gokuls's picture
End of training
7670d1c
raw
history blame
21.2 kB
{
"best_metric": 0.1256455034017563,
"best_model_checkpoint": "mobilebert_sa_GLUE_Experiment_logit_kd_data_aug_mrpc/checkpoint-80319",
"epoch": 46.0,
"global_step": 90114,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"learning_rate": 4.9e-05,
"loss": 0.2964,
"step": 1959
},
{
"epoch": 1.0,
"eval_accuracy": 0.9607843137254902,
"eval_combined_score": 0.9663076498204916,
"eval_f1": 0.9718309859154929,
"eval_loss": 0.20259901881217957,
"eval_runtime": 0.7823,
"eval_samples_per_second": 521.547,
"eval_steps_per_second": 5.113,
"step": 1959
},
{
"epoch": 2.0,
"learning_rate": 4.8e-05,
"loss": 0.2307,
"step": 3918
},
{
"epoch": 2.0,
"eval_accuracy": 0.9705882352941176,
"eval_combined_score": 0.9747678018575852,
"eval_f1": 0.9789473684210526,
"eval_loss": 0.19426828622817993,
"eval_runtime": 0.7765,
"eval_samples_per_second": 525.442,
"eval_steps_per_second": 5.151,
"step": 3918
},
{
"epoch": 3.0,
"learning_rate": 4.7e-05,
"loss": 0.2221,
"step": 5877
},
{
"epoch": 3.0,
"eval_accuracy": 0.9803921568627451,
"eval_combined_score": 0.9831038798498122,
"eval_f1": 0.9858156028368794,
"eval_loss": 0.18736393749713898,
"eval_runtime": 0.773,
"eval_samples_per_second": 527.78,
"eval_steps_per_second": 5.174,
"step": 5877
},
{
"epoch": 4.0,
"learning_rate": 4.600000000000001e-05,
"loss": 0.2163,
"step": 7836
},
{
"epoch": 4.0,
"eval_accuracy": 0.9852941176470589,
"eval_combined_score": 0.9873279098873593,
"eval_f1": 0.9893617021276596,
"eval_loss": 0.17034168541431427,
"eval_runtime": 0.7757,
"eval_samples_per_second": 526.002,
"eval_steps_per_second": 5.157,
"step": 7836
},
{
"epoch": 5.0,
"learning_rate": 4.5e-05,
"loss": 0.2115,
"step": 9795
},
{
"epoch": 5.0,
"eval_accuracy": 0.9852941176470589,
"eval_combined_score": 0.9873279098873593,
"eval_f1": 0.9893617021276596,
"eval_loss": 0.18048794567584991,
"eval_runtime": 0.7764,
"eval_samples_per_second": 525.485,
"eval_steps_per_second": 5.152,
"step": 9795
},
{
"epoch": 6.0,
"learning_rate": 4.4000000000000006e-05,
"loss": 0.2071,
"step": 11754
},
{
"epoch": 6.0,
"eval_accuracy": 0.9803921568627451,
"eval_combined_score": 0.9831289406221853,
"eval_f1": 0.9858657243816255,
"eval_loss": 0.16820354759693146,
"eval_runtime": 0.7809,
"eval_samples_per_second": 522.481,
"eval_steps_per_second": 5.122,
"step": 11754
},
{
"epoch": 7.0,
"learning_rate": 4.3e-05,
"loss": 0.2036,
"step": 13713
},
{
"epoch": 7.0,
"eval_accuracy": 0.9877450980392157,
"eval_combined_score": 0.9894320516839064,
"eval_f1": 0.9911190053285969,
"eval_loss": 0.15831029415130615,
"eval_runtime": 0.78,
"eval_samples_per_second": 523.105,
"eval_steps_per_second": 5.128,
"step": 13713
},
{
"epoch": 8.0,
"learning_rate": 4.2e-05,
"loss": 0.2007,
"step": 15672
},
{
"epoch": 8.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.16275052726268768,
"eval_runtime": 0.7788,
"eval_samples_per_second": 523.893,
"eval_steps_per_second": 5.136,
"step": 15672
},
{
"epoch": 9.0,
"learning_rate": 4.1e-05,
"loss": 0.1985,
"step": 17631
},
{
"epoch": 9.0,
"eval_accuracy": 0.9852941176470589,
"eval_combined_score": 0.9873279098873593,
"eval_f1": 0.9893617021276596,
"eval_loss": 0.15478737652301788,
"eval_runtime": 0.7768,
"eval_samples_per_second": 525.26,
"eval_steps_per_second": 5.15,
"step": 17631
},
{
"epoch": 10.0,
"learning_rate": 4e-05,
"loss": 0.1965,
"step": 19590
},
{
"epoch": 10.0,
"eval_accuracy": 0.9877450980392157,
"eval_combined_score": 0.9894162210338681,
"eval_f1": 0.9910873440285204,
"eval_loss": 0.15827800333499908,
"eval_runtime": 0.7788,
"eval_samples_per_second": 523.897,
"eval_steps_per_second": 5.136,
"step": 19590
},
{
"epoch": 11.0,
"learning_rate": 3.9000000000000006e-05,
"loss": 0.195,
"step": 21549
},
{
"epoch": 11.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915138098250054,
"eval_f1": 0.992831541218638,
"eval_loss": 0.1527128666639328,
"eval_runtime": 0.7814,
"eval_samples_per_second": 522.15,
"eval_steps_per_second": 5.119,
"step": 21549
},
{
"epoch": 12.0,
"learning_rate": 3.8e-05,
"loss": 0.1938,
"step": 23508
},
{
"epoch": 12.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915266106442577,
"eval_f1": 0.9928571428571428,
"eval_loss": 0.15122443437576294,
"eval_runtime": 0.7775,
"eval_samples_per_second": 524.756,
"eval_steps_per_second": 5.145,
"step": 23508
},
{
"epoch": 13.0,
"learning_rate": 3.7e-05,
"loss": 0.1926,
"step": 25467
},
{
"epoch": 13.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.1425897628068924,
"eval_runtime": 0.7755,
"eval_samples_per_second": 526.144,
"eval_steps_per_second": 5.158,
"step": 25467
},
{
"epoch": 14.0,
"learning_rate": 3.6e-05,
"loss": 0.1917,
"step": 27426
},
{
"epoch": 14.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.14360927045345306,
"eval_runtime": 0.7791,
"eval_samples_per_second": 523.666,
"eval_steps_per_second": 5.134,
"step": 27426
},
{
"epoch": 15.0,
"learning_rate": 3.5e-05,
"loss": 0.191,
"step": 29385
},
{
"epoch": 15.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936401662632853,
"eval_f1": 0.9946332737030412,
"eval_loss": 0.1502537876367569,
"eval_runtime": 0.7823,
"eval_samples_per_second": 521.529,
"eval_steps_per_second": 5.113,
"step": 29385
},
{
"epoch": 16.0,
"learning_rate": 3.4000000000000007e-05,
"loss": 0.1901,
"step": 31344
},
{
"epoch": 16.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.14606404304504395,
"eval_runtime": 0.7757,
"eval_samples_per_second": 525.986,
"eval_steps_per_second": 5.157,
"step": 31344
},
{
"epoch": 17.0,
"learning_rate": 3.3e-05,
"loss": 0.1894,
"step": 33303
},
{
"epoch": 17.0,
"eval_accuracy": 0.9975490196078431,
"eval_combined_score": 0.9978800554210951,
"eval_f1": 0.998211091234347,
"eval_loss": 0.1498469114303589,
"eval_runtime": 0.7758,
"eval_samples_per_second": 525.884,
"eval_steps_per_second": 5.156,
"step": 33303
},
{
"epoch": 18.0,
"learning_rate": 3.2000000000000005e-05,
"loss": 0.1888,
"step": 35262
},
{
"epoch": 18.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915266106442577,
"eval_f1": 0.9928571428571428,
"eval_loss": 0.14023421704769135,
"eval_runtime": 0.7766,
"eval_samples_per_second": 525.375,
"eval_steps_per_second": 5.151,
"step": 35262
},
{
"epoch": 19.0,
"learning_rate": 3.1e-05,
"loss": 0.1882,
"step": 37221
},
{
"epoch": 19.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936401662632853,
"eval_f1": 0.9946332737030412,
"eval_loss": 0.14202910661697388,
"eval_runtime": 0.7794,
"eval_samples_per_second": 523.449,
"eval_steps_per_second": 5.132,
"step": 37221
},
{
"epoch": 20.0,
"learning_rate": 3e-05,
"loss": 0.1876,
"step": 39180
},
{
"epoch": 20.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.13458134233951569,
"eval_runtime": 0.7781,
"eval_samples_per_second": 524.385,
"eval_steps_per_second": 5.141,
"step": 39180
},
{
"epoch": 21.0,
"learning_rate": 2.9e-05,
"loss": 0.1871,
"step": 41139
},
{
"epoch": 21.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.13956362009048462,
"eval_runtime": 0.7768,
"eval_samples_per_second": 525.249,
"eval_steps_per_second": 5.15,
"step": 41139
},
{
"epoch": 22.0,
"learning_rate": 2.8000000000000003e-05,
"loss": 0.1867,
"step": 43098
},
{
"epoch": 22.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.14428994059562683,
"eval_runtime": 0.7783,
"eval_samples_per_second": 524.192,
"eval_steps_per_second": 5.139,
"step": 43098
},
{
"epoch": 23.0,
"learning_rate": 2.7000000000000002e-05,
"loss": 0.1862,
"step": 45057
},
{
"epoch": 23.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.1346110701560974,
"eval_runtime": 0.7827,
"eval_samples_per_second": 521.253,
"eval_steps_per_second": 5.11,
"step": 45057
},
{
"epoch": 24.0,
"learning_rate": 2.6000000000000002e-05,
"loss": 0.1857,
"step": 47016
},
{
"epoch": 24.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.13611076772212982,
"eval_runtime": 0.7925,
"eval_samples_per_second": 514.847,
"eval_steps_per_second": 5.048,
"step": 47016
},
{
"epoch": 25.0,
"learning_rate": 2.5e-05,
"loss": 0.1854,
"step": 48975
},
{
"epoch": 25.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.13179393112659454,
"eval_runtime": 0.7822,
"eval_samples_per_second": 521.594,
"eval_steps_per_second": 5.114,
"step": 48975
},
{
"epoch": 26.0,
"learning_rate": 2.4e-05,
"loss": 0.185,
"step": 50934
},
{
"epoch": 26.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.13099727034568787,
"eval_runtime": 0.851,
"eval_samples_per_second": 479.464,
"eval_steps_per_second": 4.701,
"step": 50934
},
{
"epoch": 27.0,
"learning_rate": 2.3000000000000003e-05,
"loss": 0.1846,
"step": 52893
},
{
"epoch": 27.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.13022534549236298,
"eval_runtime": 0.7818,
"eval_samples_per_second": 521.85,
"eval_steps_per_second": 5.116,
"step": 52893
},
{
"epoch": 28.0,
"learning_rate": 2.2000000000000003e-05,
"loss": 0.1842,
"step": 54852
},
{
"epoch": 28.0,
"eval_accuracy": 0.9950980392156863,
"eval_combined_score": 0.9957633053221289,
"eval_f1": 0.9964285714285714,
"eval_loss": 0.13285598158836365,
"eval_runtime": 0.7818,
"eval_samples_per_second": 521.866,
"eval_steps_per_second": 5.116,
"step": 54852
},
{
"epoch": 29.0,
"learning_rate": 2.1e-05,
"loss": 0.1839,
"step": 56811
},
{
"epoch": 29.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.13004697859287262,
"eval_runtime": 0.7761,
"eval_samples_per_second": 525.714,
"eval_steps_per_second": 5.154,
"step": 56811
},
{
"epoch": 30.0,
"learning_rate": 2e-05,
"loss": 0.1836,
"step": 58770
},
{
"epoch": 30.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.1328008770942688,
"eval_runtime": 0.7782,
"eval_samples_per_second": 524.253,
"eval_steps_per_second": 5.14,
"step": 58770
},
{
"epoch": 31.0,
"learning_rate": 1.9e-05,
"loss": 0.1832,
"step": 60729
},
{
"epoch": 31.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.1326626092195511,
"eval_runtime": 0.7751,
"eval_samples_per_second": 526.407,
"eval_steps_per_second": 5.161,
"step": 60729
},
{
"epoch": 32.0,
"learning_rate": 1.8e-05,
"loss": 0.1829,
"step": 62688
},
{
"epoch": 32.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.13079187273979187,
"eval_runtime": 0.7779,
"eval_samples_per_second": 524.516,
"eval_steps_per_second": 5.142,
"step": 62688
},
{
"epoch": 33.0,
"learning_rate": 1.7000000000000003e-05,
"loss": 0.1826,
"step": 64647
},
{
"epoch": 33.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.12873521447181702,
"eval_runtime": 0.7782,
"eval_samples_per_second": 524.295,
"eval_steps_per_second": 5.14,
"step": 64647
},
{
"epoch": 34.0,
"learning_rate": 1.6000000000000003e-05,
"loss": 0.1824,
"step": 66606
},
{
"epoch": 34.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.1308642029762268,
"eval_runtime": 0.7782,
"eval_samples_per_second": 524.283,
"eval_steps_per_second": 5.14,
"step": 66606
},
{
"epoch": 35.0,
"learning_rate": 1.5e-05,
"loss": 0.1821,
"step": 68565
},
{
"epoch": 35.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.13088451325893402,
"eval_runtime": 0.7758,
"eval_samples_per_second": 525.926,
"eval_steps_per_second": 5.156,
"step": 68565
},
{
"epoch": 36.0,
"learning_rate": 1.4000000000000001e-05,
"loss": 0.1818,
"step": 70524
},
{
"epoch": 36.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.1271485686302185,
"eval_runtime": 0.7809,
"eval_samples_per_second": 522.489,
"eval_steps_per_second": 5.122,
"step": 70524
},
{
"epoch": 37.0,
"learning_rate": 1.3000000000000001e-05,
"loss": 0.1816,
"step": 72483
},
{
"epoch": 37.0,
"eval_accuracy": 0.9877450980392157,
"eval_combined_score": 0.9894320516839064,
"eval_f1": 0.9911190053285969,
"eval_loss": 0.127828910946846,
"eval_runtime": 0.7755,
"eval_samples_per_second": 526.078,
"eval_steps_per_second": 5.158,
"step": 72483
},
{
"epoch": 38.0,
"learning_rate": 1.2e-05,
"loss": 0.1813,
"step": 74442
},
{
"epoch": 38.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.1280445158481598,
"eval_runtime": 0.7777,
"eval_samples_per_second": 524.647,
"eval_steps_per_second": 5.144,
"step": 74442
},
{
"epoch": 39.0,
"learning_rate": 1.1000000000000001e-05,
"loss": 0.1811,
"step": 76401
},
{
"epoch": 39.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.12891501188278198,
"eval_runtime": 0.7798,
"eval_samples_per_second": 523.193,
"eval_steps_per_second": 5.129,
"step": 76401
},
{
"epoch": 40.0,
"learning_rate": 1e-05,
"loss": 0.1809,
"step": 78360
},
{
"epoch": 40.0,
"eval_accuracy": 0.9877450980392157,
"eval_combined_score": 0.9894320516839064,
"eval_f1": 0.9911190053285969,
"eval_loss": 0.12895023822784424,
"eval_runtime": 0.7761,
"eval_samples_per_second": 525.732,
"eval_steps_per_second": 5.154,
"step": 78360
},
{
"epoch": 41.0,
"learning_rate": 9e-06,
"loss": 0.1807,
"step": 80319
},
{
"epoch": 41.0,
"eval_accuracy": 0.9877450980392157,
"eval_combined_score": 0.9894320516839064,
"eval_f1": 0.9911190053285969,
"eval_loss": 0.1256455034017563,
"eval_runtime": 0.7742,
"eval_samples_per_second": 526.998,
"eval_steps_per_second": 5.167,
"step": 80319
},
{
"epoch": 42.0,
"learning_rate": 8.000000000000001e-06,
"loss": 0.1805,
"step": 82278
},
{
"epoch": 42.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.12676292657852173,
"eval_runtime": 0.7808,
"eval_samples_per_second": 522.525,
"eval_steps_per_second": 5.123,
"step": 82278
},
{
"epoch": 43.0,
"learning_rate": 7.000000000000001e-06,
"loss": 0.1803,
"step": 84237
},
{
"epoch": 43.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.12742425501346588,
"eval_runtime": 0.7744,
"eval_samples_per_second": 526.85,
"eval_steps_per_second": 5.165,
"step": 84237
},
{
"epoch": 44.0,
"learning_rate": 6e-06,
"loss": 0.1801,
"step": 86196
},
{
"epoch": 44.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.12773118913173676,
"eval_runtime": 0.7768,
"eval_samples_per_second": 525.233,
"eval_steps_per_second": 5.149,
"step": 86196
},
{
"epoch": 45.0,
"learning_rate": 5e-06,
"loss": 0.1799,
"step": 88155
},
{
"epoch": 45.0,
"eval_accuracy": 0.9926470588235294,
"eval_combined_score": 0.9936497326203209,
"eval_f1": 0.9946524064171123,
"eval_loss": 0.12637551128864288,
"eval_runtime": 0.7756,
"eval_samples_per_second": 526.076,
"eval_steps_per_second": 5.158,
"step": 88155
},
{
"epoch": 46.0,
"learning_rate": 4.000000000000001e-06,
"loss": 0.1797,
"step": 90114
},
{
"epoch": 46.0,
"eval_accuracy": 0.9901960784313726,
"eval_combined_score": 0.9915393203544763,
"eval_f1": 0.9928825622775801,
"eval_loss": 0.12735696136951447,
"eval_runtime": 0.7805,
"eval_samples_per_second": 522.735,
"eval_steps_per_second": 5.125,
"step": 90114
},
{
"epoch": 46.0,
"step": 90114,
"total_flos": 3.6163552369495245e+17,
"train_loss": 0.19228331322046327,
"train_runtime": 51035.534,
"train_samples_per_second": 245.648,
"train_steps_per_second": 1.919
}
],
"max_steps": 97950,
"num_train_epochs": 50,
"total_flos": 3.6163552369495245e+17,
"trial_name": null,
"trial_params": null
}