3635 lines
108 KiB
JSON
3635 lines
108 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 4.448648047179259,
|
|
"eval_steps": 100,
|
|
"global_step": 2500,
|
|
"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
|
|
},
|
|
{
|
|
"entropy": 0.660948914475739,
|
|
"epoch": 1.7975965283186826,
|
|
"grad_norm": 1.359375,
|
|
"learning_rate": 1.474814814814815e-05,
|
|
"loss": 0.7562,
|
|
"mean_token_accuracy": 0.8152279850095511,
|
|
"num_tokens": 27049648.0,
|
|
"step": 1010
|
|
},
|
|
{
|
|
"entropy": 0.6787248050794006,
|
|
"epoch": 1.8154000222543676,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 1.4674074074074076e-05,
|
|
"loss": 0.7771,
|
|
"mean_token_accuracy": 0.8101395227015018,
|
|
"num_tokens": 27315061.0,
|
|
"step": 1020
|
|
},
|
|
{
|
|
"entropy": 0.6424054227769375,
|
|
"epoch": 1.8332035161900522,
|
|
"grad_norm": 1.359375,
|
|
"learning_rate": 1.46e-05,
|
|
"loss": 0.7373,
|
|
"mean_token_accuracy": 0.8189616728574037,
|
|
"num_tokens": 27587650.0,
|
|
"step": 1030
|
|
},
|
|
{
|
|
"entropy": 0.6606515791267157,
|
|
"epoch": 1.8510070101257372,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 1.4525925925925927e-05,
|
|
"loss": 0.7652,
|
|
"mean_token_accuracy": 0.8157406080514192,
|
|
"num_tokens": 27855458.0,
|
|
"step": 1040
|
|
},
|
|
{
|
|
"entropy": 0.6464399367570877,
|
|
"epoch": 1.868810504061422,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 1.4451851851851852e-05,
|
|
"loss": 0.7464,
|
|
"mean_token_accuracy": 0.8191553246229887,
|
|
"num_tokens": 28127703.0,
|
|
"step": 1050
|
|
},
|
|
{
|
|
"entropy": 0.633565224148333,
|
|
"epoch": 1.8866139979971068,
|
|
"grad_norm": 1.140625,
|
|
"learning_rate": 1.4377777777777779e-05,
|
|
"loss": 0.7265,
|
|
"mean_token_accuracy": 0.8215953864157199,
|
|
"num_tokens": 28395128.0,
|
|
"step": 1060
|
|
},
|
|
{
|
|
"entropy": 0.651407066360116,
|
|
"epoch": 1.9044174919327919,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 1.4303703703703703e-05,
|
|
"loss": 0.752,
|
|
"mean_token_accuracy": 0.8161330189555883,
|
|
"num_tokens": 28662344.0,
|
|
"step": 1070
|
|
},
|
|
{
|
|
"entropy": 0.6537632567808032,
|
|
"epoch": 1.9222209858684767,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.4229629629629632e-05,
|
|
"loss": 0.7551,
|
|
"mean_token_accuracy": 0.8160409614443779,
|
|
"num_tokens": 28928769.0,
|
|
"step": 1080
|
|
},
|
|
{
|
|
"entropy": 0.6610247550532222,
|
|
"epoch": 1.9400244798041615,
|
|
"grad_norm": 1.5625,
|
|
"learning_rate": 1.4155555555555556e-05,
|
|
"loss": 0.7599,
|
|
"mean_token_accuracy": 0.8147491872310638,
|
|
"num_tokens": 29186214.0,
|
|
"step": 1090
|
|
},
|
|
{
|
|
"entropy": 0.6632727082818747,
|
|
"epoch": 1.9578279737398465,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 1.4081481481481483e-05,
|
|
"loss": 0.7706,
|
|
"mean_token_accuracy": 0.8126637618988752,
|
|
"num_tokens": 29446445.0,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.9578279737398465,
|
|
"eval_biology_entropy": 1.639104124546051,
|
|
"eval_biology_loss": 2.0316033363342285,
|
|
"eval_biology_mean_token_accuracy": 0.615292895078659,
|
|
"eval_biology_num_tokens": 29446445.0,
|
|
"eval_biology_runtime": 41.3569,
|
|
"eval_biology_samples_per_second": 12.09,
|
|
"eval_biology_steps_per_second": 3.022,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.9578279737398465,
|
|
"eval_chemistry_entropy": 0.6696053767204284,
|
|
"eval_chemistry_loss": 0.8191399574279785,
|
|
"eval_chemistry_mean_token_accuracy": 0.8085824422836304,
|
|
"eval_chemistry_num_tokens": 29446445.0,
|
|
"eval_chemistry_runtime": 50.7996,
|
|
"eval_chemistry_samples_per_second": 9.843,
|
|
"eval_chemistry_steps_per_second": 2.461,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.9578279737398465,
|
|
"eval_math_entropy": 0.6799165282249451,
|
|
"eval_math_loss": 1.0633617639541626,
|
|
"eval_math_mean_token_accuracy": 0.7729020061492919,
|
|
"eval_math_num_tokens": 29446445.0,
|
|
"eval_math_runtime": 52.3664,
|
|
"eval_math_samples_per_second": 9.548,
|
|
"eval_math_steps_per_second": 2.387,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 1.9578279737398465,
|
|
"eval_physics_entropy": 0.6961899037361146,
|
|
"eval_physics_loss": 0.8565983176231384,
|
|
"eval_physics_mean_token_accuracy": 0.7995611958503723,
|
|
"eval_physics_num_tokens": 29446445.0,
|
|
"eval_physics_runtime": 59.795,
|
|
"eval_physics_samples_per_second": 8.362,
|
|
"eval_physics_steps_per_second": 2.09,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"entropy": 0.6516633328050375,
|
|
"epoch": 1.9756314676755313,
|
|
"grad_norm": 1.21875,
|
|
"learning_rate": 1.400740740740741e-05,
|
|
"loss": 0.7439,
|
|
"mean_token_accuracy": 0.8187774941325188,
|
|
"num_tokens": 29711771.0,
|
|
"step": 1110
|
|
},
|
|
{
|
|
"entropy": 0.6508602360263467,
|
|
"epoch": 1.9934349616112161,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.3933333333333334e-05,
|
|
"loss": 0.7519,
|
|
"mean_token_accuracy": 0.8169589649885893,
|
|
"num_tokens": 29979994.0,
|
|
"step": 1120
|
|
},
|
|
{
|
|
"entropy": 0.6306788632946629,
|
|
"epoch": 2.010682096361411,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.385925925925926e-05,
|
|
"loss": 0.7224,
|
|
"mean_token_accuracy": 0.8239769354943306,
|
|
"num_tokens": 30240793.0,
|
|
"step": 1130
|
|
},
|
|
{
|
|
"entropy": 0.5911945894360542,
|
|
"epoch": 2.028485590297096,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.3785185185185186e-05,
|
|
"loss": 0.7,
|
|
"mean_token_accuracy": 0.8281658995896578,
|
|
"num_tokens": 30503988.0,
|
|
"step": 1140
|
|
},
|
|
{
|
|
"entropy": 0.6057484762743115,
|
|
"epoch": 2.0462890842327806,
|
|
"grad_norm": 1.2890625,
|
|
"learning_rate": 1.3711111111111112e-05,
|
|
"loss": 0.7031,
|
|
"mean_token_accuracy": 0.8266915556043386,
|
|
"num_tokens": 30771841.0,
|
|
"step": 1150
|
|
},
|
|
{
|
|
"entropy": 0.596652439981699,
|
|
"epoch": 2.0640925781684656,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 1.3637037037037037e-05,
|
|
"loss": 0.6926,
|
|
"mean_token_accuracy": 0.8293926887214184,
|
|
"num_tokens": 31045037.0,
|
|
"step": 1160
|
|
},
|
|
{
|
|
"entropy": 0.6352694857865572,
|
|
"epoch": 2.0818960721041506,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 1.3562962962962965e-05,
|
|
"loss": 0.7356,
|
|
"mean_token_accuracy": 0.8186026088893413,
|
|
"num_tokens": 31299575.0,
|
|
"step": 1170
|
|
},
|
|
{
|
|
"entropy": 0.6292696729302406,
|
|
"epoch": 2.099699566039835,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 1.3488888888888888e-05,
|
|
"loss": 0.7273,
|
|
"mean_token_accuracy": 0.8201477468013764,
|
|
"num_tokens": 31565963.0,
|
|
"step": 1180
|
|
},
|
|
{
|
|
"entropy": 0.5940109083428979,
|
|
"epoch": 2.1175030599755202,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.3414814814814817e-05,
|
|
"loss": 0.7035,
|
|
"mean_token_accuracy": 0.8278756473213434,
|
|
"num_tokens": 31834212.0,
|
|
"step": 1190
|
|
},
|
|
{
|
|
"entropy": 0.6235881226137281,
|
|
"epoch": 2.1353065539112053,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 1.3340740740740741e-05,
|
|
"loss": 0.7172,
|
|
"mean_token_accuracy": 0.8223393179476262,
|
|
"num_tokens": 32103356.0,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 2.1353065539112053,
|
|
"eval_biology_entropy": 1.4677531628608704,
|
|
"eval_biology_loss": 2.083951234817505,
|
|
"eval_biology_mean_token_accuracy": 0.6125459330081939,
|
|
"eval_biology_num_tokens": 32103356.0,
|
|
"eval_biology_runtime": 41.2722,
|
|
"eval_biology_samples_per_second": 12.115,
|
|
"eval_biology_steps_per_second": 3.029,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 2.1353065539112053,
|
|
"eval_chemistry_entropy": 0.6140689237117767,
|
|
"eval_chemistry_loss": 0.8232262134552002,
|
|
"eval_chemistry_mean_token_accuracy": 0.8088321657180786,
|
|
"eval_chemistry_num_tokens": 32103356.0,
|
|
"eval_chemistry_runtime": 50.8871,
|
|
"eval_chemistry_samples_per_second": 9.826,
|
|
"eval_chemistry_steps_per_second": 2.456,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 2.1353065539112053,
|
|
"eval_math_entropy": 0.6312986519336701,
|
|
"eval_math_loss": 1.0854209661483765,
|
|
"eval_math_mean_token_accuracy": 0.7729385108947754,
|
|
"eval_math_num_tokens": 32103356.0,
|
|
"eval_math_runtime": 52.4669,
|
|
"eval_math_samples_per_second": 9.53,
|
|
"eval_math_steps_per_second": 2.382,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 2.1353065539112053,
|
|
"eval_physics_entropy": 0.6365497968196869,
|
|
"eval_physics_loss": 0.8704355359077454,
|
|
"eval_physics_mean_token_accuracy": 0.7990298428535462,
|
|
"eval_physics_num_tokens": 32103356.0,
|
|
"eval_physics_runtime": 59.7275,
|
|
"eval_physics_samples_per_second": 8.371,
|
|
"eval_physics_steps_per_second": 2.093,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"entropy": 0.6137512635439635,
|
|
"epoch": 2.15311004784689,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 1.3266666666666668e-05,
|
|
"loss": 0.7043,
|
|
"mean_token_accuracy": 0.8242961999028922,
|
|
"num_tokens": 32374830.0,
|
|
"step": 1210
|
|
},
|
|
{
|
|
"entropy": 0.5896952305920422,
|
|
"epoch": 2.170913541782575,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.3192592592592594e-05,
|
|
"loss": 0.6922,
|
|
"mean_token_accuracy": 0.8302250444889069,
|
|
"num_tokens": 32643479.0,
|
|
"step": 1220
|
|
},
|
|
{
|
|
"entropy": 0.5944842832162976,
|
|
"epoch": 2.18871703571826,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.311851851851852e-05,
|
|
"loss": 0.6959,
|
|
"mean_token_accuracy": 0.8287800919264555,
|
|
"num_tokens": 32908761.0,
|
|
"step": 1230
|
|
},
|
|
{
|
|
"entropy": 0.6240290224552154,
|
|
"epoch": 2.2065205296539445,
|
|
"grad_norm": 1.21875,
|
|
"learning_rate": 1.3044444444444446e-05,
|
|
"loss": 0.7238,
|
|
"mean_token_accuracy": 0.8242842797189951,
|
|
"num_tokens": 33166413.0,
|
|
"step": 1240
|
|
},
|
|
{
|
|
"entropy": 0.6112470507621766,
|
|
"epoch": 2.2243240235896296,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 1.297037037037037e-05,
|
|
"loss": 0.7149,
|
|
"mean_token_accuracy": 0.8240214943885803,
|
|
"num_tokens": 33432686.0,
|
|
"step": 1250
|
|
},
|
|
{
|
|
"entropy": 0.6079921290278435,
|
|
"epoch": 2.242127517525314,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 1.2896296296296299e-05,
|
|
"loss": 0.7131,
|
|
"mean_token_accuracy": 0.8254695866256953,
|
|
"num_tokens": 33693242.0,
|
|
"step": 1260
|
|
},
|
|
{
|
|
"entropy": 0.5799209821969271,
|
|
"epoch": 2.259931011460999,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 1.2822222222222222e-05,
|
|
"loss": 0.6722,
|
|
"mean_token_accuracy": 0.834067041426897,
|
|
"num_tokens": 33970820.0,
|
|
"step": 1270
|
|
},
|
|
{
|
|
"entropy": 0.594209510087967,
|
|
"epoch": 2.277734505396684,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 1.274814814814815e-05,
|
|
"loss": 0.6977,
|
|
"mean_token_accuracy": 0.8276573233306408,
|
|
"num_tokens": 34243042.0,
|
|
"step": 1280
|
|
},
|
|
{
|
|
"entropy": 0.6097929472103715,
|
|
"epoch": 2.295537999332369,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.2674074074074075e-05,
|
|
"loss": 0.705,
|
|
"mean_token_accuracy": 0.8267240136861801,
|
|
"num_tokens": 34515236.0,
|
|
"step": 1290
|
|
},
|
|
{
|
|
"entropy": 0.5745832217857242,
|
|
"epoch": 2.313341493268054,
|
|
"grad_norm": 1.234375,
|
|
"learning_rate": 1.2600000000000001e-05,
|
|
"loss": 0.6783,
|
|
"mean_token_accuracy": 0.8345583423972129,
|
|
"num_tokens": 34785068.0,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 2.313341493268054,
|
|
"eval_biology_entropy": 1.497672739982605,
|
|
"eval_biology_loss": 2.0853865146636963,
|
|
"eval_biology_mean_token_accuracy": 0.613484436750412,
|
|
"eval_biology_num_tokens": 34785068.0,
|
|
"eval_biology_runtime": 43.469,
|
|
"eval_biology_samples_per_second": 11.502,
|
|
"eval_biology_steps_per_second": 2.876,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 2.313341493268054,
|
|
"eval_chemistry_entropy": 0.6273940706253052,
|
|
"eval_chemistry_loss": 0.8201014995574951,
|
|
"eval_chemistry_mean_token_accuracy": 0.8088535733222961,
|
|
"eval_chemistry_num_tokens": 34785068.0,
|
|
"eval_chemistry_runtime": 52.831,
|
|
"eval_chemistry_samples_per_second": 9.464,
|
|
"eval_chemistry_steps_per_second": 2.366,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 2.313341493268054,
|
|
"eval_math_entropy": 0.646281590461731,
|
|
"eval_math_loss": 1.0807055234909058,
|
|
"eval_math_mean_token_accuracy": 0.772442822933197,
|
|
"eval_math_num_tokens": 34785068.0,
|
|
"eval_math_runtime": 55.4339,
|
|
"eval_math_samples_per_second": 9.02,
|
|
"eval_math_steps_per_second": 2.255,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 2.313341493268054,
|
|
"eval_physics_entropy": 0.6572135488986969,
|
|
"eval_physics_loss": 0.8676888346672058,
|
|
"eval_physics_mean_token_accuracy": 0.7990684523582459,
|
|
"eval_physics_num_tokens": 34785068.0,
|
|
"eval_physics_runtime": 61.4064,
|
|
"eval_physics_samples_per_second": 8.142,
|
|
"eval_physics_steps_per_second": 2.036,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"entropy": 0.6191223243251442,
|
|
"epoch": 2.331144987203739,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 1.2525925925925928e-05,
|
|
"loss": 0.7122,
|
|
"mean_token_accuracy": 0.8234442934393883,
|
|
"num_tokens": 35055845.0,
|
|
"step": 1310
|
|
},
|
|
{
|
|
"entropy": 0.6000411529093981,
|
|
"epoch": 2.3489484811394234,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.2451851851851853e-05,
|
|
"loss": 0.7023,
|
|
"mean_token_accuracy": 0.8272519435733556,
|
|
"num_tokens": 35327481.0,
|
|
"step": 1320
|
|
},
|
|
{
|
|
"entropy": 0.6216201702132821,
|
|
"epoch": 2.3667519750751085,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 1.237777777777778e-05,
|
|
"loss": 0.7178,
|
|
"mean_token_accuracy": 0.8234907556325197,
|
|
"num_tokens": 35596362.0,
|
|
"step": 1330
|
|
},
|
|
{
|
|
"entropy": 0.5952032478526235,
|
|
"epoch": 2.3845554690107935,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 1.2303703703703704e-05,
|
|
"loss": 0.695,
|
|
"mean_token_accuracy": 0.827592471987009,
|
|
"num_tokens": 35869585.0,
|
|
"step": 1340
|
|
},
|
|
{
|
|
"entropy": 0.6152689423412084,
|
|
"epoch": 2.402358962946478,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 1.222962962962963e-05,
|
|
"loss": 0.7087,
|
|
"mean_token_accuracy": 0.8254514247179031,
|
|
"num_tokens": 36134072.0,
|
|
"step": 1350
|
|
},
|
|
{
|
|
"entropy": 0.5919087225571275,
|
|
"epoch": 2.420162456882163,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 1.2155555555555555e-05,
|
|
"loss": 0.6971,
|
|
"mean_token_accuracy": 0.829882950335741,
|
|
"num_tokens": 36406263.0,
|
|
"step": 1360
|
|
},
|
|
{
|
|
"entropy": 0.6080563787370921,
|
|
"epoch": 2.437965950817848,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 1.2081481481481484e-05,
|
|
"loss": 0.7027,
|
|
"mean_token_accuracy": 0.8268357224762439,
|
|
"num_tokens": 36676822.0,
|
|
"step": 1370
|
|
},
|
|
{
|
|
"entropy": 0.6048327511176467,
|
|
"epoch": 2.4557694447535328,
|
|
"grad_norm": 1.421875,
|
|
"learning_rate": 1.2007407407407408e-05,
|
|
"loss": 0.7047,
|
|
"mean_token_accuracy": 0.8271109748631715,
|
|
"num_tokens": 36942837.0,
|
|
"step": 1380
|
|
},
|
|
{
|
|
"entropy": 0.5988074177876115,
|
|
"epoch": 2.473572938689218,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 1.1933333333333335e-05,
|
|
"loss": 0.7051,
|
|
"mean_token_accuracy": 0.8256583914160729,
|
|
"num_tokens": 37210791.0,
|
|
"step": 1390
|
|
},
|
|
{
|
|
"entropy": 0.5873929725959897,
|
|
"epoch": 2.491376432624903,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 1.185925925925926e-05,
|
|
"loss": 0.6871,
|
|
"mean_token_accuracy": 0.83187070004642,
|
|
"num_tokens": 37485822.0,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 2.491376432624903,
|
|
"eval_biology_entropy": 1.4652894883155823,
|
|
"eval_biology_loss": 2.099928855895996,
|
|
"eval_biology_mean_token_accuracy": 0.6130364212989807,
|
|
"eval_biology_num_tokens": 37485822.0,
|
|
"eval_biology_runtime": 43.3395,
|
|
"eval_biology_samples_per_second": 11.537,
|
|
"eval_biology_steps_per_second": 2.884,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 2.491376432624903,
|
|
"eval_chemistry_entropy": 0.6137497539520264,
|
|
"eval_chemistry_loss": 0.8189122080802917,
|
|
"eval_chemistry_mean_token_accuracy": 0.8097302551269532,
|
|
"eval_chemistry_num_tokens": 37485822.0,
|
|
"eval_chemistry_runtime": 52.903,
|
|
"eval_chemistry_samples_per_second": 9.451,
|
|
"eval_chemistry_steps_per_second": 2.363,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 2.491376432624903,
|
|
"eval_math_entropy": 0.6390214171409607,
|
|
"eval_math_loss": 1.082951545715332,
|
|
"eval_math_mean_token_accuracy": 0.772076274394989,
|
|
"eval_math_num_tokens": 37485822.0,
|
|
"eval_math_runtime": 54.1184,
|
|
"eval_math_samples_per_second": 9.239,
|
|
"eval_math_steps_per_second": 2.31,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 2.491376432624903,
|
|
"eval_physics_entropy": 0.6406208200454712,
|
|
"eval_physics_loss": 0.8697341084480286,
|
|
"eval_physics_mean_token_accuracy": 0.7989166221618652,
|
|
"eval_physics_num_tokens": 37485822.0,
|
|
"eval_physics_runtime": 61.5427,
|
|
"eval_physics_samples_per_second": 8.124,
|
|
"eval_physics_steps_per_second": 2.031,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"entropy": 0.6010861510410905,
|
|
"epoch": 2.5091799265605874,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 1.1785185185185186e-05,
|
|
"loss": 0.7014,
|
|
"mean_token_accuracy": 0.8287522595375776,
|
|
"num_tokens": 37746150.0,
|
|
"step": 1410
|
|
},
|
|
{
|
|
"entropy": 0.5981267247349024,
|
|
"epoch": 2.5269834204962724,
|
|
"grad_norm": 1.3046875,
|
|
"learning_rate": 1.1711111111111113e-05,
|
|
"loss": 0.6962,
|
|
"mean_token_accuracy": 0.8269948281347752,
|
|
"num_tokens": 38016863.0,
|
|
"step": 1420
|
|
},
|
|
{
|
|
"entropy": 0.6277626631781459,
|
|
"epoch": 2.544786914431957,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 1.1637037037037037e-05,
|
|
"loss": 0.7222,
|
|
"mean_token_accuracy": 0.8205803319811821,
|
|
"num_tokens": 38277754.0,
|
|
"step": 1430
|
|
},
|
|
{
|
|
"entropy": 0.6118797251954675,
|
|
"epoch": 2.562590408367642,
|
|
"grad_norm": 1.421875,
|
|
"learning_rate": 1.1562962962962964e-05,
|
|
"loss": 0.7172,
|
|
"mean_token_accuracy": 0.8237183388322592,
|
|
"num_tokens": 38542761.0,
|
|
"step": 1440
|
|
},
|
|
{
|
|
"entropy": 0.6034180456772447,
|
|
"epoch": 2.580393902303327,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 1.1488888888888889e-05,
|
|
"loss": 0.6941,
|
|
"mean_token_accuracy": 0.8272098813205957,
|
|
"num_tokens": 38820999.0,
|
|
"step": 1450
|
|
},
|
|
{
|
|
"entropy": 0.5998265566304326,
|
|
"epoch": 2.598197396239012,
|
|
"grad_norm": 1.3203125,
|
|
"learning_rate": 1.1414814814814817e-05,
|
|
"loss": 0.7114,
|
|
"mean_token_accuracy": 0.8267886329442262,
|
|
"num_tokens": 39080016.0,
|
|
"step": 1460
|
|
},
|
|
{
|
|
"entropy": 0.6005334880203008,
|
|
"epoch": 2.6160008901746967,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 1.1340740740740742e-05,
|
|
"loss": 0.6949,
|
|
"mean_token_accuracy": 0.8299154710024595,
|
|
"num_tokens": 39345434.0,
|
|
"step": 1470
|
|
},
|
|
{
|
|
"entropy": 0.6003408120945096,
|
|
"epoch": 2.6338043841103818,
|
|
"grad_norm": 1.46875,
|
|
"learning_rate": 1.1266666666666668e-05,
|
|
"loss": 0.7033,
|
|
"mean_token_accuracy": 0.8266183461993932,
|
|
"num_tokens": 39610041.0,
|
|
"step": 1480
|
|
},
|
|
{
|
|
"entropy": 0.5847593653947115,
|
|
"epoch": 2.6516078780460663,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 1.1192592592592593e-05,
|
|
"loss": 0.6785,
|
|
"mean_token_accuracy": 0.8323042623698711,
|
|
"num_tokens": 39884185.0,
|
|
"step": 1490
|
|
},
|
|
{
|
|
"entropy": 0.6105206938460469,
|
|
"epoch": 2.6694113719817514,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 1.111851851851852e-05,
|
|
"loss": 0.7201,
|
|
"mean_token_accuracy": 0.8220306947827339,
|
|
"num_tokens": 40143543.0,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 2.6694113719817514,
|
|
"eval_biology_entropy": 1.490638756752014,
|
|
"eval_biology_loss": 2.0750932693481445,
|
|
"eval_biology_mean_token_accuracy": 0.6143683531284332,
|
|
"eval_biology_num_tokens": 40143543.0,
|
|
"eval_biology_runtime": 43.2253,
|
|
"eval_biology_samples_per_second": 11.567,
|
|
"eval_biology_steps_per_second": 2.892,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 2.6694113719817514,
|
|
"eval_chemistry_entropy": 0.623803097486496,
|
|
"eval_chemistry_loss": 0.8160935044288635,
|
|
"eval_chemistry_mean_token_accuracy": 0.8098534879684448,
|
|
"eval_chemistry_num_tokens": 40143543.0,
|
|
"eval_chemistry_runtime": 50.807,
|
|
"eval_chemistry_samples_per_second": 9.841,
|
|
"eval_chemistry_steps_per_second": 2.46,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 2.6694113719817514,
|
|
"eval_math_entropy": 0.645468202829361,
|
|
"eval_math_loss": 1.0840471982955933,
|
|
"eval_math_mean_token_accuracy": 0.7726303563117981,
|
|
"eval_math_num_tokens": 40143543.0,
|
|
"eval_math_runtime": 53.0054,
|
|
"eval_math_samples_per_second": 9.433,
|
|
"eval_math_steps_per_second": 2.358,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 2.6694113719817514,
|
|
"eval_physics_entropy": 0.6492172837257385,
|
|
"eval_physics_loss": 0.8670127391815186,
|
|
"eval_physics_mean_token_accuracy": 0.7991857075691223,
|
|
"eval_physics_num_tokens": 40143543.0,
|
|
"eval_physics_runtime": 63.9696,
|
|
"eval_physics_samples_per_second": 7.816,
|
|
"eval_physics_steps_per_second": 1.954,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"entropy": 0.6046247070655226,
|
|
"epoch": 2.6872148659174364,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 1.1044444444444444e-05,
|
|
"loss": 0.7007,
|
|
"mean_token_accuracy": 0.828262010216713,
|
|
"num_tokens": 40406812.0,
|
|
"step": 1510
|
|
},
|
|
{
|
|
"entropy": 0.6191225880756974,
|
|
"epoch": 2.7050183598531214,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 1.0970370370370371e-05,
|
|
"loss": 0.7269,
|
|
"mean_token_accuracy": 0.8213932007551193,
|
|
"num_tokens": 40668543.0,
|
|
"step": 1520
|
|
},
|
|
{
|
|
"entropy": 0.6065074443817139,
|
|
"epoch": 2.722821853788806,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 1.0896296296296298e-05,
|
|
"loss": 0.6989,
|
|
"mean_token_accuracy": 0.8265019495040178,
|
|
"num_tokens": 40942603.0,
|
|
"step": 1530
|
|
},
|
|
{
|
|
"entropy": 0.6138417391106487,
|
|
"epoch": 2.740625347724491,
|
|
"grad_norm": 1.34375,
|
|
"learning_rate": 1.0822222222222222e-05,
|
|
"loss": 0.7057,
|
|
"mean_token_accuracy": 0.8261400748044252,
|
|
"num_tokens": 41208093.0,
|
|
"step": 1540
|
|
},
|
|
{
|
|
"entropy": 0.5985930340364576,
|
|
"epoch": 2.7584288416601757,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 1.074814814814815e-05,
|
|
"loss": 0.7061,
|
|
"mean_token_accuracy": 0.8252453569322824,
|
|
"num_tokens": 41477562.0,
|
|
"step": 1550
|
|
},
|
|
{
|
|
"entropy": 0.5930986102670431,
|
|
"epoch": 2.7762323355958607,
|
|
"grad_norm": 1.5234375,
|
|
"learning_rate": 1.0674074074074074e-05,
|
|
"loss": 0.6935,
|
|
"mean_token_accuracy": 0.8292474631220103,
|
|
"num_tokens": 41746571.0,
|
|
"step": 1560
|
|
},
|
|
{
|
|
"entropy": 0.6006103545427323,
|
|
"epoch": 2.7940358295315457,
|
|
"grad_norm": 1.4765625,
|
|
"learning_rate": 1.0600000000000002e-05,
|
|
"loss": 0.7025,
|
|
"mean_token_accuracy": 0.8271383382380009,
|
|
"num_tokens": 42011633.0,
|
|
"step": 1570
|
|
},
|
|
{
|
|
"entropy": 0.6129128225147724,
|
|
"epoch": 2.8118393234672303,
|
|
"grad_norm": 1.328125,
|
|
"learning_rate": 1.0525925925925927e-05,
|
|
"loss": 0.7076,
|
|
"mean_token_accuracy": 0.8252990577369929,
|
|
"num_tokens": 42283087.0,
|
|
"step": 1580
|
|
},
|
|
{
|
|
"entropy": 0.5756963772699237,
|
|
"epoch": 2.8296428174029153,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 1.0451851851851853e-05,
|
|
"loss": 0.6725,
|
|
"mean_token_accuracy": 0.8329845789819956,
|
|
"num_tokens": 42561534.0,
|
|
"step": 1590
|
|
},
|
|
{
|
|
"entropy": 0.6044865250587463,
|
|
"epoch": 2.8474463113386,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 1.0377777777777778e-05,
|
|
"loss": 0.705,
|
|
"mean_token_accuracy": 0.8281082704663276,
|
|
"num_tokens": 42834266.0,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 2.8474463113386,
|
|
"eval_biology_entropy": 1.4930514478683472,
|
|
"eval_biology_loss": 2.0736353397369385,
|
|
"eval_biology_mean_token_accuracy": 0.6145954511165619,
|
|
"eval_biology_num_tokens": 42834266.0,
|
|
"eval_biology_runtime": 40.9171,
|
|
"eval_biology_samples_per_second": 12.22,
|
|
"eval_biology_steps_per_second": 3.055,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 2.8474463113386,
|
|
"eval_chemistry_entropy": 0.6211945106983184,
|
|
"eval_chemistry_loss": 0.8145589828491211,
|
|
"eval_chemistry_mean_token_accuracy": 0.810048574924469,
|
|
"eval_chemistry_num_tokens": 42834266.0,
|
|
"eval_chemistry_runtime": 50.7089,
|
|
"eval_chemistry_samples_per_second": 9.86,
|
|
"eval_chemistry_steps_per_second": 2.465,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 2.8474463113386,
|
|
"eval_math_entropy": 0.6438178873062134,
|
|
"eval_math_loss": 1.0831515789031982,
|
|
"eval_math_mean_token_accuracy": 0.772301673412323,
|
|
"eval_math_num_tokens": 42834266.0,
|
|
"eval_math_runtime": 51.7388,
|
|
"eval_math_samples_per_second": 9.664,
|
|
"eval_math_steps_per_second": 2.416,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 2.8474463113386,
|
|
"eval_physics_entropy": 0.649389057636261,
|
|
"eval_physics_loss": 0.86627197265625,
|
|
"eval_physics_mean_token_accuracy": 0.7995732502937317,
|
|
"eval_physics_num_tokens": 42834266.0,
|
|
"eval_physics_runtime": 59.549,
|
|
"eval_physics_samples_per_second": 8.396,
|
|
"eval_physics_steps_per_second": 2.099,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"entropy": 0.625230243243277,
|
|
"epoch": 2.865249805274285,
|
|
"grad_norm": 1.46875,
|
|
"learning_rate": 1.0303703703703705e-05,
|
|
"loss": 0.7228,
|
|
"mean_token_accuracy": 0.8225024774670601,
|
|
"num_tokens": 43100513.0,
|
|
"step": 1610
|
|
},
|
|
{
|
|
"entropy": 0.605365008302033,
|
|
"epoch": 2.88305329920997,
|
|
"grad_norm": 1.359375,
|
|
"learning_rate": 1.0229629629629631e-05,
|
|
"loss": 0.698,
|
|
"mean_token_accuracy": 0.8268981877714395,
|
|
"num_tokens": 43375632.0,
|
|
"step": 1620
|
|
},
|
|
{
|
|
"entropy": 0.5911450985819101,
|
|
"epoch": 2.900856793145655,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 1.0155555555555556e-05,
|
|
"loss": 0.6958,
|
|
"mean_token_accuracy": 0.829645611718297,
|
|
"num_tokens": 43644341.0,
|
|
"step": 1630
|
|
},
|
|
{
|
|
"entropy": 0.5920769222080707,
|
|
"epoch": 2.9186602870813396,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 1.0081481481481484e-05,
|
|
"loss": 0.6833,
|
|
"mean_token_accuracy": 0.8301208417862653,
|
|
"num_tokens": 43917580.0,
|
|
"step": 1640
|
|
},
|
|
{
|
|
"entropy": 0.5912358593195677,
|
|
"epoch": 2.9364637810170247,
|
|
"grad_norm": 1.359375,
|
|
"learning_rate": 1.0007407407407407e-05,
|
|
"loss": 0.6989,
|
|
"mean_token_accuracy": 0.8305816825479269,
|
|
"num_tokens": 44177325.0,
|
|
"step": 1650
|
|
},
|
|
{
|
|
"entropy": 0.6072682781144977,
|
|
"epoch": 2.9542672749527092,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 9.933333333333334e-06,
|
|
"loss": 0.7079,
|
|
"mean_token_accuracy": 0.8247454788535833,
|
|
"num_tokens": 44442947.0,
|
|
"step": 1660
|
|
},
|
|
{
|
|
"entropy": 0.6202316217124462,
|
|
"epoch": 2.9720707688883943,
|
|
"grad_norm": 1.484375,
|
|
"learning_rate": 9.85925925925926e-06,
|
|
"loss": 0.7181,
|
|
"mean_token_accuracy": 0.8235241327434778,
|
|
"num_tokens": 44700601.0,
|
|
"step": 1670
|
|
},
|
|
{
|
|
"entropy": 0.5890740664675832,
|
|
"epoch": 2.9898742628240793,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 9.785185185185187e-06,
|
|
"loss": 0.6933,
|
|
"mean_token_accuracy": 0.8307324200868607,
|
|
"num_tokens": 44971479.0,
|
|
"step": 1680
|
|
},
|
|
{
|
|
"entropy": 0.587316531712009,
|
|
"epoch": 3.007121397574274,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 9.711111111111111e-06,
|
|
"loss": 0.6682,
|
|
"mean_token_accuracy": 0.8340173459822132,
|
|
"num_tokens": 45231796.0,
|
|
"step": 1690
|
|
},
|
|
{
|
|
"entropy": 0.6013597913086415,
|
|
"epoch": 3.0249248915099587,
|
|
"grad_norm": 1.5,
|
|
"learning_rate": 9.637037037037038e-06,
|
|
"loss": 0.7048,
|
|
"mean_token_accuracy": 0.8241627678275109,
|
|
"num_tokens": 45500317.0,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 3.0249248915099587,
|
|
"eval_biology_entropy": 1.4193837118148804,
|
|
"eval_biology_loss": 2.11480712890625,
|
|
"eval_biology_mean_token_accuracy": 0.6124376230239869,
|
|
"eval_biology_num_tokens": 45500317.0,
|
|
"eval_biology_runtime": 40.3875,
|
|
"eval_biology_samples_per_second": 12.38,
|
|
"eval_biology_steps_per_second": 3.095,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 3.0249248915099587,
|
|
"eval_chemistry_entropy": 0.5911917226314545,
|
|
"eval_chemistry_loss": 0.8190613985061646,
|
|
"eval_chemistry_mean_token_accuracy": 0.8099429054260254,
|
|
"eval_chemistry_num_tokens": 45500317.0,
|
|
"eval_chemistry_runtime": 50.2786,
|
|
"eval_chemistry_samples_per_second": 9.945,
|
|
"eval_chemistry_steps_per_second": 2.486,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 3.0249248915099587,
|
|
"eval_math_entropy": 0.6176868407726288,
|
|
"eval_math_loss": 1.0993149280548096,
|
|
"eval_math_mean_token_accuracy": 0.7718941388130188,
|
|
"eval_math_num_tokens": 45500317.0,
|
|
"eval_math_runtime": 51.8195,
|
|
"eval_math_samples_per_second": 9.649,
|
|
"eval_math_steps_per_second": 2.412,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 3.0249248915099587,
|
|
"eval_physics_entropy": 0.6201707119941712,
|
|
"eval_physics_loss": 0.8766935467720032,
|
|
"eval_physics_mean_token_accuracy": 0.7985301322937012,
|
|
"eval_physics_num_tokens": 45500317.0,
|
|
"eval_physics_runtime": 59.4252,
|
|
"eval_physics_samples_per_second": 8.414,
|
|
"eval_physics_steps_per_second": 2.103,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"entropy": 0.5692525265738368,
|
|
"epoch": 3.0427283854456437,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 9.562962962962965e-06,
|
|
"loss": 0.6591,
|
|
"mean_token_accuracy": 0.8374071232974529,
|
|
"num_tokens": 45762680.0,
|
|
"step": 1710
|
|
},
|
|
{
|
|
"entropy": 0.5559542520903051,
|
|
"epoch": 3.0605318793813288,
|
|
"grad_norm": 1.625,
|
|
"learning_rate": 9.48888888888889e-06,
|
|
"loss": 0.6597,
|
|
"mean_token_accuracy": 0.8373042322695256,
|
|
"num_tokens": 46029923.0,
|
|
"step": 1720
|
|
},
|
|
{
|
|
"entropy": 0.5692304980009795,
|
|
"epoch": 3.0783353733170133,
|
|
"grad_norm": 1.328125,
|
|
"learning_rate": 9.414814814814816e-06,
|
|
"loss": 0.6629,
|
|
"mean_token_accuracy": 0.8344687603414058,
|
|
"num_tokens": 46305156.0,
|
|
"step": 1730
|
|
},
|
|
{
|
|
"entropy": 0.5508972069248557,
|
|
"epoch": 3.0961388672526984,
|
|
"grad_norm": 1.484375,
|
|
"learning_rate": 9.34074074074074e-06,
|
|
"loss": 0.6501,
|
|
"mean_token_accuracy": 0.8386828090995551,
|
|
"num_tokens": 46582874.0,
|
|
"step": 1740
|
|
},
|
|
{
|
|
"entropy": 0.5854626100510358,
|
|
"epoch": 3.1139423611883834,
|
|
"grad_norm": 1.640625,
|
|
"learning_rate": 9.266666666666667e-06,
|
|
"loss": 0.6824,
|
|
"mean_token_accuracy": 0.8301110483705998,
|
|
"num_tokens": 46850175.0,
|
|
"step": 1750
|
|
},
|
|
{
|
|
"entropy": 0.5591677505522966,
|
|
"epoch": 3.131745855124068,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 9.192592592592594e-06,
|
|
"loss": 0.6564,
|
|
"mean_token_accuracy": 0.8384724948555231,
|
|
"num_tokens": 47125899.0,
|
|
"step": 1760
|
|
},
|
|
{
|
|
"entropy": 0.5725911458954215,
|
|
"epoch": 3.149549349059753,
|
|
"grad_norm": 1.546875,
|
|
"learning_rate": 9.118518518518518e-06,
|
|
"loss": 0.6752,
|
|
"mean_token_accuracy": 0.8343269538134337,
|
|
"num_tokens": 47390194.0,
|
|
"step": 1770
|
|
},
|
|
{
|
|
"entropy": 0.5758154878392816,
|
|
"epoch": 3.167352842995438,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 9.044444444444445e-06,
|
|
"loss": 0.6738,
|
|
"mean_token_accuracy": 0.8336336623877287,
|
|
"num_tokens": 47654784.0,
|
|
"step": 1780
|
|
},
|
|
{
|
|
"entropy": 0.5612063113600015,
|
|
"epoch": 3.1851563369311227,
|
|
"grad_norm": 1.53125,
|
|
"learning_rate": 8.970370370370372e-06,
|
|
"loss": 0.668,
|
|
"mean_token_accuracy": 0.8369329560548067,
|
|
"num_tokens": 47919400.0,
|
|
"step": 1790
|
|
},
|
|
{
|
|
"entropy": 0.5750826884061098,
|
|
"epoch": 3.2029598308668077,
|
|
"grad_norm": 1.5625,
|
|
"learning_rate": 8.896296296296298e-06,
|
|
"loss": 0.6761,
|
|
"mean_token_accuracy": 0.8343805599957704,
|
|
"num_tokens": 48187796.0,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 3.2029598308668077,
|
|
"eval_biology_entropy": 1.4034225873947144,
|
|
"eval_biology_loss": 2.1275665760040283,
|
|
"eval_biology_mean_token_accuracy": 0.6119487595558166,
|
|
"eval_biology_num_tokens": 48187796.0,
|
|
"eval_biology_runtime": 40.734,
|
|
"eval_biology_samples_per_second": 12.275,
|
|
"eval_biology_steps_per_second": 3.069,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 3.2029598308668077,
|
|
"eval_chemistry_entropy": 0.5955355260372162,
|
|
"eval_chemistry_loss": 0.8196625709533691,
|
|
"eval_chemistry_mean_token_accuracy": 0.8102530264854431,
|
|
"eval_chemistry_num_tokens": 48187796.0,
|
|
"eval_chemistry_runtime": 50.2147,
|
|
"eval_chemistry_samples_per_second": 9.957,
|
|
"eval_chemistry_steps_per_second": 2.489,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 3.2029598308668077,
|
|
"eval_math_entropy": 0.6184592893123627,
|
|
"eval_math_loss": 1.1008187532424927,
|
|
"eval_math_mean_token_accuracy": 0.7713511042594909,
|
|
"eval_math_num_tokens": 48187796.0,
|
|
"eval_math_runtime": 51.7678,
|
|
"eval_math_samples_per_second": 9.659,
|
|
"eval_math_steps_per_second": 2.415,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 3.2029598308668077,
|
|
"eval_physics_entropy": 0.6223777911663055,
|
|
"eval_physics_loss": 0.8791524767875671,
|
|
"eval_physics_mean_token_accuracy": 0.7979751496315003,
|
|
"eval_physics_num_tokens": 48187796.0,
|
|
"eval_physics_runtime": 59.0178,
|
|
"eval_physics_samples_per_second": 8.472,
|
|
"eval_physics_steps_per_second": 2.118,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"entropy": 0.5663457477465272,
|
|
"epoch": 3.2207633248024923,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 8.822222222222223e-06,
|
|
"loss": 0.666,
|
|
"mean_token_accuracy": 0.8364688046276569,
|
|
"num_tokens": 48457428.0,
|
|
"step": 1810
|
|
},
|
|
{
|
|
"entropy": 0.6113261075690388,
|
|
"epoch": 3.2385668187381773,
|
|
"grad_norm": 1.515625,
|
|
"learning_rate": 8.74814814814815e-06,
|
|
"loss": 0.704,
|
|
"mean_token_accuracy": 0.8250963229686021,
|
|
"num_tokens": 48716417.0,
|
|
"step": 1820
|
|
},
|
|
{
|
|
"entropy": 0.5722766594961286,
|
|
"epoch": 3.2563703126738623,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 8.674074074074074e-06,
|
|
"loss": 0.6724,
|
|
"mean_token_accuracy": 0.8353833951056003,
|
|
"num_tokens": 48977306.0,
|
|
"step": 1830
|
|
},
|
|
{
|
|
"entropy": 0.5715915774926543,
|
|
"epoch": 3.274173806609547,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 8.6e-06,
|
|
"loss": 0.6663,
|
|
"mean_token_accuracy": 0.8353964131325483,
|
|
"num_tokens": 49249667.0,
|
|
"step": 1840
|
|
},
|
|
{
|
|
"entropy": 0.5664192967116832,
|
|
"epoch": 3.291977300545232,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 8.525925925925927e-06,
|
|
"loss": 0.6619,
|
|
"mean_token_accuracy": 0.8355418112128973,
|
|
"num_tokens": 49520900.0,
|
|
"step": 1850
|
|
},
|
|
{
|
|
"entropy": 0.5784041214734316,
|
|
"epoch": 3.309780794480917,
|
|
"grad_norm": 1.484375,
|
|
"learning_rate": 8.451851851851852e-06,
|
|
"loss": 0.6816,
|
|
"mean_token_accuracy": 0.8331022556871176,
|
|
"num_tokens": 49786411.0,
|
|
"step": 1860
|
|
},
|
|
{
|
|
"entropy": 0.5709506403654814,
|
|
"epoch": 3.3275842884166016,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 8.377777777777779e-06,
|
|
"loss": 0.6589,
|
|
"mean_token_accuracy": 0.8369544640183448,
|
|
"num_tokens": 50056207.0,
|
|
"step": 1870
|
|
},
|
|
{
|
|
"entropy": 0.5751367569901049,
|
|
"epoch": 3.3453877823522866,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 8.303703703703705e-06,
|
|
"loss": 0.6735,
|
|
"mean_token_accuracy": 0.8336179967969656,
|
|
"num_tokens": 50330876.0,
|
|
"step": 1880
|
|
},
|
|
{
|
|
"entropy": 0.577986160852015,
|
|
"epoch": 3.3631912762879717,
|
|
"grad_norm": 1.6015625,
|
|
"learning_rate": 8.229629629629632e-06,
|
|
"loss": 0.6774,
|
|
"mean_token_accuracy": 0.8319189596921206,
|
|
"num_tokens": 50602004.0,
|
|
"step": 1890
|
|
},
|
|
{
|
|
"entropy": 0.5784968495368957,
|
|
"epoch": 3.3809947702236562,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 8.155555555555556e-06,
|
|
"loss": 0.6728,
|
|
"mean_token_accuracy": 0.8320122357457876,
|
|
"num_tokens": 50873444.0,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 3.3809947702236562,
|
|
"eval_biology_entropy": 1.379322585105896,
|
|
"eval_biology_loss": 2.1383986473083496,
|
|
"eval_biology_mean_token_accuracy": 0.6115301356315613,
|
|
"eval_biology_num_tokens": 50873444.0,
|
|
"eval_biology_runtime": 40.4711,
|
|
"eval_biology_samples_per_second": 12.354,
|
|
"eval_biology_steps_per_second": 3.089,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 3.3809947702236562,
|
|
"eval_chemistry_entropy": 0.5861142361164093,
|
|
"eval_chemistry_loss": 0.8205327391624451,
|
|
"eval_chemistry_mean_token_accuracy": 0.8100807156562805,
|
|
"eval_chemistry_num_tokens": 50873444.0,
|
|
"eval_chemistry_runtime": 50.1133,
|
|
"eval_chemistry_samples_per_second": 9.977,
|
|
"eval_chemistry_steps_per_second": 2.494,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 3.3809947702236562,
|
|
"eval_math_entropy": 0.6114957261085511,
|
|
"eval_math_loss": 1.1055864095687866,
|
|
"eval_math_mean_token_accuracy": 0.7713269739151001,
|
|
"eval_math_num_tokens": 50873444.0,
|
|
"eval_math_runtime": 51.5676,
|
|
"eval_math_samples_per_second": 9.696,
|
|
"eval_math_steps_per_second": 2.424,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 3.3809947702236562,
|
|
"eval_physics_entropy": 0.6135163457393646,
|
|
"eval_physics_loss": 0.8826585412025452,
|
|
"eval_physics_mean_token_accuracy": 0.7979225769042969,
|
|
"eval_physics_num_tokens": 50873444.0,
|
|
"eval_physics_runtime": 58.9496,
|
|
"eval_physics_samples_per_second": 8.482,
|
|
"eval_physics_steps_per_second": 2.12,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"entropy": 0.5896684370934964,
|
|
"epoch": 3.3987982641593413,
|
|
"grad_norm": 1.5234375,
|
|
"learning_rate": 8.081481481481483e-06,
|
|
"loss": 0.6862,
|
|
"mean_token_accuracy": 0.8290154490619898,
|
|
"num_tokens": 51141077.0,
|
|
"step": 1910
|
|
},
|
|
{
|
|
"entropy": 0.5699383469298482,
|
|
"epoch": 3.4166017580950263,
|
|
"grad_norm": 1.578125,
|
|
"learning_rate": 8.007407407407408e-06,
|
|
"loss": 0.6705,
|
|
"mean_token_accuracy": 0.8345434635877609,
|
|
"num_tokens": 51406483.0,
|
|
"step": 1920
|
|
},
|
|
{
|
|
"entropy": 0.5792136957868934,
|
|
"epoch": 3.434405252030711,
|
|
"grad_norm": 1.546875,
|
|
"learning_rate": 7.933333333333334e-06,
|
|
"loss": 0.6753,
|
|
"mean_token_accuracy": 0.8337912186980247,
|
|
"num_tokens": 51668458.0,
|
|
"step": 1930
|
|
},
|
|
{
|
|
"entropy": 0.5839371480047703,
|
|
"epoch": 3.452208745966396,
|
|
"grad_norm": 1.53125,
|
|
"learning_rate": 7.859259259259259e-06,
|
|
"loss": 0.6891,
|
|
"mean_token_accuracy": 0.8307471171021461,
|
|
"num_tokens": 51928199.0,
|
|
"step": 1940
|
|
},
|
|
{
|
|
"entropy": 0.574542698264122,
|
|
"epoch": 3.470012239902081,
|
|
"grad_norm": 1.4140625,
|
|
"learning_rate": 7.785185185185185e-06,
|
|
"loss": 0.6653,
|
|
"mean_token_accuracy": 0.8337355487048626,
|
|
"num_tokens": 52199244.0,
|
|
"step": 1950
|
|
},
|
|
{
|
|
"entropy": 0.5655553802847862,
|
|
"epoch": 3.4878157338377656,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 7.711111111111112e-06,
|
|
"loss": 0.6606,
|
|
"mean_token_accuracy": 0.8357041519135237,
|
|
"num_tokens": 52473476.0,
|
|
"step": 1960
|
|
},
|
|
{
|
|
"entropy": 0.5800750594586134,
|
|
"epoch": 3.5056192277734506,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 7.637037037037037e-06,
|
|
"loss": 0.6777,
|
|
"mean_token_accuracy": 0.832566126063466,
|
|
"num_tokens": 52743275.0,
|
|
"step": 1970
|
|
},
|
|
{
|
|
"entropy": 0.5716741790995001,
|
|
"epoch": 3.523422721709135,
|
|
"grad_norm": 1.4375,
|
|
"learning_rate": 7.562962962962963e-06,
|
|
"loss": 0.6637,
|
|
"mean_token_accuracy": 0.8349632371217013,
|
|
"num_tokens": 53014827.0,
|
|
"step": 1980
|
|
},
|
|
{
|
|
"entropy": 0.5780819065868854,
|
|
"epoch": 3.54122621564482,
|
|
"grad_norm": 1.5859375,
|
|
"learning_rate": 7.48888888888889e-06,
|
|
"loss": 0.6791,
|
|
"mean_token_accuracy": 0.8327229652553797,
|
|
"num_tokens": 53285577.0,
|
|
"step": 1990
|
|
},
|
|
{
|
|
"entropy": 0.5571762975305319,
|
|
"epoch": 3.5590297095805052,
|
|
"grad_norm": 1.5,
|
|
"learning_rate": 7.4148148148148155e-06,
|
|
"loss": 0.656,
|
|
"mean_token_accuracy": 0.8386137288063764,
|
|
"num_tokens": 53553391.0,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 3.5590297095805052,
|
|
"eval_biology_entropy": 1.381414409160614,
|
|
"eval_biology_loss": 2.13814115524292,
|
|
"eval_biology_mean_token_accuracy": 0.6120853550434112,
|
|
"eval_biology_num_tokens": 53553391.0,
|
|
"eval_biology_runtime": 40.2984,
|
|
"eval_biology_samples_per_second": 12.407,
|
|
"eval_biology_steps_per_second": 3.102,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 3.5590297095805052,
|
|
"eval_chemistry_entropy": 0.5889808633327485,
|
|
"eval_chemistry_loss": 0.8193114995956421,
|
|
"eval_chemistry_mean_token_accuracy": 0.8103628730773926,
|
|
"eval_chemistry_num_tokens": 53553391.0,
|
|
"eval_chemistry_runtime": 49.8517,
|
|
"eval_chemistry_samples_per_second": 10.03,
|
|
"eval_chemistry_steps_per_second": 2.507,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 3.5590297095805052,
|
|
"eval_math_entropy": 0.6149049291610718,
|
|
"eval_math_loss": 1.1041083335876465,
|
|
"eval_math_mean_token_accuracy": 0.7715086970329285,
|
|
"eval_math_num_tokens": 53553391.0,
|
|
"eval_math_runtime": 51.0164,
|
|
"eval_math_samples_per_second": 9.801,
|
|
"eval_math_steps_per_second": 2.45,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 3.5590297095805052,
|
|
"eval_physics_entropy": 0.6177053947448731,
|
|
"eval_physics_loss": 0.8811244368553162,
|
|
"eval_physics_mean_token_accuracy": 0.798056589126587,
|
|
"eval_physics_num_tokens": 53553391.0,
|
|
"eval_physics_runtime": 58.0264,
|
|
"eval_physics_samples_per_second": 8.617,
|
|
"eval_physics_steps_per_second": 2.154,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"entropy": 0.5717996397987009,
|
|
"epoch": 3.5768332035161903,
|
|
"grad_norm": 1.4453125,
|
|
"learning_rate": 7.340740740740742e-06,
|
|
"loss": 0.6683,
|
|
"mean_token_accuracy": 0.835101555287838,
|
|
"num_tokens": 53826074.0,
|
|
"step": 2010
|
|
},
|
|
{
|
|
"entropy": 0.5765499664470554,
|
|
"epoch": 3.594636697451875,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 7.266666666666668e-06,
|
|
"loss": 0.6758,
|
|
"mean_token_accuracy": 0.833604721352458,
|
|
"num_tokens": 54090804.0,
|
|
"step": 2020
|
|
},
|
|
{
|
|
"entropy": 0.5620788568630815,
|
|
"epoch": 3.61244019138756,
|
|
"grad_norm": 1.4921875,
|
|
"learning_rate": 7.192592592592593e-06,
|
|
"loss": 0.6652,
|
|
"mean_token_accuracy": 0.8382882464677095,
|
|
"num_tokens": 54350633.0,
|
|
"step": 2030
|
|
},
|
|
{
|
|
"entropy": 0.5829729141667486,
|
|
"epoch": 3.6302436853232445,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 7.118518518518519e-06,
|
|
"loss": 0.6832,
|
|
"mean_token_accuracy": 0.8307051323354244,
|
|
"num_tokens": 54623090.0,
|
|
"step": 2040
|
|
},
|
|
{
|
|
"entropy": 0.5580741995945573,
|
|
"epoch": 3.6480471792589295,
|
|
"grad_norm": 1.4921875,
|
|
"learning_rate": 7.044444444444445e-06,
|
|
"loss": 0.6611,
|
|
"mean_token_accuracy": 0.8379048384726048,
|
|
"num_tokens": 54891961.0,
|
|
"step": 2050
|
|
},
|
|
{
|
|
"entropy": 0.5711165275424719,
|
|
"epoch": 3.6658506731946146,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 6.97037037037037e-06,
|
|
"loss": 0.6686,
|
|
"mean_token_accuracy": 0.83506432287395,
|
|
"num_tokens": 55160443.0,
|
|
"step": 2060
|
|
},
|
|
{
|
|
"entropy": 0.5829119691625237,
|
|
"epoch": 3.6836541671302996,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 6.896296296296297e-06,
|
|
"loss": 0.679,
|
|
"mean_token_accuracy": 0.8321164496243,
|
|
"num_tokens": 55425148.0,
|
|
"step": 2070
|
|
},
|
|
{
|
|
"entropy": 0.577633217908442,
|
|
"epoch": 3.701457661065984,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 6.8222222222222225e-06,
|
|
"loss": 0.6756,
|
|
"mean_token_accuracy": 0.831652557849884,
|
|
"num_tokens": 55701877.0,
|
|
"step": 2080
|
|
},
|
|
{
|
|
"entropy": 0.5851919570937753,
|
|
"epoch": 3.719261155001669,
|
|
"grad_norm": 1.4375,
|
|
"learning_rate": 6.748148148148149e-06,
|
|
"loss": 0.6871,
|
|
"mean_token_accuracy": 0.8311607643961907,
|
|
"num_tokens": 55969756.0,
|
|
"step": 2090
|
|
},
|
|
{
|
|
"entropy": 0.5728656422346831,
|
|
"epoch": 3.737064648937354,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 6.674074074074075e-06,
|
|
"loss": 0.6697,
|
|
"mean_token_accuracy": 0.8338924027979374,
|
|
"num_tokens": 56238576.0,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"epoch": 3.737064648937354,
|
|
"eval_biology_entropy": 1.3676475839614868,
|
|
"eval_biology_loss": 2.1407225131988525,
|
|
"eval_biology_mean_token_accuracy": 0.6122846744060516,
|
|
"eval_biology_num_tokens": 56238576.0,
|
|
"eval_biology_runtime": 39.0497,
|
|
"eval_biology_samples_per_second": 12.804,
|
|
"eval_biology_steps_per_second": 3.201,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"epoch": 3.737064648937354,
|
|
"eval_chemistry_entropy": 0.5876963279247284,
|
|
"eval_chemistry_loss": 0.8179986476898193,
|
|
"eval_chemistry_mean_token_accuracy": 0.8108155965805054,
|
|
"eval_chemistry_num_tokens": 56238576.0,
|
|
"eval_chemistry_runtime": 48.6816,
|
|
"eval_chemistry_samples_per_second": 10.271,
|
|
"eval_chemistry_steps_per_second": 2.568,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"epoch": 3.737064648937354,
|
|
"eval_math_entropy": 0.6119424107074738,
|
|
"eval_math_loss": 1.103035569190979,
|
|
"eval_math_mean_token_accuracy": 0.7718833708763122,
|
|
"eval_math_num_tokens": 56238576.0,
|
|
"eval_math_runtime": 50.0797,
|
|
"eval_math_samples_per_second": 9.984,
|
|
"eval_math_steps_per_second": 2.496,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"epoch": 3.737064648937354,
|
|
"eval_physics_entropy": 0.6151737835407257,
|
|
"eval_physics_loss": 0.8803703784942627,
|
|
"eval_physics_mean_token_accuracy": 0.7982921867370606,
|
|
"eval_physics_num_tokens": 56238576.0,
|
|
"eval_physics_runtime": 57.6104,
|
|
"eval_physics_samples_per_second": 8.679,
|
|
"eval_physics_steps_per_second": 2.17,
|
|
"step": 2100
|
|
},
|
|
{
|
|
"entropy": 0.575149905309081,
|
|
"epoch": 3.754868142873039,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 6.600000000000001e-06,
|
|
"loss": 0.6827,
|
|
"mean_token_accuracy": 0.8325412474572659,
|
|
"num_tokens": 56497905.0,
|
|
"step": 2110
|
|
},
|
|
{
|
|
"entropy": 0.5853163762018084,
|
|
"epoch": 3.772671636808724,
|
|
"grad_norm": 1.6796875,
|
|
"learning_rate": 6.525925925925927e-06,
|
|
"loss": 0.6848,
|
|
"mean_token_accuracy": 0.8329160042107105,
|
|
"num_tokens": 56757407.0,
|
|
"step": 2120
|
|
},
|
|
{
|
|
"entropy": 0.5622592825442553,
|
|
"epoch": 3.7904751307444084,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 6.4518518518518525e-06,
|
|
"loss": 0.666,
|
|
"mean_token_accuracy": 0.8372485417872667,
|
|
"num_tokens": 57019213.0,
|
|
"step": 2130
|
|
},
|
|
{
|
|
"entropy": 0.5716784641146659,
|
|
"epoch": 3.8082786246800935,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 6.377777777777778e-06,
|
|
"loss": 0.6785,
|
|
"mean_token_accuracy": 0.8330885298550129,
|
|
"num_tokens": 57274629.0,
|
|
"step": 2140
|
|
},
|
|
{
|
|
"entropy": 0.5762781355530023,
|
|
"epoch": 3.826082118615778,
|
|
"grad_norm": 1.4375,
|
|
"learning_rate": 6.303703703703704e-06,
|
|
"loss": 0.6635,
|
|
"mean_token_accuracy": 0.8338523168116808,
|
|
"num_tokens": 57549791.0,
|
|
"step": 2150
|
|
},
|
|
{
|
|
"entropy": 0.5586860096082091,
|
|
"epoch": 3.843885612551463,
|
|
"grad_norm": 1.5703125,
|
|
"learning_rate": 6.2296296296296295e-06,
|
|
"loss": 0.6637,
|
|
"mean_token_accuracy": 0.837235826253891,
|
|
"num_tokens": 57815454.0,
|
|
"step": 2160
|
|
},
|
|
{
|
|
"entropy": 0.5845056502148509,
|
|
"epoch": 3.861689106487148,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 6.155555555555556e-06,
|
|
"loss": 0.6887,
|
|
"mean_token_accuracy": 0.8297582499682903,
|
|
"num_tokens": 58076430.0,
|
|
"step": 2170
|
|
},
|
|
{
|
|
"entropy": 0.5741686699911952,
|
|
"epoch": 3.879492600422833,
|
|
"grad_norm": 1.609375,
|
|
"learning_rate": 6.081481481481482e-06,
|
|
"loss": 0.6757,
|
|
"mean_token_accuracy": 0.8332955792546273,
|
|
"num_tokens": 58346307.0,
|
|
"step": 2180
|
|
},
|
|
{
|
|
"entropy": 0.587618819437921,
|
|
"epoch": 3.8972960943585178,
|
|
"grad_norm": 1.4609375,
|
|
"learning_rate": 6.007407407407407e-06,
|
|
"loss": 0.6811,
|
|
"mean_token_accuracy": 0.8313716314733028,
|
|
"num_tokens": 58621624.0,
|
|
"step": 2190
|
|
},
|
|
{
|
|
"entropy": 0.5831481296569109,
|
|
"epoch": 3.915099588294203,
|
|
"grad_norm": 1.5,
|
|
"learning_rate": 5.933333333333335e-06,
|
|
"loss": 0.6864,
|
|
"mean_token_accuracy": 0.8321379527449608,
|
|
"num_tokens": 58874383.0,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"epoch": 3.915099588294203,
|
|
"eval_biology_entropy": 1.3810388627052308,
|
|
"eval_biology_loss": 2.1356613636016846,
|
|
"eval_biology_mean_token_accuracy": 0.6122371079921722,
|
|
"eval_biology_num_tokens": 58874383.0,
|
|
"eval_biology_runtime": 39.0687,
|
|
"eval_biology_samples_per_second": 12.798,
|
|
"eval_biology_steps_per_second": 3.199,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"epoch": 3.915099588294203,
|
|
"eval_chemistry_entropy": 0.5890640773773194,
|
|
"eval_chemistry_loss": 0.8181027770042419,
|
|
"eval_chemistry_mean_token_accuracy": 0.8106128220558166,
|
|
"eval_chemistry_num_tokens": 58874383.0,
|
|
"eval_chemistry_runtime": 48.6446,
|
|
"eval_chemistry_samples_per_second": 10.279,
|
|
"eval_chemistry_steps_per_second": 2.57,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"epoch": 3.915099588294203,
|
|
"eval_math_entropy": 0.6159807667732239,
|
|
"eval_math_loss": 1.1035534143447876,
|
|
"eval_math_mean_token_accuracy": 0.7718168897628784,
|
|
"eval_math_num_tokens": 58874383.0,
|
|
"eval_math_runtime": 50.0657,
|
|
"eval_math_samples_per_second": 9.987,
|
|
"eval_math_steps_per_second": 2.497,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"epoch": 3.915099588294203,
|
|
"eval_physics_entropy": 0.6170708584785461,
|
|
"eval_physics_loss": 0.8807304501533508,
|
|
"eval_physics_mean_token_accuracy": 0.7982254939079285,
|
|
"eval_physics_num_tokens": 58874383.0,
|
|
"eval_physics_runtime": 57.5593,
|
|
"eval_physics_samples_per_second": 8.687,
|
|
"eval_physics_steps_per_second": 2.172,
|
|
"step": 2200
|
|
},
|
|
{
|
|
"entropy": 0.5722213637083768,
|
|
"epoch": 3.9329030822298874,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 5.85925925925926e-06,
|
|
"loss": 0.669,
|
|
"mean_token_accuracy": 0.8334877517074346,
|
|
"num_tokens": 59149972.0,
|
|
"step": 2210
|
|
},
|
|
{
|
|
"entropy": 0.5695188788697123,
|
|
"epoch": 3.9507065761655724,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 5.785185185185186e-06,
|
|
"loss": 0.6716,
|
|
"mean_token_accuracy": 0.8339714191854,
|
|
"num_tokens": 59418591.0,
|
|
"step": 2220
|
|
},
|
|
{
|
|
"entropy": 0.58336979188025,
|
|
"epoch": 3.9685100701012574,
|
|
"grad_norm": 1.609375,
|
|
"learning_rate": 5.711111111111112e-06,
|
|
"loss": 0.6787,
|
|
"mean_token_accuracy": 0.8313135091215372,
|
|
"num_tokens": 59691322.0,
|
|
"step": 2230
|
|
},
|
|
{
|
|
"entropy": 0.5696559147909284,
|
|
"epoch": 3.9863135640369425,
|
|
"grad_norm": 1.625,
|
|
"learning_rate": 5.637037037037037e-06,
|
|
"loss": 0.674,
|
|
"mean_token_accuracy": 0.835492142662406,
|
|
"num_tokens": 59953838.0,
|
|
"step": 2240
|
|
},
|
|
{
|
|
"entropy": 0.5570171371583016,
|
|
"epoch": 4.003560698787137,
|
|
"grad_norm": 1.34375,
|
|
"learning_rate": 5.562962962962963e-06,
|
|
"loss": 0.6546,
|
|
"mean_token_accuracy": 0.8391922777698886,
|
|
"num_tokens": 60216049.0,
|
|
"step": 2250
|
|
},
|
|
{
|
|
"entropy": 0.5453685067594052,
|
|
"epoch": 4.021364192722822,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 5.4888888888888895e-06,
|
|
"loss": 0.6385,
|
|
"mean_token_accuracy": 0.8419853840023279,
|
|
"num_tokens": 60490937.0,
|
|
"step": 2260
|
|
},
|
|
{
|
|
"entropy": 0.553492389805615,
|
|
"epoch": 4.039167686658507,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 5.414814814814815e-06,
|
|
"loss": 0.6556,
|
|
"mean_token_accuracy": 0.8393857058137655,
|
|
"num_tokens": 60755892.0,
|
|
"step": 2270
|
|
},
|
|
{
|
|
"entropy": 0.5440751701593399,
|
|
"epoch": 4.056971180594192,
|
|
"grad_norm": 1.4765625,
|
|
"learning_rate": 5.340740740740741e-06,
|
|
"loss": 0.6398,
|
|
"mean_token_accuracy": 0.8416271641850471,
|
|
"num_tokens": 61030893.0,
|
|
"step": 2280
|
|
},
|
|
{
|
|
"entropy": 0.5338836884126067,
|
|
"epoch": 4.074774674529876,
|
|
"grad_norm": 1.4453125,
|
|
"learning_rate": 5.2666666666666665e-06,
|
|
"loss": 0.6349,
|
|
"mean_token_accuracy": 0.844103604927659,
|
|
"num_tokens": 61301458.0,
|
|
"step": 2290
|
|
},
|
|
{
|
|
"entropy": 0.5652395199984313,
|
|
"epoch": 4.092578168465561,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 5.192592592592594e-06,
|
|
"loss": 0.6572,
|
|
"mean_token_accuracy": 0.8368313021957874,
|
|
"num_tokens": 61574717.0,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"epoch": 4.092578168465561,
|
|
"eval_biology_entropy": 1.3518374223709106,
|
|
"eval_biology_loss": 2.1629726886749268,
|
|
"eval_biology_mean_token_accuracy": 0.6107748429775238,
|
|
"eval_biology_num_tokens": 61574717.0,
|
|
"eval_biology_runtime": 39.0562,
|
|
"eval_biology_samples_per_second": 12.802,
|
|
"eval_biology_steps_per_second": 3.201,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"epoch": 4.092578168465561,
|
|
"eval_chemistry_entropy": 0.5780029859542847,
|
|
"eval_chemistry_loss": 0.8214347958564758,
|
|
"eval_chemistry_mean_token_accuracy": 0.810426944732666,
|
|
"eval_chemistry_num_tokens": 61574717.0,
|
|
"eval_chemistry_runtime": 48.6785,
|
|
"eval_chemistry_samples_per_second": 10.271,
|
|
"eval_chemistry_steps_per_second": 2.568,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"epoch": 4.092578168465561,
|
|
"eval_math_entropy": 0.6076318190097809,
|
|
"eval_math_loss": 1.1111681461334229,
|
|
"eval_math_mean_token_accuracy": 0.7712901134490967,
|
|
"eval_math_num_tokens": 61574717.0,
|
|
"eval_math_runtime": 50.0691,
|
|
"eval_math_samples_per_second": 9.986,
|
|
"eval_math_steps_per_second": 2.497,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"epoch": 4.092578168465561,
|
|
"eval_physics_entropy": 0.6077080953121186,
|
|
"eval_physics_loss": 0.8865928053855896,
|
|
"eval_physics_mean_token_accuracy": 0.7976819982528687,
|
|
"eval_physics_num_tokens": 61574717.0,
|
|
"eval_physics_runtime": 57.5982,
|
|
"eval_physics_samples_per_second": 8.681,
|
|
"eval_physics_steps_per_second": 2.17,
|
|
"step": 2300
|
|
},
|
|
{
|
|
"entropy": 0.5665840411558747,
|
|
"epoch": 4.110381662401246,
|
|
"grad_norm": 1.5859375,
|
|
"learning_rate": 5.1185185185185195e-06,
|
|
"loss": 0.6708,
|
|
"mean_token_accuracy": 0.8358031015843153,
|
|
"num_tokens": 61840186.0,
|
|
"step": 2310
|
|
},
|
|
{
|
|
"entropy": 0.56264589689672,
|
|
"epoch": 4.128185156336931,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 5.044444444444445e-06,
|
|
"loss": 0.6631,
|
|
"mean_token_accuracy": 0.836660572513938,
|
|
"num_tokens": 62112844.0,
|
|
"step": 2320
|
|
},
|
|
{
|
|
"entropy": 0.5638300092890859,
|
|
"epoch": 4.145988650272616,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 4.970370370370371e-06,
|
|
"loss": 0.6511,
|
|
"mean_token_accuracy": 0.8376240566372871,
|
|
"num_tokens": 62382390.0,
|
|
"step": 2330
|
|
},
|
|
{
|
|
"entropy": 0.5563945654779673,
|
|
"epoch": 4.163792144208301,
|
|
"grad_norm": 1.3203125,
|
|
"learning_rate": 4.8962962962962965e-06,
|
|
"loss": 0.6578,
|
|
"mean_token_accuracy": 0.8388708371669054,
|
|
"num_tokens": 62658026.0,
|
|
"step": 2340
|
|
},
|
|
{
|
|
"entropy": 0.5789911892265082,
|
|
"epoch": 4.181595638143985,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 4.822222222222222e-06,
|
|
"loss": 0.6856,
|
|
"mean_token_accuracy": 0.8329568482935429,
|
|
"num_tokens": 62916724.0,
|
|
"step": 2350
|
|
},
|
|
{
|
|
"entropy": 0.555764663591981,
|
|
"epoch": 4.19939913207967,
|
|
"grad_norm": 1.484375,
|
|
"learning_rate": 4.748148148148149e-06,
|
|
"loss": 0.6465,
|
|
"mean_token_accuracy": 0.8380315884947777,
|
|
"num_tokens": 63195271.0,
|
|
"step": 2360
|
|
},
|
|
{
|
|
"entropy": 0.5678263012319803,
|
|
"epoch": 4.2172026260153554,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 4.674074074074074e-06,
|
|
"loss": 0.6671,
|
|
"mean_token_accuracy": 0.8358955714851618,
|
|
"num_tokens": 63466066.0,
|
|
"step": 2370
|
|
},
|
|
{
|
|
"entropy": 0.554285722412169,
|
|
"epoch": 4.2350061199510405,
|
|
"grad_norm": 1.4921875,
|
|
"learning_rate": 4.600000000000001e-06,
|
|
"loss": 0.6573,
|
|
"mean_token_accuracy": 0.838936285302043,
|
|
"num_tokens": 63739678.0,
|
|
"step": 2380
|
|
},
|
|
{
|
|
"entropy": 0.5789531769230962,
|
|
"epoch": 4.2528096138867255,
|
|
"grad_norm": 1.4765625,
|
|
"learning_rate": 4.5259259259259265e-06,
|
|
"loss": 0.6723,
|
|
"mean_token_accuracy": 0.833799697458744,
|
|
"num_tokens": 63999364.0,
|
|
"step": 2390
|
|
},
|
|
{
|
|
"entropy": 0.5556379873305559,
|
|
"epoch": 4.2706131078224105,
|
|
"grad_norm": 1.4609375,
|
|
"learning_rate": 4.451851851851852e-06,
|
|
"loss": 0.6557,
|
|
"mean_token_accuracy": 0.8391667053103447,
|
|
"num_tokens": 64270501.0,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"epoch": 4.2706131078224105,
|
|
"eval_biology_entropy": 1.3456227674484253,
|
|
"eval_biology_loss": 2.162630558013916,
|
|
"eval_biology_mean_token_accuracy": 0.6108367409706116,
|
|
"eval_biology_num_tokens": 64270501.0,
|
|
"eval_biology_runtime": 39.0723,
|
|
"eval_biology_samples_per_second": 12.797,
|
|
"eval_biology_steps_per_second": 3.199,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"epoch": 4.2706131078224105,
|
|
"eval_chemistry_entropy": 0.5770703101158142,
|
|
"eval_chemistry_loss": 0.8213785290718079,
|
|
"eval_chemistry_mean_token_accuracy": 0.8106174335479737,
|
|
"eval_chemistry_num_tokens": 64270501.0,
|
|
"eval_chemistry_runtime": 48.65,
|
|
"eval_chemistry_samples_per_second": 10.278,
|
|
"eval_chemistry_steps_per_second": 2.569,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"epoch": 4.2706131078224105,
|
|
"eval_math_entropy": 0.6059090893268585,
|
|
"eval_math_loss": 1.110784888267517,
|
|
"eval_math_mean_token_accuracy": 0.771416862487793,
|
|
"eval_math_num_tokens": 64270501.0,
|
|
"eval_math_runtime": 50.0542,
|
|
"eval_math_samples_per_second": 9.989,
|
|
"eval_math_steps_per_second": 2.497,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"epoch": 4.2706131078224105,
|
|
"eval_physics_entropy": 0.6061555581092835,
|
|
"eval_physics_loss": 0.8872570991516113,
|
|
"eval_physics_mean_token_accuracy": 0.797800799369812,
|
|
"eval_physics_num_tokens": 64270501.0,
|
|
"eval_physics_runtime": 57.5677,
|
|
"eval_physics_samples_per_second": 8.685,
|
|
"eval_physics_steps_per_second": 2.171,
|
|
"step": 2400
|
|
},
|
|
{
|
|
"entropy": 0.5603191582486033,
|
|
"epoch": 4.288416601758095,
|
|
"grad_norm": 1.640625,
|
|
"learning_rate": 4.377777777777778e-06,
|
|
"loss": 0.6653,
|
|
"mean_token_accuracy": 0.8375619545578956,
|
|
"num_tokens": 64533223.0,
|
|
"step": 2410
|
|
},
|
|
{
|
|
"entropy": 0.5656454924494028,
|
|
"epoch": 4.30622009569378,
|
|
"grad_norm": 1.515625,
|
|
"learning_rate": 4.3037037037037035e-06,
|
|
"loss": 0.659,
|
|
"mean_token_accuracy": 0.8366425581276417,
|
|
"num_tokens": 64799113.0,
|
|
"step": 2420
|
|
},
|
|
{
|
|
"entropy": 0.5550693856552243,
|
|
"epoch": 4.324023589629465,
|
|
"grad_norm": 1.625,
|
|
"learning_rate": 4.22962962962963e-06,
|
|
"loss": 0.6538,
|
|
"mean_token_accuracy": 0.8389651108533144,
|
|
"num_tokens": 65075459.0,
|
|
"step": 2430
|
|
},
|
|
{
|
|
"entropy": 0.5567959282547236,
|
|
"epoch": 4.34182708356515,
|
|
"grad_norm": 1.640625,
|
|
"learning_rate": 4.155555555555556e-06,
|
|
"loss": 0.66,
|
|
"mean_token_accuracy": 0.837913642078638,
|
|
"num_tokens": 65341896.0,
|
|
"step": 2440
|
|
},
|
|
{
|
|
"entropy": 0.5572125539183617,
|
|
"epoch": 4.359630577500835,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 4.081481481481482e-06,
|
|
"loss": 0.6482,
|
|
"mean_token_accuracy": 0.837642902880907,
|
|
"num_tokens": 65614487.0,
|
|
"step": 2450
|
|
},
|
|
{
|
|
"entropy": 0.5893890267238021,
|
|
"epoch": 4.37743407143652,
|
|
"grad_norm": 1.4375,
|
|
"learning_rate": 4.007407407407408e-06,
|
|
"loss": 0.6784,
|
|
"mean_token_accuracy": 0.830023642629385,
|
|
"num_tokens": 65887311.0,
|
|
"step": 2460
|
|
},
|
|
{
|
|
"entropy": 0.5546840105205775,
|
|
"epoch": 4.395237565372204,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 3.9333333333333335e-06,
|
|
"loss": 0.6537,
|
|
"mean_token_accuracy": 0.8394486285746098,
|
|
"num_tokens": 66151098.0,
|
|
"step": 2470
|
|
},
|
|
{
|
|
"entropy": 0.5494891541078687,
|
|
"epoch": 4.413041059307889,
|
|
"grad_norm": 1.5703125,
|
|
"learning_rate": 3.85925925925926e-06,
|
|
"loss": 0.6481,
|
|
"mean_token_accuracy": 0.8396681498736143,
|
|
"num_tokens": 66417928.0,
|
|
"step": 2480
|
|
},
|
|
{
|
|
"entropy": 0.5561942363157868,
|
|
"epoch": 4.430844553243574,
|
|
"grad_norm": 1.796875,
|
|
"learning_rate": 3.7851851851851857e-06,
|
|
"loss": 0.6614,
|
|
"mean_token_accuracy": 0.8361107159405947,
|
|
"num_tokens": 66684481.0,
|
|
"step": 2490
|
|
},
|
|
{
|
|
"entropy": 0.5738992135971784,
|
|
"epoch": 4.448648047179259,
|
|
"grad_norm": 1.7109375,
|
|
"learning_rate": 3.7111111111111113e-06,
|
|
"loss": 0.6772,
|
|
"mean_token_accuracy": 0.833794866502285,
|
|
"num_tokens": 66940215.0,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 4.448648047179259,
|
|
"eval_biology_entropy": 1.3500232787132262,
|
|
"eval_biology_loss": 2.1630518436431885,
|
|
"eval_biology_mean_token_accuracy": 0.6106264660358429,
|
|
"eval_biology_num_tokens": 66940215.0,
|
|
"eval_biology_runtime": 39.0451,
|
|
"eval_biology_samples_per_second": 12.806,
|
|
"eval_biology_steps_per_second": 3.201,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 4.448648047179259,
|
|
"eval_chemistry_entropy": 0.5779505209922791,
|
|
"eval_chemistry_loss": 0.8212056756019592,
|
|
"eval_chemistry_mean_token_accuracy": 0.8103259291648864,
|
|
"eval_chemistry_num_tokens": 66940215.0,
|
|
"eval_chemistry_runtime": 48.6513,
|
|
"eval_chemistry_samples_per_second": 10.277,
|
|
"eval_chemistry_steps_per_second": 2.569,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 4.448648047179259,
|
|
"eval_math_entropy": 0.6078341612815857,
|
|
"eval_math_loss": 1.111467719078064,
|
|
"eval_math_mean_token_accuracy": 0.771428505897522,
|
|
"eval_math_num_tokens": 66940215.0,
|
|
"eval_math_runtime": 50.181,
|
|
"eval_math_samples_per_second": 9.964,
|
|
"eval_math_steps_per_second": 2.491,
|
|
"step": 2500
|
|
},
|
|
{
|
|
"epoch": 4.448648047179259,
|
|
"eval_physics_entropy": 0.6069678666591645,
|
|
"eval_physics_loss": 0.8872233629226685,
|
|
"eval_physics_mean_token_accuracy": 0.7977854557037354,
|
|
"eval_physics_num_tokens": 66940215.0,
|
|
"eval_physics_runtime": 57.617,
|
|
"eval_physics_samples_per_second": 8.678,
|
|
"eval_physics_steps_per_second": 2.169,
|
|
"step": 2500
|
|
}
|
|
],
|
|
"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": 5.20654490802918e+18,
|
|
"train_batch_size": 4,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|