434875488, | |
"eval_runtime": 23.8295, | |
"eval_samples_per_second": 209.824, | |
"eval_steps_per_second": 3.315, | |
"step": 50 | |
}, | |
{ | |
"epoch": 0.15, | |
"grad_norm": 145.65431213378906, | |
"learning_rate": 4.744245524296676e-06, | |
"loss": 1.7945, | |
"step": 60 | |
}, | |
{ | |
"epoch": 0.18, | |
"grad_norm": 92.55813598632812, | |
"learning_rate": 4.701619778346122e-06, | |
"loss": 1.7563, | |
"step": 70 | |
}, | |
{ | |
"epoch": 0.2, | |
"grad_norm": 98.50062561035156, | |
"learning_rate": 4.658994032395567e-06, | |
"loss": 1.7211, | |
"step": 80 | |
}, | |
{ | |
"epoch": 0.23, | |
"grad_norm": 142.7998809814453, | |
"learning_rate": 4.616368286445013e-06, | |
"loss": 1.6574, | |
"step": 90 | |
}, | |
{ | |
"epoch": 0.26, | |
"grad_norm": 81.44950866699219, | |
"learning_rate": 4.573742540494459e-06, | |
"loss": 1.6023, | |
"step": 100 | |
}, | |
{ | |
"epoch": 0.26, | |
"eval_accuracy": 0.438, | |
"eval_f1_macro": 0.3087308992190071, | |
"eval_f1_micro": 0.438, | |
"eval_loss": 1.6011250019073486, | |
"eval_runtime": 23.8419, | |
"eval_samples_per_second": 209.715, | |
"eval_steps_per_second": 3.314, | |
"step": 100 | |
}, | |
{ | |
"epoch": 0.28, | |
"grad_norm": 57.21122360229492, | |
"learning_rate": 4.531116794543905e-06, | |
"loss": 1.5813, | |
"step": 110 | |
}, | |
{ | |
"epoch": 0.31, | |
"grad_norm": 64.49856567382812, | |
"learning_rate": 4.488491048593351e-06, | |
"loss": 1.534, | |
"step": 120 | |
}, | |
{ | |
"epoch": 0.33, | |
"grad_norm": 202.29051208496094, | |
"learning_rate": 4.445865302642797e-06, | |
"loss": 1.5035, | |
"step": 130 | |
}, | |
{ | |
"epoch": 0.36, | |
"grad_norm": 63.948612213134766, | |
"learning_rate": 4.403239556692242e-06, | |
"loss": 1.441, | |
"step": 140 | |
}, | |
{ | |
"epoch": 0.38, | |
"grad_norm": 104.93111419677734, | |
"learning_rate": 4.3606138107416884e-06, | |
"loss": 1.4113, | |
"step": 150 | |
}, | |
{ | |
"epoch": 0.38, | |
"eval_accuracy": 0.4912, | |
"eval_f1_macro": 0.35987844199885505, | |
"eval_f1_micro": 0.4912, | |
"eval_loss": 1.423912525177002, | |
"eval_runtime": 23.9884, | |
"eval_samples_per_second": 208.434, | |
"eval_steps_per_second": 3.293, | |
"step": 150 | |
}, | |
{ | |
"epoch": 0.41, | |
"grad_norm": 71.7006607055664, | |
"learning_rate": 4.317988064791135e-06, | |
"loss": 1.3883, | |
"step": 160 | |
}, | |
{ | |
"epoch": 0.43, | |
"grad_norm": 89.34910583496094, | |
"learning_rate": 4.27536231884058e-06, | |
"loss": 1.4598, | |
"step": 170 | |
}, | |
{ | |
"epoch": 0.46, | |
"grad_norm": 75.64799499511719, | |
"learning_rate": 4.232736572890026e-06, | |
"loss": 1.3648, | |
"step": 180 | |
}, | |
{ | |
"epoch": 0.49, | |
"grad_norm": 52.86457824707031, | |
"learning_rate": 4.190110826939472e-06, | |
"loss": 1.3918, | |
"step": 190 | |
}, | |
{ | |
"epoch": 0.51, | |
"grad_norm": 138.5376434326172, | |
"learning_rate": 4.1474850809889175e-06, | |
"loss": 1.3062, | |
"step": 200 | |
}, | |
{ | |
"epoch": 0.51, | |
"eval_accuracy": 0.5498, | |
"eval_f1_macro": 0.4108688176572631, | |
"eval_f1_micro": 0.5498, | |
"eval_loss": 1.2827750444412231, | |
"eval_runtime": 23.889, | |
"eval_samples_per_second": 209.302, | |
"eval_steps_per_second": 3.307, | |
"step": 200 | |
}, | |
{ | |
"epoch": 0.54, | |
"grad_norm": 78.88294219970703, | |
"learning_rate": 4.104859335038364e-06, | |
"loss": 1.3029, | |
"step": 210 | |
}, | |
{ | |
"epoch": 0.56, | |
"grad_norm": 78.97193145751953, | |
"learning_rate": 4.06223358908781e-06, | |
"loss": 1.2592, | |
"step": 220 | |
}, | |
{ | |
"epoch": 0.59, | |
"grad_norm": 72.6739273071289, | |
"learning_rate": 4.019607843137255e-06, | |
"loss": 1.2129, | |
"step": 230 | |
}, | |
{ | |
"epoch": 0.61, | |
"grad_norm": 118.53155517578125, | |
"learning_rate": 3.976982097186701e-06, | |
"loss": 1.235, | |
"step": 240 | |
}, | |
{ | |
"epoch": 0.64, | |
"grad_norm": 63.675079345703125, | |
"learning_rate": 3.934356351236147e-06, | |
"loss": 1.2574, | |
"step": 250 | |
}, | |
{ | |
"epoch": 0.64, | |
"eval_accuracy": 0.5822, | |
"eval_f1_macro": 0.4702416378449123, | |
"eval_f1_micro": 0.5822, | |
"eval_loss": 1.1801375150680542, | |
"eval_runtime": 23.9096, | |
"eval_samples_per_second": 209.121, | |
"eval_steps_per_second": 3.304, | |
"step": 250 | |
}, | |
{ | |
"epoch": 0.66, | |
"grad_norm": 62.65312576293945, | |
"learning_rate": 3.891730605285593e-06, | |
"loss": 1.2318, | |
"step": 260 | |
}, | |
{ | |
"epoch": 0.69, | |
"grad_norm": 66.97611236572266, | |
"learning_rate": 3.849104859335039e-06, | |
"loss": 1.1959, | |
"step": 270 | |
}, | |
{ | |
"epoch": 0.72, | |
"grad_norm": 81.28976440429688, | |
"learning_rate": 3.8064791133844846e-06, | |
"loss": 1.1418, | |
"step": 280 | |
}, | |
{ | |
"epoch": 0.74, | |
"grad_norm": 45.58428955078125, | |
"learning_rate": 3.7638533674339308e-06, | |
"loss": 1.1826, | |
"step": 290 | |
}, | |
{ | |
"epoch": 0.77, | |
"grad_norm": 60.475181579589844, | |
"learning_rate": 3.7212276214833765e-06, | |
"loss": 1.1687, | |
"step": 300 | |
}, | |
{ | |
"epoch": 0.77, | |
"eval_accuracy": 0.6032, | |
"eval_f1_macro": 0.482480554692146, | |
"eval_f1_micro": 0.6032, | |
"eval_loss": 1.140125036239624, | |
"eval_runtime": 23.9126, | |
"eval_samples_per_second": 209.095, | |
"eval_steps_per_second": 3.304, | |
"step": 300 | |
}, | |
{ | |
"epoch": 0.79, | |
"grad_norm": 87.48420715332031, | |
"learning_rate": 3.678601875532822e-06, | |
"loss": 1.1279, | |
"step": 310 | |
}, | |
{ | |
"epoch": 0.82, | |
"grad_norm": 112.53974914550781, | |
"learning_rate": 3.6359761295822684e-06, | |
"loss": 1.1576, | |
"step": 320 | |
}, | |
{ | |
"epoch": 0.84, | |
"grad_norm": 70.61732482910156, | |
"learning_rate": 3.593350383631714e-06, | |
"loss": 1.127, | |
"step": 330 | |
}, | |
{ | |
"epoch": 0.87, | |
"grad_norm": 78.03123474121094, | |
"learning_rate": 3.55072463768116e-06, | |
"loss": 1.1408, | |
"step": 340 | |
}, | |
{ | |
"epoch": 0.9, | |
"grad_norm": 65.87451934814453, | |
"learning_rate": 3.508098891730606e-06, | |
"loss": 1.1396, | |
"step": 350 | |
}, | |
{ | |
"epoch": 0.9, | |
"eval_accuracy": 0.613, | |
"eval_f1_macro": 0.5246169078724559, | |
"eval_f1_micro": 0.613, | |
"eval_loss": 1.085337519645691, | |
"eval_runtime": 23.9084, | |
"eval_samples_per_second": 209.132, | |
"eval_steps_per_second": 3.304, | |
"step": 350 | |
}, | |
{ | |
"epoch": 0.92, | |
"grad_norm": 58.78898239135742, | |
"learning_rate": 3.4654731457800517e-06, | |
"loss": 1.0928, | |
"step": 360 | |
}, | |
{ | |
"epoch": 0.95, | |
"grad_norm": 61.19722366333008, | |
"learning_rate": 3.4228473998294974e-06, | |
"loss": 1.101, | |
"step": 370 | |
}, | |
{ | |
"epoch": 0.97, | |
"grad_norm": 58.480716705322266, | |
"learning_rate": 3.3802216538789436e-06, | |
"loss": 1.0672, | |
"step": 380 | |
}, | |
{ | |
"epoch": 1.0, | |
"grad_norm": 60.74428176879883, | |
"learning_rate": 3.3375959079283893e-06, | |
"loss": 1.1035, | |
"step": 390 | |
}, | |
{ | |
"epoch": 1.02, | |
"grad_norm": 100.67351531982422, | |
"learning_rate": 3.294970161977835e-06, | |
"loss": 1.0529, | |
"step": 400 | |
}, | |
{ | |
"epoch": 1.02, | |
"eval_accuracy": 0.6226, | |
"eval_f1_macro": 0.5367966974361359, | |
"eval_f1_micro": 0.6226, | |
"eval_loss": 1.0639499425888062, | |
"eval_runtime": 24.0464, | |
"eval_samples_per_second": 207.931, | |
"eval_steps_per_second": 3.285, | |
"step": 400 | |
}, | |
{ | |
"epoch": 1.05, | |
"grad_norm": 74.74591064453125, | |
"learning_rate": 3.252344416027281e-06, | |
"loss": 1.0479, | |
"step": 410 | |
}, | |
{ | |
"epoch": 1.07, | |
"grad_norm": 66.17108917236328, | |
"learning_rate": 3.2097186700767265e-06, | |
"loss": 1.0605, | |
"step": 420 | |
}, | |
{ | |
"epoch": 1.1, | |
"grad_norm": 75.01661682128906, | |
"learning_rate": 3.1670929241261722e-06, | |
"loss": 1.0996, | |
"step": 430 | |
}, | |
{ | |
"epoch": 1.13, | |
"grad_norm": 77.53501892089844, | |
"learning_rate": 3.124467178175618e-06, | |
"loss": 1.1199, | |
"step": 440 | |
}, | |
{ | |
"epoch": 1.15, | |
"grad_norm": 93.35897827148438, | |
"learning_rate": 3.081841432225064e-06, | |
"loss": 1.0261, | |
"step": 450 | |
}, | |
{ | |
"epoch": 1.15, | |
"eval_accuracy": 0.6304, | |
"eval_f1_macro": 0.5448931120233109, | |
"eval_f1_micro": 0.6304, | |
"eval_loss": 1.0742125511169434, | |
"eval_runtime": 23.93, | |
"eval_samples_per_second": 208.943, | |
"eval_steps_per_second": 3.301, | |
"step": 450 | |
}, | |
{ | |
"epoch": 1.18, | |
"grad_norm": 94.84565734863281, | |
"learning_rate": 3.03921568627451e-06, | |
"loss": 1.0846, | |
"step": 460 | |
}, | |
{ | |
"epoch": 1.2, | |
"grad_norm": 83.83853912353516, | |
"learning_rate": 2.9965899403239555e-06, | |
"loss": 1.084, | |
"step": 470 | |
}, | |
{ | |
"epoch": 1.23, | |
"grad_norm": 76.9677734375, | |
"learning_rate": 2.9539641943734017e-06, | |
"loss": 1.004, | |
"step": 480 | |
}, | |
{ | |
"epoch": 1.25, | |
"grad_norm": 102.35611724853516, | |
"learning_rate": 2.9113384484228474e-06, | |
"loss": 1.0121, | |
"step": 490 | |
}, | |
{ | |
"epoch": 1.28, | |
"grad_norm": 66.14274597167969, | |
"learning_rate": 2.868712702472293e-06, | |
"loss": 1.0068, | |
"step": 500 | |
}, | |
{ | |
"epoch": 1.28, | |
"eval_accuracy": 0.6444, | |
"eval_f1_macro": 0.5824552772225025, | |
"eval_f1_micro": 0.6444, | |
"eval_loss": 1.0339875221252441, | |
"eval_runtime": 24.073, | |
"eval_samples_per_second": 207.702, | |
"eval_steps_per_second": 3.282, | |
"step": 500 | |
}, | |
{ | |
"epoch": 1.3, | |
"grad_norm": 42.732051849365234, | |
"learning_rate": 2.8260869565217393e-06, | |
"loss": 1.0389, | |
"step": 510 | |
}, | |
{ | |
"epoch": 1.33, | |
"grad_norm": 55.783145904541016, | |
"learning_rate": 2.783461210571185e-06, | |
"loss": 1.0863, | |
"step": 520 | |
}, | |
{ | |
"epoch": 1.36, | |
"grad_norm": 78.20547485351562, | |
"learning_rate": 2.7408354646206308e-06, | |
"loss": 1.0111, | |
"step": 530 | |
}, | |
{ | |
"epoch": 1.38, | |
"grad_norm": 60.08778381347656, | |
"learning_rate": 2.698209718670077e-06, | |
"loss": 1.0131, | |
"step": 540 | |
}, | |
{ | |
"epoch": 1.41, | |
"grad_norm": 69.90200805664062, | |
"learning_rate": 2.6555839727195226e-06, | |
"loss": 0.975, | |
"step": 550 | |
}, | |
{ | |
"epoch": 1.41, | |
"eval_accuracy": 0.65, | |
"eval_f1_macro": 0.5776773747566525, | |
"eval_f1_micro": 0.65, | |
"eval_loss": 1.0150874853134155, | |
"eval_runtime": 23.9081, | |
"eval_samples_per_second": 209.134, | |
"eval_steps_per_second": 3.304, | |
"step": 550 | |
}, | |
{ | |
"epoch": 1.43, | |
"grad_norm": 56.22085952758789, | |
"learning_rate": 2.6129582267689684e-06, | |
"loss": 1.0537, | |
"step": 560 | |
}, | |
{ | |
"epoch": 1.46, | |
"grad_norm": 58.212154388427734, | |
"learning_rate": 2.5703324808184145e-06, | |
"loss": 0.9584, | |
"step": 570 | |
}, | |
{ | |
"epoch": 1.48, | |
"grad_norm": 81.3975830078125, | |
"learning_rate": 2.5277067348678602e-06, | |
"loss": 0.9631, | |
"step": 580 | |
}, | |
{ | |
"epoch": 1.51, | |
"grad_norm": 55.56730270385742, | |
"learning_rate": 2.4850809889173064e-06, | |
"loss": 0.8791, | |
"step": 590 | |
}, | |
{ | |
"epoch": 1.53, | |
"grad_norm": 55.6596565246582, | |
"learning_rate": 2.442455242966752e-06, | |
"loss": 0.966, | |
"step": 600 | |
}, | |
{ | |
"epoch": 1.53, | |
"eval_accuracy": 0.6498, | |
"eval_f1_macro": 0.5922541588134865, | |
"eval_f1_micro": 0.6498, | |
"eval_loss": 1.0022250413894653, | |
"eval_runtime": 23.9354, | |
"eval_samples_per_second": 208.896, | |
"eval_steps_per_second": 3.301, | |
"step": 600 | |
}, | |
{ | |
"epoch": 1.56, | |
"grad_norm": 64.6776123046875, | |
"learning_rate": 2.3998294970161983e-06, | |
"loss": 1.0297, | |
"step": 610 | |
}, | |
{ | |
"epoch": 1.59, | |
"grad_norm": 52.14447784423828, | |
"learning_rate": 2.357203751065644e-06, | |
"loss": 1.0012, | |
"step": 620 | |
}, | |
{ | |
"epoch": 1.61, | |
"grad_norm": 60.170494079589844, | |
"learning_rate": 2.3145780051150897e-06, | |
"loss": 1.101, | |
"step": 630 | |
}, | |
{ | |
"epoch": 1.64, | |
"grad_norm": 28.052824020385742, | |
"learning_rate": 2.2719522591645355e-06, | |
"loss": 0.9088, | |
"step": 640 | |
}, | |
{ | |
"epoch": 1.66, | |
"grad_norm": 42.27618408203125, | |
"learning_rate": 2.229326513213981e-06, | |
"loss": 1.0201, | |
"step": 650 | |
}, | |
{ | |
"epoch": 1.66, | |
"eval_accuracy": 0.6562, | |
"eval_f1_macro": 0.5854114142305936, | |
"eval_f1_micro": 0.6562, | |
"eval_loss": 0.9899250268936157, | |
"eval_runtime": 24.0682, | |
"eval_samples_per_second": 207.743, | |
"eval_steps_per_second": 3.282, | |
"step": 650 | |
}, | |
{ | |
"epoch": 1.69, | |
"grad_norm": 39.84022521972656, | |
"learning_rate": 2.1867007672634273e-06, | |
"loss": 0.948, | |
"step": 660 | |
}, | |
{ | |
"epoch": 1.71, | |
"grad_norm": 60.004764556884766, | |
"learning_rate": 2.144075021312873e-06, | |
"loss": 1.0174, | |
"step": 670 | |
}, | |
{ | |
"epoch": 1.74, | |
"grad_norm": 50.25641632080078, | |
"learning_rate": 2.101449275362319e-06, | |
"loss": 1.043, | |
"step": 680 | |
}, | |
{ | |
"epoch": 1.76, | |
"grad_norm": 53.1125602722168, | |
"learning_rate": 2.058823529411765e-06, | |
"loss": 0.9387, | |
"step": 690 | |
}, | |
{ | |
"epoch": 1.79, | |
"grad_norm": 31.242416381835938, | |
"learning_rate": 2.0161977834612107e-06, | |
"loss": 0.9346, | |
"step": 700 | |
}, | |
{ | |
"epoch": 1.79, | |
"eval_accuracy": 0.6598, | |
"eval_f1_macro": 0.573523321862691, | |
"eval_f1_micro": 0.6598, | |
"eval_loss": 0.9807375073432922, | |
"eval_runtime": 23.9497, | |
"eval_samples_per_second": 208.771, | |
"eval_steps_per_second": 3.299, | |
"step": 700 | |
}, | |
{ | |
"epoch": 1.82, | |
"grad_norm": 39.843658447265625, | |
"learning_rate": 1.9735720375106564e-06, | |
"loss": 0.9857, | |
"step": 710 | |
}, | |
{ | |
"epoch": 1.84, | |
"grad_norm": 45.25575256347656, | |
"learning_rate": 1.9309462915601026e-06, | |
"loss": 1.0664, | |
"step": 720 | |
}, | |
{ | |
"epoch": 1.87, | |
"grad_norm": 36.48699951171875, | |
"learning_rate": 1.8883205456095483e-06, | |
"loss": 0.9861, | |
"step": 730 | |
}, | |
{ | |
"epoch": 1.89, | |
"grad_norm": 45.37526321411133, | |
"learning_rate": 1.8456947996589942e-06, | |
"loss": 1.0301, | |
"step": 740 | |
}, | |
{ | |
"epoch": 1.92, | |
"grad_norm": 34.89789962768555, | |
"learning_rate": 1.80306905370844e-06, | |
"loss": 0.9807, | |
"step": 750 | |
}, | |
{ | |
"epoch": 1.92, | |
"eval_accuracy": 0.6586, | |
"eval_f1_macro": 0.6004389739883, | |
"eval_f1_micro": 0.6586, | |
"eval_loss": 0.9694374799728394, | |
"eval_runtime": 24.081, | |
"eval_samples_per_second": 207.632, | |
"eval_steps_per_second": 3.281, | |
"step": 750 | |
}, | |
{ | |
"epoch": 1.94, | |
"grad_norm": 45.91880798339844, | |
"learning_rate": 1.7604433077578859e-06, | |
"loss": 1.049, | |
"step": 760 | |
}, | |
{ | |
"epoch": 1.97, | |
"grad_norm": 33.828102111816406, | |
"learning_rate": 1.7178175618073316e-06, | |
"loss": 0.9516, | |
"step": 770 | |
}, | |
{ | |
"epoch": 1.99, | |
"grad_norm": 46.1928825378418, | |
"learning_rate": 1.6751918158567776e-06, | |
"loss": 0.998, | |
"step": 780 | |
}, | |
{ | |
"epoch": 2.02, | |
"grad_norm": 41.60182189941406, | |
"learning_rate": 1.6325660699062235e-06, | |
"loss": 0.882, | |
"step": 790 | |
}, | |
{ | |
"epoch": 2.05, | |
"grad_norm": 39.84925842285156, | |
"learning_rate": 1.5899403239556692e-06, | |
"loss": 0.917, | |
"step": 800 | |
}, | |
{ | |
"epoch": 2.05, | |
"eval_accuracy": 0.6608, | |
"eval_f1_macro": 0.608606788032407, | |
"eval_f1_micro": 0.6608, | |
"eval_loss": 0.9664250016212463, | |
"eval_runtime": 23.955, | |
"eval_samples_per_second": 208.725, | |
"eval_steps_per_second": 3.298, | |
"step": 800 | |
}, | |
{ | |
"epoch": 2.07, | |
"grad_norm": 44.208858489990234, | |
"learning_rate": 1.5473145780051152e-06, | |
"loss": 0.9264, | |
"step": 810 | |
}, | |
{ | |
"epoch": 2.1, | |
"grad_norm": 28.902544021606445, | |
"learning_rate": 1.504688832054561e-06, | |
"loss": 0.841, | |
"step": 820 | |
}, | |
{ | |
"epoch": 2.12, | |
"grad_norm": 98.9646224975586, | |
"learning_rate": 1.4620630861040068e-06, | |
"loss": 0.9324, | |
"step": 830 | |
}, | |
{ | |
"epoch": 2.15, | |
"grad_norm": 45.69807815551758, | |
"learning_rate": 1.4194373401534528e-06, | |
"loss": 0.9197, | |
"step": 840 | |
}, | |
{ | |
"epoch": 2.17, | |
"grad_norm": 64.01890563964844, | |
"learning_rate": 1.3768115942028987e-06, | |
"loss": 0.9268, | |
"step": 850 | |
}, | |
{ | |
"epoch": 2.17, | |
"eval_accuracy": 0.6626, | |
"eval_f1_macro": 0.6107213059026647, | |
"eval_f1_micro": 0.6626, | |
"eval_loss": 0.9618874788284302, | |
"eval_runtime": 23.9512, | |
"eval_samples_per_second": 208.758, | |
"eval_steps_per_second": 3.298, | |
"step": 850 | |
}, | |
{ | |
"epoch": 2.2, | |
"grad_norm": 35.80299377441406, | |
"learning_rate": 1.3341858482523444e-06, | |
"loss": 1.0117, | |
"step": 860 | |
}, | |
{ | |
"epoch": 2.23, | |
"grad_norm": 53.93342971801758, | |
"learning_rate": 1.2915601023017904e-06, | |
"loss": 0.883, | |
"step": 870 | |
}, | |
{ | |
"epoch": 2.25, | |
"grad_norm": 77.43561553955078, | |
"learning_rate": 1.2489343563512363e-06, | |
"loss": 0.9797, | |
"step": 880 | |
}, | |
{ | |
"epoch": 2.28, | |
"grad_norm": 66.75629425048828, | |
"learning_rate": 1.206308610400682e-06, | |
"loss": 0.9676, | |
"step": 890 | |
}, | |
{ | |
"epoch": 2.3, | |
"grad_norm": 43.43849182128906, | |
"learning_rate": 1.163682864450128e-06, | |
"loss": 1.0107, | |
"step": 900 | |
}, | |
{ | |
"epoch": 2.3, | |
"eval_accuracy": 0.6648, | |
"eval_f1_macro": 0.6156214213853304, | |
"eval_f1_micro": 0.6648, | |
"eval_loss": 0.9547874927520752, | |
"eval_runtime": 23.9691, | |
"eval_samples_per_second": 208.602, | |
"eval_steps_per_second": 3.296, | |
"step": 900 | |
}, | |
{ | |
"epoch": 2.33, | |
"grad_norm": 44.34565734863281, | |
"learning_rate": 1.121057118499574e-06, | |
"loss": 0.9422, | |
"step": 910 | |
}, | |
{ | |
"epoch": 2.35, | |
"grad_norm": 29.199148178100586, | |
"learning_rate": 1.0784313725490197e-06, | |
"loss": 0.8689, | |
"step": 920 | |
}, | |
{ | |
"epoch": 2.38, | |
"grad_norm": 66.50476837158203, | |
"learning_rate": 1.0358056265984656e-06, | |
"loss": 0.8928, | |
"step": 930 | |
}, | |
{ | |
"epoch": 2.4, | |
"grad_norm": 60.33115768432617, | |
"learning_rate": 9.931798806479115e-07, | |
"loss": 0.9049, | |
"step": 940 | |
}, | |
{ | |
"epoch": 2.43, | |
"grad_norm": 34.405818939208984, | |
"learning_rate": 9.505541346973573e-07, | |
"loss": 0.9378, | |
"step": 950 | |
}, | |
{ | |
"epoch": 2.43, | |
"eval_accuracy": 0.6656, | |
"eval_f1_macro": 0.6109423474694455, | |
"eval_f1_micro": 0.6656, | |
"eval_loss": 0.9559375047683716, | |
"eval_runtime": 23.9546, | |
"eval_samples_per_second": 208.728, | |
"eval_steps_per_second": 3.298, | |
"step": 950 | |
}, | |
{ | |
"epoch": 2.46, | |
"grad_norm": 36.140838623046875, | |
"learning_rate": 9.079283887468031e-07, | |
"loss": 0.9625, | |
"step": 960 | |
}, | |
{ | |
"epoch": 2.48, | |
"grad_norm": 53.852909088134766, | |
"learning_rate": 8.653026427962489e-07, | |
"loss": 0.9463, | |
"step": 970 | |
}, | |
{ | |
"epoch": 2.51, | |
"grad_norm": 33.10171127319336, | |
"learning_rate": 8.226768968456949e-07, | |
"loss": 0.857, | |
"step": 980 | |
}, | |
{ | |
"epoch": 2.53, | |
"grad_norm": 37.32207489013672, | |
"learning_rate": 7.800511508951407e-07, | |
"loss": 0.9545, | |
"step": 990 | |
}, | |
{ | |
"epoch": 2.56, | |
"grad_norm": 54.08137512207031, | |
"learning_rate": 7.374254049445865e-07, | |
"loss": 0.9199, | |
"step": 1000 | |
}, | |
{ | |
"epoch": 2.56, | |
"eval_accuracy": 0.6658, | |
"eval_f1_macro": 0.6164743673742167, | |
"eval_f1_micro": 0.6658, | |
"eval_loss": 0.9514374732971191, | |
"eval_runtime": 24.1172, | |
"eval_samples_per_second": 207.321, | |
"eval_steps_per_second": 3.276, | |
"step": 1000 | |
}, | |
{ | |
"epoch": 2.58, | |
"grad_norm": 38.91159439086914, | |
"learning_rate": 6.947996589940325e-07, | |
"loss": 0.934, | |
"step": 1010 | |
}, | |
{ | |
"epoch": 2.61, | |
"grad_norm": 43.181209564208984, | |
"learning_rate": 6.521739130434783e-07, | |
"loss": 0.883, | |
"step": 1020 | |
}, | |
{ | |
"epoch": 2.63, | |
"grad_norm": 43.725006103515625, | |
"learning_rate": 6.095481670929241e-07, | |
"loss": 0.8582, | |
"step": 1030 | |
}, | |
{ | |
"epoch": 2.66, | |
"grad_norm": 34.141021728515625, | |
"learning_rate": 5.669224211423701e-07, | |
"loss": 0.8987, | |
"step": 1040 | |
}, | |
{ | |
"epoch": 2.69, | |
"grad_norm": 40.176025390625, | |
"learning_rate": 5.242966751918159e-07, | |
"loss": 0.8467, | |
"step": 1050 | |
}, | |
{ | |
"epoch": 2.69, | |
"eval_accuracy": 0.6714, | |
"eval_f1_macro": 0.6203159302594646, | |
"eval_f1_micro": 0.6714, | |
"eval_loss": 0.945437490940094, | |
"eval_runtime": 23.9548, | |
"eval_samples_per_second": 208.727, | |
"eval_steps_per_second": 3.298, | |
"step": 1050 | |
}, | |
{ | |
"epoch": 2.71, | |
"grad_norm": 24.048503875732422, | |
"learning_rate": 4.816709292412618e-07, | |
"loss": 0.8916, | |
"step": 1060 | |
}, | |
{ | |
"epoch": 2.74, | |
"grad_norm": 51.164852142333984, | |
"learning_rate": 4.3904518329070764e-07, | |
"loss": 0.9561, | |
"step": 1070 | |
}, | |
{ | |
"epoch": 2.76, | |
"grad_norm": 57.753108978271484, | |
"learning_rate": 3.964194373401535e-07, | |
"loss": 0.9553, | |
"step": 1080 | |
}, | |
{ | |
"epoch": 2.79, | |
"grad_norm": 40.737831115722656, | |
"learning_rate": 3.5379369138959936e-07, | |
"loss": 0.9115, | |
"step": 1090 | |
}, | |
{ | |
"epoch": 2.81, | |
"grad_norm": 43.91144943237305, | |
"learning_rate": 3.111679454390452e-07, | |
"loss": 0.8923, | |
"step": 1100 | |
}, | |
{ | |
"epoch": 2.81, | |
"eval_accuracy": 0.67, | |
"eval_f1_macro": 0.6205757591067013, | |
"eval_f1_micro": 0.67, | |
"eval_loss": 0.9412500262260437, | |
"eval_runtime": 23.9783, | |
"eval_samples_per_second": 208.522, | |
"eval_steps_per_second": 3.295, | |
"step": 1100 | |
}, | |
{ | |
"epoch": 2.84, | |
"grad_norm": 27.63128662109375, | |
"learning_rate": 2.685421994884911e-07, | |
"loss": 0.9275, | |
"step": 1110 | |
}, | |
{ | |
"epoch": 2.86, | |
"grad_norm": 38.49683380126953, | |
"learning_rate": 2.2591645353793694e-07, | |
"loss": 0.933, | |
"step": 1120 | |
}, | |
{ | |
"epoch": 2.89, | |
"grad_norm": 28.47804832458496, | |
"learning_rate": 1.8329070758738277e-07, | |
"loss": 0.9074, | |
"step": 1130 | |
}, | |
{ | |
"epoch": 2.92, | |
"grad_norm": 44.488765716552734, | |
"learning_rate": 1.4066496163682866e-07, | |
"loss": 0.9654, | |
"step": 1140 | |
}, | |
{ | |
"epoch": 2.94, | |
"grad_norm": 49.465614318847656, | |
"learning_rate": 9.803921568627452e-08, | |
"loss": 0.9545, | |
"step": 1150 | |
}, | |
{ | |
"epoch": 2.94, | |
"eval_accuracy": 0.6678, | |
"eval_f1_macro": 0.6212650261638899, | |
"eval_f1_micro": 0.6678, | |
"eval_loss": 0.9403125047683716, | |
"eval_runtime": 23.9809, | |
"eval_samples_per_second": 208.5, | |
"eval_steps_per_second": 3.294, | |
"step": 1150 | |
}, | |
{ | |
"epoch": 2.97, | |
"grad_norm": 38.096954345703125, | |
"learning_rate": 5.541346973572038e-08, | |
"loss": 0.9039, | |
"step": 1160 | |
}, | |
{ | |
"epoch": 2.99, | |
"grad_norm": 41.38774871826172, | |
"learning_rate": 1.2787723785166243e-08, | |
"loss": 0.8791, | |
"step": 1170 | |
}, | |
{ | |
"epoch": 3.0, | |
"step": 1173, | |
"total_flos": 1.451478791393116e+17, | |
"train_loss": 1.123963495044757, | |
"train_runtime": 3214.4035, | |
"train_samples_per_second": 23.332, | |
"train_steps_per_second": 0.365 | |
} | |
], | |
"logging_steps": 10, | |
"max_steps": 1173, | |
"num_input_tokens_seen": 0, | |
"num_train_epochs": 3, | |
"save_steps": 50, | |
"total_flos": 1.451478791393116e+17, | |
"train_batch_size": 32, | |
"trial_name": null, | |
"trial_params": null | |
} | |