mistral-7b-ift / trainer_state.json
NicholasCorrado's picture
Model save
35884d0 verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9989023051591658,
"eval_steps": 500,
"global_step": 455,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0021953896816684962,
"grad_norm": 37.21439565141779,
"learning_rate": 1.0869565217391303e-08,
"loss": 1.4277,
"step": 1
},
{
"epoch": 0.010976948408342482,
"grad_norm": 31.799276785765738,
"learning_rate": 5.434782608695652e-08,
"loss": 1.4875,
"step": 5
},
{
"epoch": 0.021953896816684963,
"grad_norm": 37.128332006170986,
"learning_rate": 1.0869565217391303e-07,
"loss": 1.4405,
"step": 10
},
{
"epoch": 0.03293084522502744,
"grad_norm": 28.763704469610055,
"learning_rate": 1.6304347826086955e-07,
"loss": 1.4495,
"step": 15
},
{
"epoch": 0.043907793633369926,
"grad_norm": 24.21109200751535,
"learning_rate": 2.1739130434782607e-07,
"loss": 1.348,
"step": 20
},
{
"epoch": 0.054884742041712405,
"grad_norm": 17.806110439198463,
"learning_rate": 2.717391304347826e-07,
"loss": 1.2804,
"step": 25
},
{
"epoch": 0.06586169045005488,
"grad_norm": 7.093114542451441,
"learning_rate": 3.260869565217391e-07,
"loss": 1.2082,
"step": 30
},
{
"epoch": 0.07683863885839737,
"grad_norm": 4.867218022023519,
"learning_rate": 3.8043478260869567e-07,
"loss": 1.1694,
"step": 35
},
{
"epoch": 0.08781558726673985,
"grad_norm": 4.163949798778549,
"learning_rate": 4.3478260869565214e-07,
"loss": 1.1523,
"step": 40
},
{
"epoch": 0.09879253567508232,
"grad_norm": 3.75376057274967,
"learning_rate": 4.891304347826087e-07,
"loss": 1.1297,
"step": 45
},
{
"epoch": 0.10976948408342481,
"grad_norm": 3.964842079535969,
"learning_rate": 4.998820089787287e-07,
"loss": 1.108,
"step": 50
},
{
"epoch": 0.1207464324917673,
"grad_norm": 2.7520967442544055,
"learning_rate": 4.99402861330474e-07,
"loss": 1.1033,
"step": 55
},
{
"epoch": 0.13172338090010977,
"grad_norm": 3.5589483412826977,
"learning_rate": 4.985558887302488e-07,
"loss": 1.0951,
"step": 60
},
{
"epoch": 0.14270032930845225,
"grad_norm": 3.119698599724887,
"learning_rate": 4.973423403123028e-07,
"loss": 1.0783,
"step": 65
},
{
"epoch": 0.15367727771679474,
"grad_norm": 2.949317383115411,
"learning_rate": 4.957640058451501e-07,
"loss": 1.0669,
"step": 70
},
{
"epoch": 0.16465422612513722,
"grad_norm": 2.618140408739008,
"learning_rate": 4.938232130919785e-07,
"loss": 1.0726,
"step": 75
},
{
"epoch": 0.1756311745334797,
"grad_norm": 2.6209648500734697,
"learning_rate": 4.915228243776124e-07,
"loss": 1.0662,
"step": 80
},
{
"epoch": 0.18660812294182216,
"grad_norm": 2.867532808719891,
"learning_rate": 4.888662323670913e-07,
"loss": 1.0555,
"step": 85
},
{
"epoch": 0.19758507135016465,
"grad_norm": 2.762665890450646,
"learning_rate": 4.858573550620908e-07,
"loss": 1.0669,
"step": 90
},
{
"epoch": 0.20856201975850713,
"grad_norm": 2.4214228636048833,
"learning_rate": 4.825006300225665e-07,
"loss": 1.0714,
"step": 95
},
{
"epoch": 0.21953896816684962,
"grad_norm": 2.316773071192448,
"learning_rate": 4.7880100782214e-07,
"loss": 1.0594,
"step": 100
},
{
"epoch": 0.2305159165751921,
"grad_norm": 2.2001269225986864,
"learning_rate": 4.747639447468816e-07,
"loss": 1.0621,
"step": 105
},
{
"epoch": 0.2414928649835346,
"grad_norm": 2.4429626086547014,
"learning_rate": 4.7039539474825683e-07,
"loss": 1.0262,
"step": 110
},
{
"epoch": 0.2524698133918771,
"grad_norm": 2.505629215080317,
"learning_rate": 4.657018006621053e-07,
"loss": 1.0288,
"step": 115
},
{
"epoch": 0.26344676180021953,
"grad_norm": 2.34925727459466,
"learning_rate": 4.6069008470660057e-07,
"loss": 1.0671,
"step": 120
},
{
"epoch": 0.27442371020856204,
"grad_norm": 2.1952559509776575,
"learning_rate": 4.5536763827320803e-07,
"loss": 1.033,
"step": 125
},
{
"epoch": 0.2854006586169045,
"grad_norm": 2.2990824111311987,
"learning_rate": 4.4974231102569355e-07,
"loss": 1.0245,
"step": 130
},
{
"epoch": 0.29637760702524696,
"grad_norm": 2.2303125915616313,
"learning_rate": 4.438223993232634e-07,
"loss": 1.0095,
"step": 135
},
{
"epoch": 0.30735455543358947,
"grad_norm": 2.2638284322428412,
"learning_rate": 4.3761663398490634e-07,
"loss": 1.016,
"step": 140
},
{
"epoch": 0.31833150384193193,
"grad_norm": 2.4912960841051657,
"learning_rate": 4.3113416741298616e-07,
"loss": 1.0232,
"step": 145
},
{
"epoch": 0.32930845225027444,
"grad_norm": 2.35071868002771,
"learning_rate": 4.2438456009507195e-07,
"loss": 1.0094,
"step": 150
},
{
"epoch": 0.3402854006586169,
"grad_norm": 2.325666914884671,
"learning_rate": 4.1737776650391625e-07,
"loss": 1.0186,
"step": 155
},
{
"epoch": 0.3512623490669594,
"grad_norm": 2.492980306835079,
"learning_rate": 4.101241204163748e-07,
"loss": 0.9975,
"step": 160
},
{
"epoch": 0.36223929747530187,
"grad_norm": 2.2095279955781835,
"learning_rate": 4.0263431967291934e-07,
"loss": 1.0269,
"step": 165
},
{
"epoch": 0.3732162458836443,
"grad_norm": 2.200947139945517,
"learning_rate": 3.949194104002224e-07,
"loss": 1.0078,
"step": 170
},
{
"epoch": 0.38419319429198684,
"grad_norm": 2.2892065158763333,
"learning_rate": 3.8699077072008085e-07,
"loss": 1.0105,
"step": 175
},
{
"epoch": 0.3951701427003293,
"grad_norm": 2.3262445212448912,
"learning_rate": 3.7886009396870564e-07,
"loss": 1.0032,
"step": 180
},
{
"epoch": 0.4061470911086718,
"grad_norm": 3.3437108002828713,
"learning_rate": 3.705393714511268e-07,
"loss": 1.0007,
"step": 185
},
{
"epoch": 0.41712403951701427,
"grad_norm": 2.271617618137847,
"learning_rate": 3.6204087475614676e-07,
"loss": 1.0161,
"step": 190
},
{
"epoch": 0.4281009879253567,
"grad_norm": 2.2040001544409704,
"learning_rate": 3.533771376579249e-07,
"loss": 1.0115,
"step": 195
},
{
"epoch": 0.43907793633369924,
"grad_norm": 2.208144019801884,
"learning_rate": 3.445609376308857e-07,
"loss": 0.9991,
"step": 200
},
{
"epoch": 0.4500548847420417,
"grad_norm": 2.2625225642348608,
"learning_rate": 3.356052770052119e-07,
"loss": 1.0002,
"step": 205
},
{
"epoch": 0.4610318331503842,
"grad_norm": 2.17968991949006,
"learning_rate": 3.2652336379071506e-07,
"loss": 1.003,
"step": 210
},
{
"epoch": 0.47200878155872666,
"grad_norm": 2.2484730404674367,
"learning_rate": 3.173285921973657e-07,
"loss": 0.9967,
"step": 215
},
{
"epoch": 0.4829857299670692,
"grad_norm": 2.2023781045941297,
"learning_rate": 3.0803452288121113e-07,
"loss": 0.9934,
"step": 220
},
{
"epoch": 0.49396267837541163,
"grad_norm": 2.1760991166567925,
"learning_rate": 2.986548629448146e-07,
"loss": 1.0118,
"step": 225
},
{
"epoch": 0.5049396267837541,
"grad_norm": 2.1634685814392567,
"learning_rate": 2.892034457217119e-07,
"loss": 0.9803,
"step": 230
},
{
"epoch": 0.5159165751920965,
"grad_norm": 2.1190899654664324,
"learning_rate": 2.7969421037470033e-07,
"loss": 1.0109,
"step": 235
},
{
"epoch": 0.5268935236004391,
"grad_norm": 2.2152330549300867,
"learning_rate": 2.70141181338047e-07,
"loss": 0.9835,
"step": 240
},
{
"epoch": 0.5378704720087816,
"grad_norm": 2.274189691558315,
"learning_rate": 2.60558447633938e-07,
"loss": 0.9994,
"step": 245
},
{
"epoch": 0.5488474204171241,
"grad_norm": 2.242337338350247,
"learning_rate": 2.509601420936717e-07,
"loss": 1.0056,
"step": 250
},
{
"epoch": 0.5598243688254665,
"grad_norm": 2.334006849829965,
"learning_rate": 2.413604205142404e-07,
"loss": 0.9976,
"step": 255
},
{
"epoch": 0.570801317233809,
"grad_norm": 2.761308414970569,
"learning_rate": 2.3177344078104305e-07,
"loss": 1.0227,
"step": 260
},
{
"epoch": 0.5817782656421515,
"grad_norm": 2.33919437556703,
"learning_rate": 2.2221334198751717e-07,
"loss": 0.9997,
"step": 265
},
{
"epoch": 0.5927552140504939,
"grad_norm": 2.20272202664836,
"learning_rate": 2.1269422358248534e-07,
"loss": 0.9765,
"step": 270
},
{
"epoch": 0.6037321624588364,
"grad_norm": 2.417200784540872,
"learning_rate": 2.0323012457597113e-07,
"loss": 0.9912,
"step": 275
},
{
"epoch": 0.6147091108671789,
"grad_norm": 2.144247790210012,
"learning_rate": 1.9383500283415127e-07,
"loss": 1.0085,
"step": 280
},
{
"epoch": 0.6256860592755215,
"grad_norm": 2.1598500041507576,
"learning_rate": 1.8452271449398015e-07,
"loss": 0.9936,
"step": 285
},
{
"epoch": 0.6366630076838639,
"grad_norm": 2.2513888869939875,
"learning_rate": 1.753069935278477e-07,
"loss": 0.9983,
"step": 290
},
{
"epoch": 0.6476399560922064,
"grad_norm": 2.21672483120269,
"learning_rate": 1.662014314884074e-07,
"loss": 0.9823,
"step": 295
},
{
"epoch": 0.6586169045005489,
"grad_norm": 2.167838745580858,
"learning_rate": 1.5721945746344914e-07,
"loss": 1.003,
"step": 300
},
{
"epoch": 0.6695938529088913,
"grad_norm": 2.2265387290087824,
"learning_rate": 1.4837431827037786e-07,
"loss": 0.9844,
"step": 305
},
{
"epoch": 0.6805708013172338,
"grad_norm": 2.210867575625144,
"learning_rate": 1.3967905891950936e-07,
"loss": 0.9894,
"step": 310
},
{
"epoch": 0.6915477497255763,
"grad_norm": 2.221663064011151,
"learning_rate": 1.3114650337499578e-07,
"loss": 0.9932,
"step": 315
},
{
"epoch": 0.7025246981339188,
"grad_norm": 2.207207427668881,
"learning_rate": 1.227892356417542e-07,
"loss": 0.9801,
"step": 320
},
{
"epoch": 0.7135016465422612,
"grad_norm": 2.229969820276736,
"learning_rate": 1.1461958120629345e-07,
"loss": 0.9749,
"step": 325
},
{
"epoch": 0.7244785949506037,
"grad_norm": 2.2050752947618797,
"learning_rate": 1.0664958885880901e-07,
"loss": 0.9995,
"step": 330
},
{
"epoch": 0.7354555433589463,
"grad_norm": 2.220875622216501,
"learning_rate": 9.889101292335625e-08,
"loss": 0.9882,
"step": 335
},
{
"epoch": 0.7464324917672887,
"grad_norm": 2.1962028300045415,
"learning_rate": 9.13552959223089e-08,
"loss": 0.9963,
"step": 340
},
{
"epoch": 0.7574094401756312,
"grad_norm": 2.2493255089237665,
"learning_rate": 8.405355170066925e-08,
"loss": 0.9945,
"step": 345
},
{
"epoch": 0.7683863885839737,
"grad_norm": 2.194427672592623,
"learning_rate": 7.699654903511971e-08,
"loss": 0.9868,
"step": 350
},
{
"epoch": 0.7793633369923162,
"grad_norm": 2.1546138188685906,
"learning_rate": 7.01946957519886e-08,
"loss": 0.9761,
"step": 355
},
{
"epoch": 0.7903402854006586,
"grad_norm": 2.131901247198089,
"learning_rate": 6.365802337755364e-08,
"loss": 1.0021,
"step": 360
},
{
"epoch": 0.8013172338090011,
"grad_norm": 2.35581904723087,
"learning_rate": 5.739617234332131e-08,
"loss": 0.9962,
"step": 365
},
{
"epoch": 0.8122941822173436,
"grad_norm": 2.1749901587074336,
"learning_rate": 5.14183777681014e-08,
"loss": 0.9792,
"step": 370
},
{
"epoch": 0.823271130625686,
"grad_norm": 2.270280240224792,
"learning_rate": 4.5733455837846325e-08,
"loss": 1.0028,
"step": 375
},
{
"epoch": 0.8342480790340285,
"grad_norm": 2.160330737233165,
"learning_rate": 4.0349790803341274e-08,
"loss": 1.0058,
"step": 380
},
{
"epoch": 0.845225027442371,
"grad_norm": 2.1252739289722427,
"learning_rate": 3.527532261492272e-08,
"loss": 0.9877,
"step": 385
},
{
"epoch": 0.8562019758507134,
"grad_norm": 2.118987787417721,
"learning_rate": 3.0517535212460946e-08,
"loss": 0.9786,
"step": 390
},
{
"epoch": 0.867178924259056,
"grad_norm": 2.1714044073053373,
"learning_rate": 2.608344548787722e-08,
"loss": 0.9714,
"step": 395
},
{
"epoch": 0.8781558726673985,
"grad_norm": 2.185899254100342,
"learning_rate": 2.19795929364735e-08,
"loss": 0.9897,
"step": 400
},
{
"epoch": 0.889132821075741,
"grad_norm": 2.1647439375274438,
"learning_rate": 1.8212030012337704e-08,
"loss": 0.9983,
"step": 405
},
{
"epoch": 0.9001097694840834,
"grad_norm": 2.149723670222018,
"learning_rate": 1.4786313202048456e-08,
"loss": 1.0078,
"step": 410
},
{
"epoch": 0.9110867178924259,
"grad_norm": 2.0770986497941584,
"learning_rate": 1.1707494829843207e-08,
"loss": 0.9856,
"step": 415
},
{
"epoch": 0.9220636663007684,
"grad_norm": 2.127123760568819,
"learning_rate": 8.980115606337046e-09,
"loss": 0.9959,
"step": 420
},
{
"epoch": 0.9330406147091108,
"grad_norm": 2.119993193338152,
"learning_rate": 6.608197931780496e-09,
"loss": 0.9771,
"step": 425
},
{
"epoch": 0.9440175631174533,
"grad_norm": 2.1728444045432176,
"learning_rate": 4.595239963733011e-09,
"loss": 0.9854,
"step": 430
},
{
"epoch": 0.9549945115257958,
"grad_norm": 2.1073358990092603,
"learning_rate": 2.9442104579016356e-09,
"loss": 0.9918,
"step": 435
},
{
"epoch": 0.9659714599341384,
"grad_norm": 2.2775104380213635,
"learning_rate": 1.6575443897531294e-09,
"loss": 0.9911,
"step": 440
},
{
"epoch": 0.9769484083424808,
"grad_norm": 2.1495528942632705,
"learning_rate": 7.371393633574252e-10,
"loss": 0.9846,
"step": 445
},
{
"epoch": 0.9879253567508233,
"grad_norm": 2.3281869972319096,
"learning_rate": 1.843528127584981e-10,
"loss": 0.9961,
"step": 450
},
{
"epoch": 0.9989023051591658,
"grad_norm": 2.22432656366925,
"learning_rate": 0.0,
"loss": 0.9719,
"step": 455
},
{
"epoch": 0.9989023051591658,
"eval_loss": 1.0896308422088623,
"eval_runtime": 3.6044,
"eval_samples_per_second": 74.353,
"eval_steps_per_second": 1.387,
"step": 455
},
{
"epoch": 0.9989023051591658,
"step": 455,
"total_flos": 190430796840960.0,
"train_loss": 1.0385671806859447,
"train_runtime": 5397.8209,
"train_samples_per_second": 21.597,
"train_steps_per_second": 0.084
}
],
"logging_steps": 5,
"max_steps": 455,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 190430796840960.0,
"train_batch_size": 8,
"trial_name": null,
"trial_params": null
}