1475 lines
44 KiB
JSON
1475 lines
44 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 1.7797930343829975,
|
|
"eval_steps": 100,
|
|
"global_step": 1000,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"entropy": 1.3253214355558156,
|
|
"epoch": 0.017803493935684877,
|
|
"grad_norm": 14.5,
|
|
"learning_rate": 6.000000000000001e-07,
|
|
"loss": 2.1959,
|
|
"mean_token_accuracy": 0.5975642748177051,
|
|
"num_tokens": 259048.0,
|
|
"step": 10
|
|
},
|
|
{
|
|
"entropy": 1.3036176152527332,
|
|
"epoch": 0.035606987871369754,
|
|
"grad_norm": 12.75,
|
|
"learning_rate": 1.2666666666666669e-06,
|
|
"loss": 2.1499,
|
|
"mean_token_accuracy": 0.6023047739639878,
|
|
"num_tokens": 532019.0,
|
|
"step": 20
|
|
},
|
|
{
|
|
"entropy": 1.3484113823622466,
|
|
"epoch": 0.053410481807054634,
|
|
"grad_norm": 10.8125,
|
|
"learning_rate": 1.9333333333333336e-06,
|
|
"loss": 2.108,
|
|
"mean_token_accuracy": 0.6078463492915034,
|
|
"num_tokens": 807220.0,
|
|
"step": 30
|
|
},
|
|
{
|
|
"entropy": 1.4311681490391492,
|
|
"epoch": 0.07121397574273951,
|
|
"grad_norm": 7.875,
|
|
"learning_rate": 2.6e-06,
|
|
"loss": 2.0349,
|
|
"mean_token_accuracy": 0.6107733339071274,
|
|
"num_tokens": 1068429.0,
|
|
"step": 40
|
|
},
|
|
{
|
|
"entropy": 1.479667642340064,
|
|
"epoch": 0.08901746967842439,
|
|
"grad_norm": 6.4375,
|
|
"learning_rate": 3.266666666666667e-06,
|
|
"loss": 1.8649,
|
|
"mean_token_accuracy": 0.630409375578165,
|
|
"num_tokens": 1336698.0,
|
|
"step": 50
|
|
},
|
|
{
|
|
"entropy": 1.5030938494950532,
|
|
"epoch": 0.10682096361410927,
|
|
"grad_norm": 4.1875,
|
|
"learning_rate": 3.9333333333333335e-06,
|
|
"loss": 1.7217,
|
|
"mean_token_accuracy": 0.6490663634613156,
|
|
"num_tokens": 1608574.0,
|
|
"step": 60
|
|
},
|
|
{
|
|
"entropy": 1.5066865295171739,
|
|
"epoch": 0.12462445754979415,
|
|
"grad_norm": 2.71875,
|
|
"learning_rate": 4.600000000000001e-06,
|
|
"loss": 1.5716,
|
|
"mean_token_accuracy": 0.6711439102888107,
|
|
"num_tokens": 1872747.0,
|
|
"step": 70
|
|
},
|
|
{
|
|
"entropy": 1.3747932206839324,
|
|
"epoch": 0.14242795148547902,
|
|
"grad_norm": 2.234375,
|
|
"learning_rate": 5.2666666666666665e-06,
|
|
"loss": 1.4403,
|
|
"mean_token_accuracy": 0.6936415441334247,
|
|
"num_tokens": 2133173.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"entropy": 1.1736511185765266,
|
|
"epoch": 0.1602314454211639,
|
|
"grad_norm": 1.921875,
|
|
"learning_rate": 5.933333333333335e-06,
|
|
"loss": 1.2902,
|
|
"mean_token_accuracy": 0.7222886722534895,
|
|
"num_tokens": 2406145.0,
|
|
"step": 90
|
|
},
|
|
{
|
|
"entropy": 1.1317794900387526,
|
|
"epoch": 0.17803493935684878,
|
|
"grad_norm": 1.8046875,
|
|
"learning_rate": 6.600000000000001e-06,
|
|
"loss": 1.2424,
|
|
"mean_token_accuracy": 0.7295983772724867,
|
|
"num_tokens": 2663709.0,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.17803493935684878,
|
|
"eval_biology_entropy": 2.3133783559799195,
|
|
"eval_biology_loss": 2.7910587787628174,
|
|
"eval_biology_mean_token_accuracy": 0.5260232911109924,
|
|
"eval_biology_num_tokens": 2663709.0,
|
|
"eval_biology_runtime": 39.1656,
|
|
"eval_biology_samples_per_second": 12.766,
|
|
"eval_biology_steps_per_second": 3.192,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.17803493935684878,
|
|
"eval_chemistry_entropy": 1.0788086318969727,
|
|
"eval_chemistry_loss": 1.2403336763381958,
|
|
"eval_chemistry_mean_token_accuracy": 0.7357268543243408,
|
|
"eval_chemistry_num_tokens": 2663709.0,
|
|
"eval_chemistry_runtime": 48.7822,
|
|
"eval_chemistry_samples_per_second": 10.25,
|
|
"eval_chemistry_steps_per_second": 2.562,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.17803493935684878,
|
|
"eval_math_entropy": 0.938195451259613,
|
|
"eval_math_loss": 1.3248047828674316,
|
|
"eval_math_mean_token_accuracy": 0.7312402629852295,
|
|
"eval_math_num_tokens": 2663709.0,
|
|
"eval_math_runtime": 50.197,
|
|
"eval_math_samples_per_second": 9.961,
|
|
"eval_math_steps_per_second": 2.49,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.17803493935684878,
|
|
"eval_physics_entropy": 1.022246531009674,
|
|
"eval_physics_loss": 1.176153302192688,
|
|
"eval_physics_mean_token_accuracy": 0.7475795021057129,
|
|
"eval_physics_num_tokens": 2663709.0,
|
|
"eval_physics_runtime": 57.6895,
|
|
"eval_physics_samples_per_second": 8.667,
|
|
"eval_physics_steps_per_second": 2.167,
|
|
"step": 100
|
|
},
|
|
{
|
|
"entropy": 1.0380497876554728,
|
|
"epoch": 0.19583843329253367,
|
|
"grad_norm": 1.7421875,
|
|
"learning_rate": 7.266666666666668e-06,
|
|
"loss": 1.157,
|
|
"mean_token_accuracy": 0.7445634145289659,
|
|
"num_tokens": 2934980.0,
|
|
"step": 110
|
|
},
|
|
{
|
|
"entropy": 1.0149445166811346,
|
|
"epoch": 0.21364192722821854,
|
|
"grad_norm": 1.6015625,
|
|
"learning_rate": 7.933333333333334e-06,
|
|
"loss": 1.1167,
|
|
"mean_token_accuracy": 0.7485828153789044,
|
|
"num_tokens": 3208702.0,
|
|
"step": 120
|
|
},
|
|
{
|
|
"entropy": 0.9639404848217964,
|
|
"epoch": 0.2314454211639034,
|
|
"grad_norm": 1.6953125,
|
|
"learning_rate": 8.6e-06,
|
|
"loss": 1.0979,
|
|
"mean_token_accuracy": 0.7553199239075183,
|
|
"num_tokens": 3480119.0,
|
|
"step": 130
|
|
},
|
|
{
|
|
"entropy": 0.9687324531376362,
|
|
"epoch": 0.2492489150995883,
|
|
"grad_norm": 1.5625,
|
|
"learning_rate": 9.266666666666667e-06,
|
|
"loss": 1.0762,
|
|
"mean_token_accuracy": 0.757096454873681,
|
|
"num_tokens": 3746392.0,
|
|
"step": 140
|
|
},
|
|
{
|
|
"entropy": 0.9682734729722142,
|
|
"epoch": 0.26705240903527316,
|
|
"grad_norm": 1.875,
|
|
"learning_rate": 9.933333333333334e-06,
|
|
"loss": 1.0798,
|
|
"mean_token_accuracy": 0.7543047085404396,
|
|
"num_tokens": 4019487.0,
|
|
"step": 150
|
|
},
|
|
{
|
|
"entropy": 0.9270784268155694,
|
|
"epoch": 0.28485590297095803,
|
|
"grad_norm": 1.6015625,
|
|
"learning_rate": 1.0600000000000002e-05,
|
|
"loss": 1.047,
|
|
"mean_token_accuracy": 0.7619738396257162,
|
|
"num_tokens": 4298100.0,
|
|
"step": 160
|
|
},
|
|
{
|
|
"entropy": 0.9063327137380839,
|
|
"epoch": 0.30265939690664295,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 1.1266666666666668e-05,
|
|
"loss": 1.0199,
|
|
"mean_token_accuracy": 0.7664302848279476,
|
|
"num_tokens": 4568116.0,
|
|
"step": 170
|
|
},
|
|
{
|
|
"entropy": 0.9037324048578739,
|
|
"epoch": 0.3204628908423278,
|
|
"grad_norm": 1.7109375,
|
|
"learning_rate": 1.1933333333333335e-05,
|
|
"loss": 1.0144,
|
|
"mean_token_accuracy": 0.7657378144562245,
|
|
"num_tokens": 4831334.0,
|
|
"step": 180
|
|
},
|
|
{
|
|
"entropy": 0.8886868219822646,
|
|
"epoch": 0.3382663847780127,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 1.2600000000000001e-05,
|
|
"loss": 0.9874,
|
|
"mean_token_accuracy": 0.7726409930735827,
|
|
"num_tokens": 5103301.0,
|
|
"step": 190
|
|
},
|
|
{
|
|
"entropy": 0.8936551520600915,
|
|
"epoch": 0.35606987871369755,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 1.3266666666666668e-05,
|
|
"loss": 1.019,
|
|
"mean_token_accuracy": 0.7648640714585782,
|
|
"num_tokens": 5362915.0,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.35606987871369755,
|
|
"eval_biology_entropy": 2.0171296310424807,
|
|
"eval_biology_loss": 2.3780677318573,
|
|
"eval_biology_mean_token_accuracy": 0.5706260974407196,
|
|
"eval_biology_num_tokens": 5362915.0,
|
|
"eval_biology_runtime": 39.1203,
|
|
"eval_biology_samples_per_second": 12.781,
|
|
"eval_biology_steps_per_second": 3.195,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.35606987871369755,
|
|
"eval_chemistry_entropy": 0.8995157589912415,
|
|
"eval_chemistry_loss": 1.022930383682251,
|
|
"eval_chemistry_mean_token_accuracy": 0.7701377034187317,
|
|
"eval_chemistry_num_tokens": 5362915.0,
|
|
"eval_chemistry_runtime": 48.7263,
|
|
"eval_chemistry_samples_per_second": 10.261,
|
|
"eval_chemistry_steps_per_second": 2.565,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.35606987871369755,
|
|
"eval_math_entropy": 0.8352043969631195,
|
|
"eval_math_loss": 1.1696399450302124,
|
|
"eval_math_mean_token_accuracy": 0.7531440143585205,
|
|
"eval_math_num_tokens": 5362915.0,
|
|
"eval_math_runtime": 50.1494,
|
|
"eval_math_samples_per_second": 9.97,
|
|
"eval_math_steps_per_second": 2.493,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.35606987871369755,
|
|
"eval_physics_entropy": 0.8889413847923279,
|
|
"eval_physics_loss": 0.9962027072906494,
|
|
"eval_physics_mean_token_accuracy": 0.775176504611969,
|
|
"eval_physics_num_tokens": 5362915.0,
|
|
"eval_physics_runtime": 57.6533,
|
|
"eval_physics_samples_per_second": 8.673,
|
|
"eval_physics_steps_per_second": 2.168,
|
|
"step": 200
|
|
},
|
|
{
|
|
"entropy": 0.8509893301874399,
|
|
"epoch": 0.3738733726493824,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 1.3933333333333334e-05,
|
|
"loss": 0.9619,
|
|
"mean_token_accuracy": 0.7772600032389164,
|
|
"num_tokens": 5625455.0,
|
|
"step": 210
|
|
},
|
|
{
|
|
"entropy": 0.8560685662552714,
|
|
"epoch": 0.39167686658506734,
|
|
"grad_norm": 1.4609375,
|
|
"learning_rate": 1.46e-05,
|
|
"loss": 0.9666,
|
|
"mean_token_accuracy": 0.7764565829187632,
|
|
"num_tokens": 5890806.0,
|
|
"step": 220
|
|
},
|
|
{
|
|
"entropy": 0.8282079607248306,
|
|
"epoch": 0.4094803605207522,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 1.5266666666666667e-05,
|
|
"loss": 0.9353,
|
|
"mean_token_accuracy": 0.7817027729004622,
|
|
"num_tokens": 6158751.0,
|
|
"step": 230
|
|
},
|
|
{
|
|
"entropy": 0.8447791969403624,
|
|
"epoch": 0.4272838544564371,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 1.5933333333333336e-05,
|
|
"loss": 0.9428,
|
|
"mean_token_accuracy": 0.7795210376381874,
|
|
"num_tokens": 6423747.0,
|
|
"step": 240
|
|
},
|
|
{
|
|
"entropy": 0.8478817906230688,
|
|
"epoch": 0.44508734839212194,
|
|
"grad_norm": 1.8984375,
|
|
"learning_rate": 1.66e-05,
|
|
"loss": 0.9542,
|
|
"mean_token_accuracy": 0.7772787544876337,
|
|
"num_tokens": 6684517.0,
|
|
"step": 250
|
|
},
|
|
{
|
|
"entropy": 0.8135418241843582,
|
|
"epoch": 0.4628908423278068,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 1.726666666666667e-05,
|
|
"loss": 0.9297,
|
|
"mean_token_accuracy": 0.783412617444992,
|
|
"num_tokens": 6945659.0,
|
|
"step": 260
|
|
},
|
|
{
|
|
"entropy": 0.8251218661665917,
|
|
"epoch": 0.48069433626349173,
|
|
"grad_norm": 1.6640625,
|
|
"learning_rate": 1.7933333333333333e-05,
|
|
"loss": 0.9307,
|
|
"mean_token_accuracy": 0.7831020273268223,
|
|
"num_tokens": 7217797.0,
|
|
"step": 270
|
|
},
|
|
{
|
|
"entropy": 0.7879339458420873,
|
|
"epoch": 0.4984978301991766,
|
|
"grad_norm": 1.328125,
|
|
"learning_rate": 1.86e-05,
|
|
"loss": 0.8795,
|
|
"mean_token_accuracy": 0.7897050529718399,
|
|
"num_tokens": 7498182.0,
|
|
"step": 280
|
|
},
|
|
{
|
|
"entropy": 0.8145477209240198,
|
|
"epoch": 0.5163013241348615,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 1.926666666666667e-05,
|
|
"loss": 0.9187,
|
|
"mean_token_accuracy": 0.7838029216974973,
|
|
"num_tokens": 7761686.0,
|
|
"step": 290
|
|
},
|
|
{
|
|
"entropy": 0.8231891892850399,
|
|
"epoch": 0.5341048180705463,
|
|
"grad_norm": 1.5625,
|
|
"learning_rate": 1.9933333333333334e-05,
|
|
"loss": 0.9271,
|
|
"mean_token_accuracy": 0.7819279827177524,
|
|
"num_tokens": 8026923.0,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.5341048180705463,
|
|
"eval_biology_entropy": 1.8821040544509888,
|
|
"eval_biology_loss": 2.226550340652466,
|
|
"eval_biology_mean_token_accuracy": 0.5873979187011719,
|
|
"eval_biology_num_tokens": 8026923.0,
|
|
"eval_biology_runtime": 39.1096,
|
|
"eval_biology_samples_per_second": 12.785,
|
|
"eval_biology_steps_per_second": 3.196,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.5341048180705463,
|
|
"eval_chemistry_entropy": 0.7922393922805786,
|
|
"eval_chemistry_loss": 0.9438055157661438,
|
|
"eval_chemistry_mean_token_accuracy": 0.7841580080986023,
|
|
"eval_chemistry_num_tokens": 8026923.0,
|
|
"eval_chemistry_runtime": 48.7268,
|
|
"eval_chemistry_samples_per_second": 10.261,
|
|
"eval_chemistry_steps_per_second": 2.565,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.5341048180705463,
|
|
"eval_math_entropy": 0.7649447858333588,
|
|
"eval_math_loss": 1.1145466566085815,
|
|
"eval_math_mean_token_accuracy": 0.763204463005066,
|
|
"eval_math_num_tokens": 8026923.0,
|
|
"eval_math_runtime": 50.1205,
|
|
"eval_math_samples_per_second": 9.976,
|
|
"eval_math_steps_per_second": 2.494,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.5341048180705463,
|
|
"eval_physics_entropy": 0.7964776389598847,
|
|
"eval_physics_loss": 0.9373180866241455,
|
|
"eval_physics_mean_token_accuracy": 0.784841582775116,
|
|
"eval_physics_num_tokens": 8026923.0,
|
|
"eval_physics_runtime": 57.6222,
|
|
"eval_physics_samples_per_second": 8.677,
|
|
"eval_physics_steps_per_second": 2.169,
|
|
"step": 300
|
|
},
|
|
{
|
|
"entropy": 0.800599187426269,
|
|
"epoch": 0.5519083120062312,
|
|
"grad_norm": 1.6640625,
|
|
"learning_rate": 1.9933333333333334e-05,
|
|
"loss": 0.9071,
|
|
"mean_token_accuracy": 0.7856069568544626,
|
|
"num_tokens": 8301506.0,
|
|
"step": 310
|
|
},
|
|
{
|
|
"entropy": 0.7965721761807799,
|
|
"epoch": 0.5697118059419161,
|
|
"grad_norm": 1.234375,
|
|
"learning_rate": 1.985925925925926e-05,
|
|
"loss": 0.9139,
|
|
"mean_token_accuracy": 0.786777526140213,
|
|
"num_tokens": 8559126.0,
|
|
"step": 320
|
|
},
|
|
{
|
|
"entropy": 0.7925007566809654,
|
|
"epoch": 0.587515299877601,
|
|
"grad_norm": 1.4453125,
|
|
"learning_rate": 1.9785185185185187e-05,
|
|
"loss": 0.906,
|
|
"mean_token_accuracy": 0.7864235289394855,
|
|
"num_tokens": 8821088.0,
|
|
"step": 330
|
|
},
|
|
{
|
|
"entropy": 0.7728251311928034,
|
|
"epoch": 0.6053187938132859,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 1.971111111111111e-05,
|
|
"loss": 0.8764,
|
|
"mean_token_accuracy": 0.7919386122375727,
|
|
"num_tokens": 9089866.0,
|
|
"step": 340
|
|
},
|
|
{
|
|
"entropy": 0.7705478269606829,
|
|
"epoch": 0.6231222877489707,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 1.963703703703704e-05,
|
|
"loss": 0.8837,
|
|
"mean_token_accuracy": 0.7896027479320764,
|
|
"num_tokens": 9358866.0,
|
|
"step": 350
|
|
},
|
|
{
|
|
"entropy": 0.7656249281018972,
|
|
"epoch": 0.6409257816846556,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 1.9562962962962964e-05,
|
|
"loss": 0.8626,
|
|
"mean_token_accuracy": 0.7931958142668009,
|
|
"num_tokens": 9629977.0,
|
|
"step": 360
|
|
},
|
|
{
|
|
"entropy": 0.7598591955378652,
|
|
"epoch": 0.6587292756203404,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.948888888888889e-05,
|
|
"loss": 0.8608,
|
|
"mean_token_accuracy": 0.7945109561085701,
|
|
"num_tokens": 9897757.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"entropy": 0.7398761732503771,
|
|
"epoch": 0.6765327695560254,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 1.9414814814814817e-05,
|
|
"loss": 0.848,
|
|
"mean_token_accuracy": 0.7985429972410202,
|
|
"num_tokens": 10168367.0,
|
|
"step": 380
|
|
},
|
|
{
|
|
"entropy": 0.7331193052232265,
|
|
"epoch": 0.6943362634917103,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 1.9340740740740743e-05,
|
|
"loss": 0.8366,
|
|
"mean_token_accuracy": 0.8001827634871006,
|
|
"num_tokens": 10439039.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"entropy": 0.7496613372117281,
|
|
"epoch": 0.7121397574273951,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 1.926666666666667e-05,
|
|
"loss": 0.8517,
|
|
"mean_token_accuracy": 0.7967161379754544,
|
|
"num_tokens": 10709549.0,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.7121397574273951,
|
|
"eval_biology_entropy": 1.823248610496521,
|
|
"eval_biology_loss": 2.159681558609009,
|
|
"eval_biology_mean_token_accuracy": 0.5972214796543122,
|
|
"eval_biology_num_tokens": 10709549.0,
|
|
"eval_biology_runtime": 39.2024,
|
|
"eval_biology_samples_per_second": 12.754,
|
|
"eval_biology_steps_per_second": 3.189,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.7121397574273951,
|
|
"eval_chemistry_entropy": 0.7672218294143677,
|
|
"eval_chemistry_loss": 0.8971230387687683,
|
|
"eval_chemistry_mean_token_accuracy": 0.7928389143943787,
|
|
"eval_chemistry_num_tokens": 10709549.0,
|
|
"eval_chemistry_runtime": 48.8296,
|
|
"eval_chemistry_samples_per_second": 10.24,
|
|
"eval_chemistry_steps_per_second": 2.56,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.7121397574273951,
|
|
"eval_math_entropy": 0.7494234163761139,
|
|
"eval_math_loss": 1.0849554538726807,
|
|
"eval_math_mean_token_accuracy": 0.7666959943771362,
|
|
"eval_math_num_tokens": 10709549.0,
|
|
"eval_math_runtime": 50.2325,
|
|
"eval_math_samples_per_second": 9.954,
|
|
"eval_math_steps_per_second": 2.488,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.7121397574273951,
|
|
"eval_physics_entropy": 0.7731240901947022,
|
|
"eval_physics_loss": 0.9018430709838867,
|
|
"eval_physics_mean_token_accuracy": 0.7909709796905517,
|
|
"eval_physics_num_tokens": 10709549.0,
|
|
"eval_physics_runtime": 57.7809,
|
|
"eval_physics_samples_per_second": 8.653,
|
|
"eval_physics_steps_per_second": 2.163,
|
|
"step": 400
|
|
},
|
|
{
|
|
"entropy": 0.7354462265968322,
|
|
"epoch": 0.72994325136308,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.9192592592592593e-05,
|
|
"loss": 0.8411,
|
|
"mean_token_accuracy": 0.8003354988992214,
|
|
"num_tokens": 10976011.0,
|
|
"step": 410
|
|
},
|
|
{
|
|
"entropy": 0.7435122076421976,
|
|
"epoch": 0.7477467452987648,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 1.911851851851852e-05,
|
|
"loss": 0.8438,
|
|
"mean_token_accuracy": 0.7971596572548151,
|
|
"num_tokens": 11245251.0,
|
|
"step": 420
|
|
},
|
|
{
|
|
"entropy": 0.759309609234333,
|
|
"epoch": 0.7655502392344498,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 1.9044444444444446e-05,
|
|
"loss": 0.8579,
|
|
"mean_token_accuracy": 0.792990094423294,
|
|
"num_tokens": 11518514.0,
|
|
"step": 430
|
|
},
|
|
{
|
|
"entropy": 0.7545185785740613,
|
|
"epoch": 0.7833537331701347,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.8970370370370372e-05,
|
|
"loss": 0.8574,
|
|
"mean_token_accuracy": 0.794682926684618,
|
|
"num_tokens": 11787117.0,
|
|
"step": 440
|
|
},
|
|
{
|
|
"entropy": 0.751254404336214,
|
|
"epoch": 0.8011572271058195,
|
|
"grad_norm": 1.515625,
|
|
"learning_rate": 1.8896296296296295e-05,
|
|
"loss": 0.8495,
|
|
"mean_token_accuracy": 0.7972703959792853,
|
|
"num_tokens": 12050993.0,
|
|
"step": 450
|
|
},
|
|
{
|
|
"entropy": 0.7543802969157696,
|
|
"epoch": 0.8189607210415044,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 1.8822222222222225e-05,
|
|
"loss": 0.8527,
|
|
"mean_token_accuracy": 0.7960003320127725,
|
|
"num_tokens": 12311688.0,
|
|
"step": 460
|
|
},
|
|
{
|
|
"entropy": 0.7186065271496773,
|
|
"epoch": 0.8367642149771892,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 1.874814814814815e-05,
|
|
"loss": 0.8096,
|
|
"mean_token_accuracy": 0.8027401898056269,
|
|
"num_tokens": 12598408.0,
|
|
"step": 470
|
|
},
|
|
{
|
|
"entropy": 0.7257912319153548,
|
|
"epoch": 0.8545677089128741,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 1.8674074074074075e-05,
|
|
"loss": 0.8299,
|
|
"mean_token_accuracy": 0.8012042317539454,
|
|
"num_tokens": 12869642.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"entropy": 0.7173270899802446,
|
|
"epoch": 0.8723712028485591,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 1.86e-05,
|
|
"loss": 0.8146,
|
|
"mean_token_accuracy": 0.8032398864626884,
|
|
"num_tokens": 13139555.0,
|
|
"step": 490
|
|
},
|
|
{
|
|
"entropy": 0.7487573977559805,
|
|
"epoch": 0.8901746967842439,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 1.8525925925925928e-05,
|
|
"loss": 0.8562,
|
|
"mean_token_accuracy": 0.7955521304160357,
|
|
"num_tokens": 13396545.0,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.8901746967842439,
|
|
"eval_biology_entropy": 1.8204833917617798,
|
|
"eval_biology_loss": 2.1014022827148438,
|
|
"eval_biology_mean_token_accuracy": 0.6051526765823364,
|
|
"eval_biology_num_tokens": 13396545.0,
|
|
"eval_biology_runtime": 39.1543,
|
|
"eval_biology_samples_per_second": 12.77,
|
|
"eval_biology_steps_per_second": 3.192,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.8901746967842439,
|
|
"eval_chemistry_entropy": 0.7538458592891694,
|
|
"eval_chemistry_loss": 0.8699584007263184,
|
|
"eval_chemistry_mean_token_accuracy": 0.7975720787048339,
|
|
"eval_chemistry_num_tokens": 13396545.0,
|
|
"eval_chemistry_runtime": 48.8229,
|
|
"eval_chemistry_samples_per_second": 10.241,
|
|
"eval_chemistry_steps_per_second": 2.56,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.8901746967842439,
|
|
"eval_math_entropy": 0.7398368701934814,
|
|
"eval_math_loss": 1.0758970975875854,
|
|
"eval_math_mean_token_accuracy": 0.7687290472984314,
|
|
"eval_math_num_tokens": 13396545.0,
|
|
"eval_math_runtime": 50.0916,
|
|
"eval_math_samples_per_second": 9.982,
|
|
"eval_math_steps_per_second": 2.495,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.8901746967842439,
|
|
"eval_physics_entropy": 0.7607374501228332,
|
|
"eval_physics_loss": 0.8848790526390076,
|
|
"eval_physics_mean_token_accuracy": 0.7937801394462586,
|
|
"eval_physics_num_tokens": 13396545.0,
|
|
"eval_physics_runtime": 57.6599,
|
|
"eval_physics_samples_per_second": 8.672,
|
|
"eval_physics_steps_per_second": 2.168,
|
|
"step": 500
|
|
},
|
|
{
|
|
"entropy": 0.7129743231460453,
|
|
"epoch": 0.9079781907199288,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 1.8451851851851855e-05,
|
|
"loss": 0.8109,
|
|
"mean_token_accuracy": 0.8060642663389445,
|
|
"num_tokens": 13667857.0,
|
|
"step": 510
|
|
},
|
|
{
|
|
"entropy": 0.7393450669944286,
|
|
"epoch": 0.9257816846556136,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 1.8377777777777778e-05,
|
|
"loss": 0.8325,
|
|
"mean_token_accuracy": 0.7997861213982105,
|
|
"num_tokens": 13932771.0,
|
|
"step": 520
|
|
},
|
|
{
|
|
"entropy": 0.7191115407273173,
|
|
"epoch": 0.9435851785912985,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 1.8303703703703704e-05,
|
|
"loss": 0.8233,
|
|
"mean_token_accuracy": 0.8034567594528198,
|
|
"num_tokens": 14194632.0,
|
|
"step": 530
|
|
},
|
|
{
|
|
"entropy": 0.7434917811304331,
|
|
"epoch": 0.9613886725269835,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 1.822962962962963e-05,
|
|
"loss": 0.8435,
|
|
"mean_token_accuracy": 0.7969928599894047,
|
|
"num_tokens": 14457650.0,
|
|
"step": 540
|
|
},
|
|
{
|
|
"entropy": 0.7100320702418685,
|
|
"epoch": 0.9791921664626683,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 1.8155555555555557e-05,
|
|
"loss": 0.8149,
|
|
"mean_token_accuracy": 0.803739969804883,
|
|
"num_tokens": 14724967.0,
|
|
"step": 550
|
|
},
|
|
{
|
|
"entropy": 0.737933493219316,
|
|
"epoch": 0.9969956603983532,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 1.8081481481481484e-05,
|
|
"loss": 0.8421,
|
|
"mean_token_accuracy": 0.8001800544559956,
|
|
"num_tokens": 14993272.0,
|
|
"step": 560
|
|
},
|
|
{
|
|
"entropy": 0.6655864379098339,
|
|
"epoch": 1.014242795148548,
|
|
"grad_norm": 1.1171875,
|
|
"learning_rate": 1.800740740740741e-05,
|
|
"loss": 0.7668,
|
|
"mean_token_accuracy": 0.814733551779101,
|
|
"num_tokens": 15254936.0,
|
|
"step": 570
|
|
},
|
|
{
|
|
"entropy": 0.6937626471742988,
|
|
"epoch": 1.0320462890842328,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 1.7933333333333333e-05,
|
|
"loss": 0.7942,
|
|
"mean_token_accuracy": 0.806296594068408,
|
|
"num_tokens": 15515007.0,
|
|
"step": 580
|
|
},
|
|
{
|
|
"entropy": 0.6596947606652975,
|
|
"epoch": 1.0498497830199176,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 1.785925925925926e-05,
|
|
"loss": 0.7623,
|
|
"mean_token_accuracy": 0.8152227636426688,
|
|
"num_tokens": 15794203.0,
|
|
"step": 590
|
|
},
|
|
{
|
|
"entropy": 0.6868148937821388,
|
|
"epoch": 1.0676532769556026,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 1.7785185185185186e-05,
|
|
"loss": 0.7827,
|
|
"mean_token_accuracy": 0.8101296100765467,
|
|
"num_tokens": 16063664.0,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.0676532769556026,
|
|
"eval_biology_entropy": 1.6867316131591796,
|
|
"eval_biology_loss": 2.077732801437378,
|
|
"eval_biology_mean_token_accuracy": 0.6078621828556061,
|
|
"eval_biology_num_tokens": 16063664.0,
|
|
"eval_biology_runtime": 39.1079,
|
|
"eval_biology_samples_per_second": 12.785,
|
|
"eval_biology_steps_per_second": 3.196,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.0676532769556026,
|
|
"eval_chemistry_entropy": 0.6882545161247253,
|
|
"eval_chemistry_loss": 0.8549924492835999,
|
|
"eval_chemistry_mean_token_accuracy": 0.8009334373474121,
|
|
"eval_chemistry_num_tokens": 16063664.0,
|
|
"eval_chemistry_runtime": 48.732,
|
|
"eval_chemistry_samples_per_second": 10.26,
|
|
"eval_chemistry_steps_per_second": 2.565,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.0676532769556026,
|
|
"eval_math_entropy": 0.6883157949447631,
|
|
"eval_math_loss": 1.075770616531372,
|
|
"eval_math_mean_token_accuracy": 0.7705092186927796,
|
|
"eval_math_num_tokens": 16063664.0,
|
|
"eval_math_runtime": 50.1345,
|
|
"eval_math_samples_per_second": 9.973,
|
|
"eval_math_steps_per_second": 2.493,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 1.0676532769556026,
|
|
"eval_physics_entropy": 0.7055722939968109,
|
|
"eval_physics_loss": 0.8771266937255859,
|
|
"eval_physics_mean_token_accuracy": 0.7955313258171082,
|
|
"eval_physics_num_tokens": 16063664.0,
|
|
"eval_physics_runtime": 57.6658,
|
|
"eval_physics_samples_per_second": 8.671,
|
|
"eval_physics_steps_per_second": 2.168,
|
|
"step": 600
|
|
},
|
|
{
|
|
"entropy": 0.6597639428451657,
|
|
"epoch": 1.0854567708912874,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 1.7711111111111113e-05,
|
|
"loss": 0.7655,
|
|
"mean_token_accuracy": 0.8151189126074314,
|
|
"num_tokens": 16338303.0,
|
|
"step": 610
|
|
},
|
|
{
|
|
"entropy": 0.6991181125864386,
|
|
"epoch": 1.1032602648269723,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.763703703703704e-05,
|
|
"loss": 0.7934,
|
|
"mean_token_accuracy": 0.8071035247296094,
|
|
"num_tokens": 16606302.0,
|
|
"step": 620
|
|
},
|
|
{
|
|
"entropy": 0.6774254593998194,
|
|
"epoch": 1.121063758762657,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 1.7562962962962962e-05,
|
|
"loss": 0.7706,
|
|
"mean_token_accuracy": 0.8098537240177393,
|
|
"num_tokens": 16877071.0,
|
|
"step": 630
|
|
},
|
|
{
|
|
"entropy": 0.6678685389459134,
|
|
"epoch": 1.138867252698342,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 1.7488888888888892e-05,
|
|
"loss": 0.7714,
|
|
"mean_token_accuracy": 0.8127609983086586,
|
|
"num_tokens": 17145775.0,
|
|
"step": 640
|
|
},
|
|
{
|
|
"entropy": 0.6851805068552494,
|
|
"epoch": 1.156670746634027,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 1.7414814814814815e-05,
|
|
"loss": 0.777,
|
|
"mean_token_accuracy": 0.8110932789742946,
|
|
"num_tokens": 17410775.0,
|
|
"step": 650
|
|
},
|
|
{
|
|
"entropy": 0.7042925998568534,
|
|
"epoch": 1.1744742405697117,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 1.7340740740740742e-05,
|
|
"loss": 0.7979,
|
|
"mean_token_accuracy": 0.8038393501192331,
|
|
"num_tokens": 17677791.0,
|
|
"step": 660
|
|
},
|
|
{
|
|
"entropy": 0.6810087833553553,
|
|
"epoch": 1.1922777345053968,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 1.726666666666667e-05,
|
|
"loss": 0.7688,
|
|
"mean_token_accuracy": 0.8104576140642166,
|
|
"num_tokens": 17949441.0,
|
|
"step": 670
|
|
},
|
|
{
|
|
"entropy": 0.662551742233336,
|
|
"epoch": 1.2100812284410816,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 1.7192592592592595e-05,
|
|
"loss": 0.7653,
|
|
"mean_token_accuracy": 0.8126230895519256,
|
|
"num_tokens": 18217742.0,
|
|
"step": 680
|
|
},
|
|
{
|
|
"entropy": 0.6803617887198925,
|
|
"epoch": 1.2278847223767664,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 1.711851851851852e-05,
|
|
"loss": 0.7771,
|
|
"mean_token_accuracy": 0.8095329128205776,
|
|
"num_tokens": 18487769.0,
|
|
"step": 690
|
|
},
|
|
{
|
|
"entropy": 0.6707866100594402,
|
|
"epoch": 1.2456882163124514,
|
|
"grad_norm": 1.34375,
|
|
"learning_rate": 1.7044444444444445e-05,
|
|
"loss": 0.7687,
|
|
"mean_token_accuracy": 0.8113945126533508,
|
|
"num_tokens": 18760296.0,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 1.2456882163124514,
|
|
"eval_biology_entropy": 1.6589141931533813,
|
|
"eval_biology_loss": 2.081825017929077,
|
|
"eval_biology_mean_token_accuracy": 0.608892562866211,
|
|
"eval_biology_num_tokens": 18760296.0,
|
|
"eval_biology_runtime": 39.8083,
|
|
"eval_biology_samples_per_second": 12.56,
|
|
"eval_biology_steps_per_second": 3.14,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 1.2456882163124514,
|
|
"eval_chemistry_entropy": 0.6833467047214508,
|
|
"eval_chemistry_loss": 0.8464663028717041,
|
|
"eval_chemistry_mean_token_accuracy": 0.80278346824646,
|
|
"eval_chemistry_num_tokens": 18760296.0,
|
|
"eval_chemistry_runtime": 49.1467,
|
|
"eval_chemistry_samples_per_second": 10.174,
|
|
"eval_chemistry_steps_per_second": 2.543,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 1.2456882163124514,
|
|
"eval_math_entropy": 0.6947779231071473,
|
|
"eval_math_loss": 1.0786137580871582,
|
|
"eval_math_mean_token_accuracy": 0.7709644784927369,
|
|
"eval_math_num_tokens": 18760296.0,
|
|
"eval_math_runtime": 52.5755,
|
|
"eval_math_samples_per_second": 9.51,
|
|
"eval_math_steps_per_second": 2.378,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 1.2456882163124514,
|
|
"eval_physics_entropy": 0.704109364271164,
|
|
"eval_physics_loss": 0.8724388480186462,
|
|
"eval_physics_mean_token_accuracy": 0.796642897605896,
|
|
"eval_physics_num_tokens": 18760296.0,
|
|
"eval_physics_runtime": 61.3658,
|
|
"eval_physics_samples_per_second": 8.148,
|
|
"eval_physics_steps_per_second": 2.037,
|
|
"step": 700
|
|
},
|
|
{
|
|
"entropy": 0.6705936685204505,
|
|
"epoch": 1.2634917102481362,
|
|
"grad_norm": 1.140625,
|
|
"learning_rate": 1.697037037037037e-05,
|
|
"loss": 0.7618,
|
|
"mean_token_accuracy": 0.8144830856472254,
|
|
"num_tokens": 19035733.0,
|
|
"step": 710
|
|
},
|
|
{
|
|
"entropy": 0.6870324395596981,
|
|
"epoch": 1.281295204183821,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 1.6896296296296298e-05,
|
|
"loss": 0.7849,
|
|
"mean_token_accuracy": 0.807737946882844,
|
|
"num_tokens": 19311432.0,
|
|
"step": 720
|
|
},
|
|
{
|
|
"entropy": 0.6551721351221204,
|
|
"epoch": 1.299098698119506,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 1.6822222222222224e-05,
|
|
"loss": 0.7538,
|
|
"mean_token_accuracy": 0.8164622981101275,
|
|
"num_tokens": 19593016.0,
|
|
"step": 730
|
|
},
|
|
{
|
|
"entropy": 0.6645344085991383,
|
|
"epoch": 1.3169021920551909,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 1.6748148148148147e-05,
|
|
"loss": 0.759,
|
|
"mean_token_accuracy": 0.8134492743760348,
|
|
"num_tokens": 19860476.0,
|
|
"step": 740
|
|
},
|
|
{
|
|
"entropy": 0.6684485200792551,
|
|
"epoch": 1.3347056859908757,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 1.6674074074074077e-05,
|
|
"loss": 0.7632,
|
|
"mean_token_accuracy": 0.8128790140151978,
|
|
"num_tokens": 20124436.0,
|
|
"step": 750
|
|
},
|
|
{
|
|
"entropy": 0.6736280497163534,
|
|
"epoch": 1.3525091799265607,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 1.66e-05,
|
|
"loss": 0.7771,
|
|
"mean_token_accuracy": 0.813029420748353,
|
|
"num_tokens": 20388027.0,
|
|
"step": 760
|
|
},
|
|
{
|
|
"entropy": 0.6706711670383811,
|
|
"epoch": 1.3703126738622455,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 1.6525925925925927e-05,
|
|
"loss": 0.7697,
|
|
"mean_token_accuracy": 0.8115393359214067,
|
|
"num_tokens": 20656076.0,
|
|
"step": 770
|
|
},
|
|
{
|
|
"entropy": 0.6996258163824678,
|
|
"epoch": 1.3881161677979303,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 1.6451851851851853e-05,
|
|
"loss": 0.8077,
|
|
"mean_token_accuracy": 0.8058754079043865,
|
|
"num_tokens": 20911633.0,
|
|
"step": 780
|
|
},
|
|
{
|
|
"entropy": 0.6905225297436118,
|
|
"epoch": 1.4059196617336152,
|
|
"grad_norm": 1.171875,
|
|
"learning_rate": 1.637777777777778e-05,
|
|
"loss": 0.7955,
|
|
"mean_token_accuracy": 0.8070559553802014,
|
|
"num_tokens": 21176227.0,
|
|
"step": 790
|
|
},
|
|
{
|
|
"entropy": 0.6718146029859782,
|
|
"epoch": 1.4237231556693,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 1.6303703703703706e-05,
|
|
"loss": 0.7716,
|
|
"mean_token_accuracy": 0.8126390781253576,
|
|
"num_tokens": 21445242.0,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 1.4237231556693,
|
|
"eval_biology_entropy": 1.6255396161079407,
|
|
"eval_biology_loss": 2.0511693954467773,
|
|
"eval_biology_mean_token_accuracy": 0.6119143316745758,
|
|
"eval_biology_num_tokens": 21445242.0,
|
|
"eval_biology_runtime": 42.4445,
|
|
"eval_biology_samples_per_second": 11.78,
|
|
"eval_biology_steps_per_second": 2.945,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 1.4237231556693,
|
|
"eval_chemistry_entropy": 0.6763756844997406,
|
|
"eval_chemistry_loss": 0.8368268609046936,
|
|
"eval_chemistry_mean_token_accuracy": 0.8050474629402161,
|
|
"eval_chemistry_num_tokens": 21445242.0,
|
|
"eval_chemistry_runtime": 51.5104,
|
|
"eval_chemistry_samples_per_second": 9.707,
|
|
"eval_chemistry_steps_per_second": 2.427,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 1.4237231556693,
|
|
"eval_math_entropy": 0.6853804645538331,
|
|
"eval_math_loss": 1.0698176622390747,
|
|
"eval_math_mean_token_accuracy": 0.7714386968612671,
|
|
"eval_math_num_tokens": 21445242.0,
|
|
"eval_math_runtime": 54.9388,
|
|
"eval_math_samples_per_second": 9.101,
|
|
"eval_math_steps_per_second": 2.275,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 1.4237231556693,
|
|
"eval_physics_entropy": 0.6991692934036254,
|
|
"eval_physics_loss": 0.8673914074897766,
|
|
"eval_physics_mean_token_accuracy": 0.7979785161018371,
|
|
"eval_physics_num_tokens": 21445242.0,
|
|
"eval_physics_runtime": 60.9527,
|
|
"eval_physics_samples_per_second": 8.203,
|
|
"eval_physics_steps_per_second": 2.051,
|
|
"step": 800
|
|
},
|
|
{
|
|
"entropy": 0.6695453573018313,
|
|
"epoch": 1.441526649604985,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.622962962962963e-05,
|
|
"loss": 0.77,
|
|
"mean_token_accuracy": 0.8109262142330408,
|
|
"num_tokens": 21718455.0,
|
|
"step": 810
|
|
},
|
|
{
|
|
"entropy": 0.6753543723374605,
|
|
"epoch": 1.4593301435406698,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.6155555555555556e-05,
|
|
"loss": 0.7782,
|
|
"mean_token_accuracy": 0.8130301378667355,
|
|
"num_tokens": 21974076.0,
|
|
"step": 820
|
|
},
|
|
{
|
|
"entropy": 0.6758085643872619,
|
|
"epoch": 1.4771336374763546,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 1.6081481481481482e-05,
|
|
"loss": 0.7768,
|
|
"mean_token_accuracy": 0.8110865034162998,
|
|
"num_tokens": 22241599.0,
|
|
"step": 830
|
|
},
|
|
{
|
|
"entropy": 0.6868746858090162,
|
|
"epoch": 1.4949371314120397,
|
|
"grad_norm": 1.34375,
|
|
"learning_rate": 1.600740740740741e-05,
|
|
"loss": 0.7818,
|
|
"mean_token_accuracy": 0.8090599600225687,
|
|
"num_tokens": 22510725.0,
|
|
"step": 840
|
|
},
|
|
{
|
|
"entropy": 0.6554480012506246,
|
|
"epoch": 1.5127406253477245,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 1.5933333333333336e-05,
|
|
"loss": 0.766,
|
|
"mean_token_accuracy": 0.8149232871830463,
|
|
"num_tokens": 22778514.0,
|
|
"step": 850
|
|
},
|
|
{
|
|
"entropy": 0.6574800366535782,
|
|
"epoch": 1.5305441192834093,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.5859259259259262e-05,
|
|
"loss": 0.7517,
|
|
"mean_token_accuracy": 0.8147028144448996,
|
|
"num_tokens": 23051230.0,
|
|
"step": 860
|
|
},
|
|
{
|
|
"entropy": 0.6707898294553161,
|
|
"epoch": 1.5483476132190943,
|
|
"grad_norm": 1.21875,
|
|
"learning_rate": 1.5785185185185185e-05,
|
|
"loss": 0.776,
|
|
"mean_token_accuracy": 0.8121768556535244,
|
|
"num_tokens": 23306327.0,
|
|
"step": 870
|
|
},
|
|
{
|
|
"entropy": 0.6759747974574566,
|
|
"epoch": 1.5661511071547791,
|
|
"grad_norm": 1.234375,
|
|
"learning_rate": 1.571111111111111e-05,
|
|
"loss": 0.7717,
|
|
"mean_token_accuracy": 0.810706127434969,
|
|
"num_tokens": 23573127.0,
|
|
"step": 880
|
|
},
|
|
{
|
|
"entropy": 0.6696014609187841,
|
|
"epoch": 1.583954601090464,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 1.5637037037037038e-05,
|
|
"loss": 0.7635,
|
|
"mean_token_accuracy": 0.8123903065919876,
|
|
"num_tokens": 23841914.0,
|
|
"step": 890
|
|
},
|
|
{
|
|
"entropy": 0.6680066948756576,
|
|
"epoch": 1.601758095026149,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 1.5562962962962965e-05,
|
|
"loss": 0.7763,
|
|
"mean_token_accuracy": 0.8117439385503531,
|
|
"num_tokens": 24105114.0,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.601758095026149,
|
|
"eval_biology_entropy": 1.6559728245735168,
|
|
"eval_biology_loss": 2.0387370586395264,
|
|
"eval_biology_mean_token_accuracy": 0.6136598820686341,
|
|
"eval_biology_num_tokens": 24105114.0,
|
|
"eval_biology_runtime": 41.4479,
|
|
"eval_biology_samples_per_second": 12.063,
|
|
"eval_biology_steps_per_second": 3.016,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.601758095026149,
|
|
"eval_chemistry_entropy": 0.683923549413681,
|
|
"eval_chemistry_loss": 0.8298754692077637,
|
|
"eval_chemistry_mean_token_accuracy": 0.8061454033851624,
|
|
"eval_chemistry_num_tokens": 24105114.0,
|
|
"eval_chemistry_runtime": 50.8984,
|
|
"eval_chemistry_samples_per_second": 9.823,
|
|
"eval_chemistry_steps_per_second": 2.456,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.601758095026149,
|
|
"eval_math_entropy": 0.6932375190258027,
|
|
"eval_math_loss": 1.06475830078125,
|
|
"eval_math_mean_token_accuracy": 0.7720690259933471,
|
|
"eval_math_num_tokens": 24105114.0,
|
|
"eval_math_runtime": 52.5544,
|
|
"eval_math_samples_per_second": 9.514,
|
|
"eval_math_steps_per_second": 2.378,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 1.601758095026149,
|
|
"eval_physics_entropy": 0.7051843535900116,
|
|
"eval_physics_loss": 0.8616264462471008,
|
|
"eval_physics_mean_token_accuracy": 0.7987073979377747,
|
|
"eval_physics_num_tokens": 24105114.0,
|
|
"eval_physics_runtime": 59.8508,
|
|
"eval_physics_samples_per_second": 8.354,
|
|
"eval_physics_steps_per_second": 2.089,
|
|
"step": 900
|
|
},
|
|
{
|
|
"entropy": 0.6430077504366636,
|
|
"epoch": 1.6195615889618338,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 1.548888888888889e-05,
|
|
"loss": 0.743,
|
|
"mean_token_accuracy": 0.8184590026736259,
|
|
"num_tokens": 24384008.0,
|
|
"step": 910
|
|
},
|
|
{
|
|
"entropy": 0.6775279760360717,
|
|
"epoch": 1.6373650828975186,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.5414814814814814e-05,
|
|
"loss": 0.7704,
|
|
"mean_token_accuracy": 0.8114439699798822,
|
|
"num_tokens": 24638205.0,
|
|
"step": 920
|
|
},
|
|
{
|
|
"entropy": 0.6244934551417828,
|
|
"epoch": 1.6551685768332036,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 1.5340740740740744e-05,
|
|
"loss": 0.7253,
|
|
"mean_token_accuracy": 0.8222480688244105,
|
|
"num_tokens": 24910104.0,
|
|
"step": 930
|
|
},
|
|
{
|
|
"entropy": 0.6539383618161082,
|
|
"epoch": 1.6729720707688884,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 1.5266666666666667e-05,
|
|
"loss": 0.7471,
|
|
"mean_token_accuracy": 0.8160006206482648,
|
|
"num_tokens": 25181306.0,
|
|
"step": 940
|
|
},
|
|
{
|
|
"entropy": 0.6607510834932327,
|
|
"epoch": 1.6907755647045732,
|
|
"grad_norm": 1.3203125,
|
|
"learning_rate": 1.5192592592592594e-05,
|
|
"loss": 0.7589,
|
|
"mean_token_accuracy": 0.8142590031027794,
|
|
"num_tokens": 25449138.0,
|
|
"step": 950
|
|
},
|
|
{
|
|
"entropy": 0.6612858925014734,
|
|
"epoch": 1.7085790586402583,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 1.5118518518518519e-05,
|
|
"loss": 0.766,
|
|
"mean_token_accuracy": 0.8137270450592041,
|
|
"num_tokens": 25709030.0,
|
|
"step": 960
|
|
},
|
|
{
|
|
"entropy": 0.6701876953244209,
|
|
"epoch": 1.7263825525759429,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.5044444444444445e-05,
|
|
"loss": 0.7702,
|
|
"mean_token_accuracy": 0.8120660580694675,
|
|
"num_tokens": 25972807.0,
|
|
"step": 970
|
|
},
|
|
{
|
|
"entropy": 0.6566708590835333,
|
|
"epoch": 1.744186046511628,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 1.497037037037037e-05,
|
|
"loss": 0.748,
|
|
"mean_token_accuracy": 0.8170768674463034,
|
|
"num_tokens": 26244955.0,
|
|
"step": 980
|
|
},
|
|
{
|
|
"entropy": 0.624184824153781,
|
|
"epoch": 1.761989540447313,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 1.4896296296296298e-05,
|
|
"loss": 0.7284,
|
|
"mean_token_accuracy": 0.8223610106855631,
|
|
"num_tokens": 26515548.0,
|
|
"step": 990
|
|
},
|
|
{
|
|
"entropy": 0.6476762825623155,
|
|
"epoch": 1.7797930343829975,
|
|
"grad_norm": 1.21875,
|
|
"learning_rate": 1.4822222222222225e-05,
|
|
"loss": 0.7422,
|
|
"mean_token_accuracy": 0.8172390758991241,
|
|
"num_tokens": 26782376.0,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.7797930343829975,
|
|
"eval_biology_entropy": 1.6157320532798767,
|
|
"eval_biology_loss": 2.0338852405548096,
|
|
"eval_biology_mean_token_accuracy": 0.6143801593780518,
|
|
"eval_biology_num_tokens": 26782376.0,
|
|
"eval_biology_runtime": 41.5849,
|
|
"eval_biology_samples_per_second": 12.024,
|
|
"eval_biology_steps_per_second": 3.006,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.7797930343829975,
|
|
"eval_chemistry_entropy": 0.6670743160247803,
|
|
"eval_chemistry_loss": 0.8244208097457886,
|
|
"eval_chemistry_mean_token_accuracy": 0.8077364330291747,
|
|
"eval_chemistry_num_tokens": 26782376.0,
|
|
"eval_chemistry_runtime": 53.0274,
|
|
"eval_chemistry_samples_per_second": 9.429,
|
|
"eval_chemistry_steps_per_second": 2.357,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.7797930343829975,
|
|
"eval_math_entropy": 0.6760952143669129,
|
|
"eval_math_loss": 1.0642753839492798,
|
|
"eval_math_mean_token_accuracy": 0.7728619589805603,
|
|
"eval_math_num_tokens": 26782376.0,
|
|
"eval_math_runtime": 53.8997,
|
|
"eval_math_samples_per_second": 9.276,
|
|
"eval_math_steps_per_second": 2.319,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 1.7797930343829975,
|
|
"eval_physics_entropy": 0.6914605097770691,
|
|
"eval_physics_loss": 0.8597292304039001,
|
|
"eval_physics_mean_token_accuracy": 0.7992245469093323,
|
|
"eval_physics_num_tokens": 26782376.0,
|
|
"eval_physics_runtime": 62.8355,
|
|
"eval_physics_samples_per_second": 7.957,
|
|
"eval_physics_steps_per_second": 1.989,
|
|
"step": 1000
|
|
}
|
|
],
|
|
"logging_steps": 10,
|
|
"max_steps": 3000,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 6,
|
|
"save_steps": 500,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": false
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 2.0811511522503905e+18,
|
|
"train_batch_size": 4,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|