reward-deberta-v3-base-aspect / trainer_state.json
theblackcat102
[init] 1500 iterations
fbccd8e
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.8444512757454659,
"global_step": 1500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 1.765775730203337e-05,
"loss": 0.878,
"step": 10
},
{
"epoch": 0.02,
"learning_rate": 2.257992073472628e-05,
"loss": 0.8493,
"step": 20
},
{
"epoch": 0.04,
"learning_rate": 2.5553565295129308e-05,
"loss": 0.8135,
"step": 30
},
{
"epoch": 0.05,
"learning_rate": 2.7895435338793016e-05,
"loss": 0.7467,
"step": 40
},
{
"epoch": 0.06,
"learning_rate": 2.9686949735748396e-05,
"loss": 0.6927,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 2.992465466722478e-05,
"loss": 0.6988,
"step": 60
},
{
"epoch": 0.09,
"learning_rate": 2.9811636668061953e-05,
"loss": 0.6967,
"step": 70
},
{
"epoch": 0.1,
"learning_rate": 2.9686061113436587e-05,
"loss": 0.6456,
"step": 80
},
{
"epoch": 0.11,
"learning_rate": 2.956048555881122e-05,
"loss": 0.6476,
"step": 90
},
{
"epoch": 0.12,
"learning_rate": 2.944746755964839e-05,
"loss": 0.6664,
"step": 100
},
{
"epoch": 0.14,
"learning_rate": 2.9321892005023023e-05,
"loss": 0.6349,
"step": 110
},
{
"epoch": 0.15,
"learning_rate": 2.9196316450397654e-05,
"loss": 0.6442,
"step": 120
},
{
"epoch": 0.16,
"learning_rate": 2.908329845123483e-05,
"loss": 0.6758,
"step": 130
},
{
"epoch": 0.17,
"learning_rate": 2.8957722896609463e-05,
"loss": 0.6686,
"step": 140
},
{
"epoch": 0.18,
"learning_rate": 2.8832147341984094e-05,
"loss": 0.6853,
"step": 150
},
{
"epoch": 0.2,
"learning_rate": 2.870657178735873e-05,
"loss": 0.6428,
"step": 160
},
{
"epoch": 0.21,
"learning_rate": 2.8580996232733363e-05,
"loss": 0.6093,
"step": 170
},
{
"epoch": 0.22,
"learning_rate": 2.8455420678107998e-05,
"loss": 0.6382,
"step": 180
},
{
"epoch": 0.23,
"learning_rate": 2.832984512348263e-05,
"loss": 0.6236,
"step": 190
},
{
"epoch": 0.25,
"learning_rate": 2.820426956885726e-05,
"loss": 0.6172,
"step": 200
},
{
"epoch": 0.26,
"learning_rate": 2.8078694014231895e-05,
"loss": 0.6481,
"step": 210
},
{
"epoch": 0.27,
"learning_rate": 2.795311845960653e-05,
"loss": 0.6091,
"step": 220
},
{
"epoch": 0.28,
"learning_rate": 2.7827542904981164e-05,
"loss": 0.6343,
"step": 230
},
{
"epoch": 0.3,
"learning_rate": 2.7701967350355798e-05,
"loss": 0.6248,
"step": 240
},
{
"epoch": 0.31,
"learning_rate": 2.7576391795730433e-05,
"loss": 0.6516,
"step": 250
},
{
"epoch": 0.31,
"eval_oasst_export_w_label_accuracy": 0.7197802197802198,
"eval_oasst_export_w_label_kendalltau": 0.35809523809523786,
"eval_oasst_export_w_label_loss": 0.64404296875,
"eval_oasst_export_w_label_neg_score": -0.31201171875,
"eval_oasst_export_w_label_pos_score": 0.261962890625,
"eval_oasst_export_w_label_runtime": 69.9812,
"eval_oasst_export_w_label_samples_per_second": 26.007,
"eval_oasst_export_w_label_score_diff": 0.57421875,
"eval_oasst_export_w_label_steps_per_second": 6.502,
"step": 251
},
{
"epoch": 0.32,
"learning_rate": 2.7450816241105064e-05,
"loss": 0.6195,
"step": 260
},
{
"epoch": 0.33,
"learning_rate": 2.73252406864797e-05,
"loss": 0.5911,
"step": 270
},
{
"epoch": 0.34,
"learning_rate": 2.7199665131854333e-05,
"loss": 0.621,
"step": 280
},
{
"epoch": 0.36,
"learning_rate": 2.7074089577228968e-05,
"loss": 0.6102,
"step": 290
},
{
"epoch": 0.37,
"learning_rate": 2.6948514022603602e-05,
"loss": 0.5947,
"step": 300
},
{
"epoch": 0.38,
"learning_rate": 2.6822938467978233e-05,
"loss": 0.6253,
"step": 310
},
{
"epoch": 0.39,
"learning_rate": 2.6697362913352868e-05,
"loss": 0.57,
"step": 320
},
{
"epoch": 0.41,
"learning_rate": 2.6571787358727503e-05,
"loss": 0.6439,
"step": 330
},
{
"epoch": 0.42,
"learning_rate": 2.6446211804102137e-05,
"loss": 0.611,
"step": 340
},
{
"epoch": 0.43,
"learning_rate": 2.632063624947677e-05,
"loss": 0.6451,
"step": 350
},
{
"epoch": 0.44,
"learning_rate": 2.6195060694851403e-05,
"loss": 0.6299,
"step": 360
},
{
"epoch": 0.45,
"learning_rate": 2.6069485140226037e-05,
"loss": 0.6554,
"step": 370
},
{
"epoch": 0.47,
"learning_rate": 2.594390958560067e-05,
"loss": 0.6024,
"step": 380
},
{
"epoch": 0.48,
"learning_rate": 2.5818334030975303e-05,
"loss": 0.6215,
"step": 390
},
{
"epoch": 0.49,
"learning_rate": 2.5705316031812474e-05,
"loss": 0.594,
"step": 400
},
{
"epoch": 0.5,
"learning_rate": 2.5579740477187108e-05,
"loss": 0.5889,
"step": 410
},
{
"epoch": 0.52,
"learning_rate": 2.5454164922561743e-05,
"loss": 0.6088,
"step": 420
},
{
"epoch": 0.53,
"learning_rate": 2.5328589367936377e-05,
"loss": 0.6793,
"step": 430
},
{
"epoch": 0.54,
"learning_rate": 2.5203013813311012e-05,
"loss": 0.6468,
"step": 440
},
{
"epoch": 0.55,
"learning_rate": 2.5077438258685643e-05,
"loss": 0.6457,
"step": 450
},
{
"epoch": 0.57,
"learning_rate": 2.4951862704060278e-05,
"loss": 0.6246,
"step": 460
},
{
"epoch": 0.58,
"learning_rate": 2.4826287149434912e-05,
"loss": 0.5629,
"step": 470
},
{
"epoch": 0.59,
"learning_rate": 2.4700711594809547e-05,
"loss": 0.6508,
"step": 480
},
{
"epoch": 0.6,
"learning_rate": 2.457513604018418e-05,
"loss": 0.635,
"step": 490
},
{
"epoch": 0.61,
"learning_rate": 2.444956048555881e-05,
"loss": 0.6358,
"step": 500
},
{
"epoch": 0.62,
"eval_oasst_export_w_label_accuracy": 0.7346153846153847,
"eval_oasst_export_w_label_kendalltau": 0.3878754578754578,
"eval_oasst_export_w_label_loss": 0.6220703125,
"eval_oasst_export_w_label_neg_score": 0.30078125,
"eval_oasst_export_w_label_pos_score": 0.83984375,
"eval_oasst_export_w_label_runtime": 69.9591,
"eval_oasst_export_w_label_samples_per_second": 26.015,
"eval_oasst_export_w_label_score_diff": 0.5390625,
"eval_oasst_export_w_label_steps_per_second": 6.504,
"step": 502
},
{
"epoch": 0.63,
"learning_rate": 2.4323984930933444e-05,
"loss": 0.6526,
"step": 510
},
{
"epoch": 0.64,
"learning_rate": 2.4198409376308078e-05,
"loss": 0.6391,
"step": 520
},
{
"epoch": 0.65,
"learning_rate": 2.4072833821682713e-05,
"loss": 0.6083,
"step": 530
},
{
"epoch": 0.66,
"learning_rate": 2.3947258267057347e-05,
"loss": 0.6283,
"step": 540
},
{
"epoch": 0.68,
"learning_rate": 2.382168271243198e-05,
"loss": 0.6115,
"step": 550
},
{
"epoch": 0.69,
"learning_rate": 2.3696107157806613e-05,
"loss": 0.6069,
"step": 560
},
{
"epoch": 0.7,
"learning_rate": 2.3570531603181248e-05,
"loss": 0.6462,
"step": 570
},
{
"epoch": 0.71,
"learning_rate": 2.3444956048555882e-05,
"loss": 0.6353,
"step": 580
},
{
"epoch": 0.73,
"learning_rate": 2.3319380493930517e-05,
"loss": 0.6194,
"step": 590
},
{
"epoch": 0.74,
"learning_rate": 2.3193804939305148e-05,
"loss": 0.618,
"step": 600
},
{
"epoch": 0.75,
"learning_rate": 2.3068229384679782e-05,
"loss": 0.5972,
"step": 610
},
{
"epoch": 0.76,
"learning_rate": 2.2942653830054417e-05,
"loss": 0.6001,
"step": 620
},
{
"epoch": 0.77,
"learning_rate": 2.281707827542905e-05,
"loss": 0.6341,
"step": 630
},
{
"epoch": 0.79,
"learning_rate": 2.2691502720803686e-05,
"loss": 0.6168,
"step": 640
},
{
"epoch": 0.8,
"learning_rate": 2.256592716617832e-05,
"loss": 0.6507,
"step": 650
},
{
"epoch": 0.81,
"learning_rate": 2.2440351611552952e-05,
"loss": 0.6237,
"step": 660
},
{
"epoch": 0.82,
"learning_rate": 2.2314776056927586e-05,
"loss": 0.6004,
"step": 670
},
{
"epoch": 0.84,
"learning_rate": 2.218920050230222e-05,
"loss": 0.6198,
"step": 680
},
{
"epoch": 0.85,
"learning_rate": 2.2063624947676852e-05,
"loss": 0.6166,
"step": 690
},
{
"epoch": 0.86,
"learning_rate": 2.1938049393051487e-05,
"loss": 0.6492,
"step": 700
},
{
"epoch": 0.87,
"learning_rate": 2.1812473838426118e-05,
"loss": 0.6194,
"step": 710
},
{
"epoch": 0.89,
"learning_rate": 2.1686898283800752e-05,
"loss": 0.5555,
"step": 720
},
{
"epoch": 0.9,
"learning_rate": 2.1561322729175387e-05,
"loss": 0.6118,
"step": 730
},
{
"epoch": 0.91,
"learning_rate": 2.143574717455002e-05,
"loss": 0.6099,
"step": 740
},
{
"epoch": 0.92,
"learning_rate": 2.1310171619924656e-05,
"loss": 0.5656,
"step": 750
},
{
"epoch": 0.93,
"eval_oasst_export_w_label_accuracy": 0.7307692307692307,
"eval_oasst_export_w_label_kendalltau": 0.3831868131868137,
"eval_oasst_export_w_label_loss": 0.63671875,
"eval_oasst_export_w_label_neg_score": -0.3466796875,
"eval_oasst_export_w_label_pos_score": 0.58740234375,
"eval_oasst_export_w_label_runtime": 70.0188,
"eval_oasst_export_w_label_samples_per_second": 25.993,
"eval_oasst_export_w_label_score_diff": 0.93408203125,
"eval_oasst_export_w_label_steps_per_second": 6.498,
"step": 753
},
{
"epoch": 0.93,
"learning_rate": 2.1184596065299287e-05,
"loss": 0.6576,
"step": 760
},
{
"epoch": 0.95,
"learning_rate": 2.1059020510673922e-05,
"loss": 0.59,
"step": 770
},
{
"epoch": 0.96,
"learning_rate": 2.0933444956048556e-05,
"loss": 0.6391,
"step": 780
},
{
"epoch": 0.97,
"learning_rate": 2.080786940142319e-05,
"loss": 0.6061,
"step": 790
},
{
"epoch": 0.98,
"learning_rate": 2.0682293846797825e-05,
"loss": 0.5858,
"step": 800
},
{
"epoch": 1.0,
"learning_rate": 2.0556718292172457e-05,
"loss": 0.6246,
"step": 810
},
{
"epoch": 1.01,
"learning_rate": 2.043114273754709e-05,
"loss": 0.5878,
"step": 820
},
{
"epoch": 1.02,
"learning_rate": 2.0305567182921726e-05,
"loss": 0.578,
"step": 830
},
{
"epoch": 1.03,
"learning_rate": 2.017999162829636e-05,
"loss": 0.6247,
"step": 840
},
{
"epoch": 1.05,
"learning_rate": 2.0054416073670995e-05,
"loss": 0.5841,
"step": 850
},
{
"epoch": 1.06,
"learning_rate": 1.9928840519045626e-05,
"loss": 0.6293,
"step": 860
},
{
"epoch": 1.07,
"learning_rate": 1.980326496442026e-05,
"loss": 0.5858,
"step": 870
},
{
"epoch": 1.08,
"learning_rate": 1.9677689409794892e-05,
"loss": 0.597,
"step": 880
},
{
"epoch": 1.09,
"learning_rate": 1.9552113855169526e-05,
"loss": 0.5869,
"step": 890
},
{
"epoch": 1.11,
"learning_rate": 1.942653830054416e-05,
"loss": 0.5672,
"step": 900
},
{
"epoch": 1.12,
"learning_rate": 1.9300962745918795e-05,
"loss": 0.6281,
"step": 910
},
{
"epoch": 1.13,
"learning_rate": 1.9175387191293427e-05,
"loss": 0.5356,
"step": 920
},
{
"epoch": 1.14,
"learning_rate": 1.904981163666806e-05,
"loss": 0.5277,
"step": 930
},
{
"epoch": 1.16,
"learning_rate": 1.8924236082042696e-05,
"loss": 0.5968,
"step": 940
},
{
"epoch": 1.17,
"learning_rate": 1.879866052741733e-05,
"loss": 0.6129,
"step": 950
},
{
"epoch": 1.18,
"learning_rate": 1.8673084972791965e-05,
"loss": 0.5914,
"step": 960
},
{
"epoch": 1.19,
"learning_rate": 1.8547509418166596e-05,
"loss": 0.5648,
"step": 970
},
{
"epoch": 1.21,
"learning_rate": 1.842193386354123e-05,
"loss": 0.5783,
"step": 980
},
{
"epoch": 1.22,
"learning_rate": 1.8296358308915865e-05,
"loss": 0.5895,
"step": 990
},
{
"epoch": 1.23,
"learning_rate": 1.81707827542905e-05,
"loss": 0.589,
"step": 1000
},
{
"epoch": 1.23,
"eval_oasst_export_w_label_accuracy": 0.7329670329670329,
"eval_oasst_export_w_label_kendalltau": 0.36846153846153845,
"eval_oasst_export_w_label_loss": 0.63330078125,
"eval_oasst_export_w_label_neg_score": -0.88427734375,
"eval_oasst_export_w_label_pos_score": 0.0426025390625,
"eval_oasst_export_w_label_runtime": 70.038,
"eval_oasst_export_w_label_samples_per_second": 25.986,
"eval_oasst_export_w_label_score_diff": 0.9267578125,
"eval_oasst_export_w_label_steps_per_second": 6.496,
"step": 1004
},
{
"epoch": 1.24,
"learning_rate": 1.8045207199665134e-05,
"loss": 0.5239,
"step": 1010
},
{
"epoch": 1.25,
"learning_rate": 1.7919631645039765e-05,
"loss": 0.5674,
"step": 1020
},
{
"epoch": 1.27,
"learning_rate": 1.77940560904144e-05,
"loss": 0.5643,
"step": 1030
},
{
"epoch": 1.28,
"learning_rate": 1.7668480535789035e-05,
"loss": 0.5631,
"step": 1040
},
{
"epoch": 1.29,
"learning_rate": 1.754290498116367e-05,
"loss": 0.5808,
"step": 1050
},
{
"epoch": 1.3,
"learning_rate": 1.7417329426538304e-05,
"loss": 0.5936,
"step": 1060
},
{
"epoch": 1.32,
"learning_rate": 1.729175387191293e-05,
"loss": 0.597,
"step": 1070
},
{
"epoch": 1.33,
"learning_rate": 1.7166178317287566e-05,
"loss": 0.5744,
"step": 1080
},
{
"epoch": 1.34,
"learning_rate": 1.70406027626622e-05,
"loss": 0.6338,
"step": 1090
},
{
"epoch": 1.35,
"learning_rate": 1.6915027208036835e-05,
"loss": 0.5946,
"step": 1100
},
{
"epoch": 1.36,
"learning_rate": 1.678945165341147e-05,
"loss": 0.5915,
"step": 1110
},
{
"epoch": 1.38,
"learning_rate": 1.6663876098786104e-05,
"loss": 0.6319,
"step": 1120
},
{
"epoch": 1.39,
"learning_rate": 1.6538300544160735e-05,
"loss": 0.5423,
"step": 1130
},
{
"epoch": 1.4,
"learning_rate": 1.641272498953537e-05,
"loss": 0.5818,
"step": 1140
},
{
"epoch": 1.41,
"learning_rate": 1.6287149434910005e-05,
"loss": 0.5576,
"step": 1150
},
{
"epoch": 1.43,
"learning_rate": 1.616157388028464e-05,
"loss": 0.5835,
"step": 1160
},
{
"epoch": 1.44,
"learning_rate": 1.6035998325659274e-05,
"loss": 0.6021,
"step": 1170
},
{
"epoch": 1.45,
"learning_rate": 1.5910422771033905e-05,
"loss": 0.6185,
"step": 1180
},
{
"epoch": 1.46,
"learning_rate": 1.578484721640854e-05,
"loss": 0.5637,
"step": 1190
},
{
"epoch": 1.48,
"learning_rate": 1.5659271661783174e-05,
"loss": 0.5624,
"step": 1200
},
{
"epoch": 1.49,
"learning_rate": 1.553369610715781e-05,
"loss": 0.5981,
"step": 1210
},
{
"epoch": 1.5,
"learning_rate": 1.5408120552532443e-05,
"loss": 0.5924,
"step": 1220
},
{
"epoch": 1.51,
"learning_rate": 1.5282544997907074e-05,
"loss": 0.5456,
"step": 1230
},
{
"epoch": 1.52,
"learning_rate": 1.5156969443281707e-05,
"loss": 0.6087,
"step": 1240
},
{
"epoch": 1.54,
"learning_rate": 1.5031393888656342e-05,
"loss": 0.5329,
"step": 1250
},
{
"epoch": 1.54,
"eval_oasst_export_w_label_accuracy": 0.7263736263736263,
"eval_oasst_export_w_label_kendalltau": 0.37285714285714255,
"eval_oasst_export_w_label_loss": 0.6298828125,
"eval_oasst_export_w_label_neg_score": -0.1357421875,
"eval_oasst_export_w_label_pos_score": 0.84716796875,
"eval_oasst_export_w_label_runtime": 70.012,
"eval_oasst_export_w_label_samples_per_second": 25.996,
"eval_oasst_export_w_label_score_diff": 0.98291015625,
"eval_oasst_export_w_label_steps_per_second": 6.499,
"step": 1255
},
{
"epoch": 1.55,
"learning_rate": 1.4905818334030976e-05,
"loss": 0.5531,
"step": 1260
},
{
"epoch": 1.56,
"learning_rate": 1.478024277940561e-05,
"loss": 0.5672,
"step": 1270
},
{
"epoch": 1.57,
"learning_rate": 1.4654667224780244e-05,
"loss": 0.5994,
"step": 1280
},
{
"epoch": 1.59,
"learning_rate": 1.4529091670154878e-05,
"loss": 0.5801,
"step": 1290
},
{
"epoch": 1.6,
"learning_rate": 1.4403516115529511e-05,
"loss": 0.5672,
"step": 1300
},
{
"epoch": 1.61,
"learning_rate": 1.4277940560904144e-05,
"loss": 0.5731,
"step": 1310
},
{
"epoch": 1.62,
"learning_rate": 1.4152365006278777e-05,
"loss": 0.5448,
"step": 1320
},
{
"epoch": 1.64,
"learning_rate": 1.4026789451653411e-05,
"loss": 0.617,
"step": 1330
},
{
"epoch": 1.65,
"learning_rate": 1.3901213897028046e-05,
"loss": 0.6464,
"step": 1340
},
{
"epoch": 1.66,
"learning_rate": 1.3775638342402679e-05,
"loss": 0.568,
"step": 1350
},
{
"epoch": 1.67,
"learning_rate": 1.3650062787777313e-05,
"loss": 0.6232,
"step": 1360
},
{
"epoch": 1.68,
"learning_rate": 1.3524487233151948e-05,
"loss": 0.5646,
"step": 1370
},
{
"epoch": 1.7,
"learning_rate": 1.3398911678526581e-05,
"loss": 0.563,
"step": 1380
},
{
"epoch": 1.71,
"learning_rate": 1.3273336123901215e-05,
"loss": 0.5761,
"step": 1390
},
{
"epoch": 1.72,
"learning_rate": 1.3147760569275847e-05,
"loss": 0.5987,
"step": 1400
},
{
"epoch": 1.73,
"learning_rate": 1.3022185014650481e-05,
"loss": 0.5446,
"step": 1410
},
{
"epoch": 1.75,
"learning_rate": 1.2896609460025116e-05,
"loss": 0.5977,
"step": 1420
},
{
"epoch": 1.76,
"learning_rate": 1.2771033905399749e-05,
"loss": 0.5778,
"step": 1430
},
{
"epoch": 1.77,
"learning_rate": 1.2645458350774383e-05,
"loss": 0.5553,
"step": 1440
},
{
"epoch": 1.78,
"learning_rate": 1.2519882796149016e-05,
"loss": 0.5804,
"step": 1450
},
{
"epoch": 1.8,
"learning_rate": 1.239430724152365e-05,
"loss": 0.5993,
"step": 1460
},
{
"epoch": 1.81,
"learning_rate": 1.2268731686898285e-05,
"loss": 0.5553,
"step": 1470
},
{
"epoch": 1.82,
"learning_rate": 1.2143156132272918e-05,
"loss": 0.5741,
"step": 1480
},
{
"epoch": 1.83,
"learning_rate": 1.2017580577647551e-05,
"loss": 0.5665,
"step": 1490
},
{
"epoch": 1.84,
"learning_rate": 1.1892005023022185e-05,
"loss": 0.6043,
"step": 1500
}
],
"max_steps": 2439,
"num_train_epochs": 3,
"total_flos": 0.0,
"trial_name": null,
"trial_params": null
}