2915 lines
84 KiB
JSON
2915 lines
84 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 1.6,
|
|
"eval_steps": 100,
|
|
"global_step": 2000,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"entropy": 0.6430254079401493,
|
|
"epoch": 0.008,
|
|
"grad_norm": 4.53125,
|
|
"learning_rate": 6.000000000000001e-07,
|
|
"loss": 1.8435,
|
|
"mean_token_accuracy": 0.6683766476809978,
|
|
"num_tokens": 133939.0,
|
|
"step": 10
|
|
},
|
|
{
|
|
"entropy": 0.6358832061290741,
|
|
"epoch": 0.016,
|
|
"grad_norm": 4.3125,
|
|
"learning_rate": 1.2666666666666669e-06,
|
|
"loss": 1.8392,
|
|
"mean_token_accuracy": 0.6651193369179964,
|
|
"num_tokens": 273600.0,
|
|
"step": 20
|
|
},
|
|
{
|
|
"entropy": 0.6754261270165444,
|
|
"epoch": 0.024,
|
|
"grad_norm": 4.9375,
|
|
"learning_rate": 1.9333333333333336e-06,
|
|
"loss": 1.9101,
|
|
"mean_token_accuracy": 0.6521823685616255,
|
|
"num_tokens": 408034.0,
|
|
"step": 30
|
|
},
|
|
{
|
|
"entropy": 0.6558344909921289,
|
|
"epoch": 0.032,
|
|
"grad_norm": 4.28125,
|
|
"learning_rate": 2.6e-06,
|
|
"loss": 1.8472,
|
|
"mean_token_accuracy": 0.6629668604582548,
|
|
"num_tokens": 542836.0,
|
|
"step": 40
|
|
},
|
|
{
|
|
"entropy": 0.6494104014709592,
|
|
"epoch": 0.04,
|
|
"grad_norm": 3.8125,
|
|
"learning_rate": 3.266666666666667e-06,
|
|
"loss": 1.8027,
|
|
"mean_token_accuracy": 0.6676773574203253,
|
|
"num_tokens": 678419.0,
|
|
"step": 50
|
|
},
|
|
{
|
|
"entropy": 0.662838158942759,
|
|
"epoch": 0.048,
|
|
"grad_norm": 4.03125,
|
|
"learning_rate": 3.9333333333333335e-06,
|
|
"loss": 1.8213,
|
|
"mean_token_accuracy": 0.6647731238976121,
|
|
"num_tokens": 813510.0,
|
|
"step": 60
|
|
},
|
|
{
|
|
"entropy": 0.6506876813247799,
|
|
"epoch": 0.056,
|
|
"grad_norm": 3.71875,
|
|
"learning_rate": 4.600000000000001e-06,
|
|
"loss": 1.7955,
|
|
"mean_token_accuracy": 0.6665123915299773,
|
|
"num_tokens": 954682.0,
|
|
"step": 70
|
|
},
|
|
{
|
|
"entropy": 0.6252049600705505,
|
|
"epoch": 0.064,
|
|
"grad_norm": 3.65625,
|
|
"learning_rate": 5.2666666666666665e-06,
|
|
"loss": 1.7861,
|
|
"mean_token_accuracy": 0.6728611093014478,
|
|
"num_tokens": 1096071.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"entropy": 0.6236795110628008,
|
|
"epoch": 0.072,
|
|
"grad_norm": 3.171875,
|
|
"learning_rate": 5.933333333333335e-06,
|
|
"loss": 1.6458,
|
|
"mean_token_accuracy": 0.6831523377448321,
|
|
"num_tokens": 1239910.0,
|
|
"step": 90
|
|
},
|
|
{
|
|
"entropy": 0.6565364092588425,
|
|
"epoch": 0.08,
|
|
"grad_norm": 3.125,
|
|
"learning_rate": 6.600000000000001e-06,
|
|
"loss": 1.6948,
|
|
"mean_token_accuracy": 0.6728412076830864,
|
|
"num_tokens": 1376045.0,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"eval_biology_entropy": 1.821127786874771,
|
|
"eval_biology_loss": 4.8697123527526855,
|
|
"eval_biology_mean_token_accuracy": 0.3407035154700279,
|
|
"eval_biology_num_tokens": 1376045.0,
|
|
"eval_biology_runtime": 26.9612,
|
|
"eval_biology_samples_per_second": 18.545,
|
|
"eval_biology_steps_per_second": 9.273,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"eval_chemistry_entropy": 1.1214371861219405,
|
|
"eval_chemistry_loss": 2.9583740234375,
|
|
"eval_chemistry_mean_token_accuracy": 0.5230503298044205,
|
|
"eval_chemistry_num_tokens": 1376045.0,
|
|
"eval_chemistry_runtime": 34.4743,
|
|
"eval_chemistry_samples_per_second": 14.504,
|
|
"eval_chemistry_steps_per_second": 7.252,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"eval_math_entropy": 0.6455623766183853,
|
|
"eval_math_loss": 1.6610009670257568,
|
|
"eval_math_mean_token_accuracy": 0.6816041300296783,
|
|
"eval_math_num_tokens": 1376045.0,
|
|
"eval_math_runtime": 34.9863,
|
|
"eval_math_samples_per_second": 14.291,
|
|
"eval_math_steps_per_second": 7.146,
|
|
"step": 100
|
|
},
|
|
{
|
|
"epoch": 0.08,
|
|
"eval_physics_entropy": 0.8894976880550385,
|
|
"eval_physics_loss": 2.340852975845337,
|
|
"eval_physics_mean_token_accuracy": 0.5970065190792083,
|
|
"eval_physics_num_tokens": 1376045.0,
|
|
"eval_physics_runtime": 40.5871,
|
|
"eval_physics_samples_per_second": 12.319,
|
|
"eval_physics_steps_per_second": 6.16,
|
|
"step": 100
|
|
},
|
|
{
|
|
"entropy": 0.6108825646340847,
|
|
"epoch": 0.088,
|
|
"grad_norm": 3.0625,
|
|
"learning_rate": 7.266666666666668e-06,
|
|
"loss": 1.534,
|
|
"mean_token_accuracy": 0.6979989364743233,
|
|
"num_tokens": 1518257.0,
|
|
"step": 110
|
|
},
|
|
{
|
|
"entropy": 0.6626681972295045,
|
|
"epoch": 0.096,
|
|
"grad_norm": 3.0,
|
|
"learning_rate": 7.933333333333334e-06,
|
|
"loss": 1.4858,
|
|
"mean_token_accuracy": 0.6921830836683511,
|
|
"num_tokens": 1656215.0,
|
|
"step": 120
|
|
},
|
|
{
|
|
"entropy": 0.6807309044525027,
|
|
"epoch": 0.104,
|
|
"grad_norm": 2.4375,
|
|
"learning_rate": 8.6e-06,
|
|
"loss": 1.3897,
|
|
"mean_token_accuracy": 0.7023331839591265,
|
|
"num_tokens": 1791507.0,
|
|
"step": 130
|
|
},
|
|
{
|
|
"entropy": 0.7359636599197984,
|
|
"epoch": 0.112,
|
|
"grad_norm": 1.9765625,
|
|
"learning_rate": 9.266666666666667e-06,
|
|
"loss": 1.3356,
|
|
"mean_token_accuracy": 0.7036307867616415,
|
|
"num_tokens": 1928153.0,
|
|
"step": 140
|
|
},
|
|
{
|
|
"entropy": 0.8070579996332526,
|
|
"epoch": 0.12,
|
|
"grad_norm": 1.7265625,
|
|
"learning_rate": 9.933333333333334e-06,
|
|
"loss": 1.2682,
|
|
"mean_token_accuracy": 0.712115453183651,
|
|
"num_tokens": 2071008.0,
|
|
"step": 150
|
|
},
|
|
{
|
|
"entropy": 0.874446533434093,
|
|
"epoch": 0.128,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 1.0600000000000002e-05,
|
|
"loss": 1.1443,
|
|
"mean_token_accuracy": 0.7246435590088367,
|
|
"num_tokens": 2216042.0,
|
|
"step": 160
|
|
},
|
|
{
|
|
"entropy": 0.9639016635715961,
|
|
"epoch": 0.136,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 1.1266666666666668e-05,
|
|
"loss": 1.1005,
|
|
"mean_token_accuracy": 0.730505557358265,
|
|
"num_tokens": 2350589.0,
|
|
"step": 170
|
|
},
|
|
{
|
|
"entropy": 1.0533128878101707,
|
|
"epoch": 0.144,
|
|
"grad_norm": 0.90234375,
|
|
"learning_rate": 1.1933333333333335e-05,
|
|
"loss": 1.0494,
|
|
"mean_token_accuracy": 0.7364416709169745,
|
|
"num_tokens": 2483116.0,
|
|
"step": 180
|
|
},
|
|
{
|
|
"entropy": 1.0085792429745197,
|
|
"epoch": 0.152,
|
|
"grad_norm": 0.96484375,
|
|
"learning_rate": 1.2600000000000001e-05,
|
|
"loss": 0.9875,
|
|
"mean_token_accuracy": 0.749194710329175,
|
|
"num_tokens": 2617958.0,
|
|
"step": 190
|
|
},
|
|
{
|
|
"entropy": 1.000020370259881,
|
|
"epoch": 0.16,
|
|
"grad_norm": 0.90234375,
|
|
"learning_rate": 1.3266666666666668e-05,
|
|
"loss": 0.9952,
|
|
"mean_token_accuracy": 0.7486346989870072,
|
|
"num_tokens": 2752258.0,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"eval_biology_entropy": 3.2312401638031005,
|
|
"eval_biology_loss": 3.4984476566314697,
|
|
"eval_biology_mean_token_accuracy": 0.4106140378713608,
|
|
"eval_biology_num_tokens": 2752258.0,
|
|
"eval_biology_runtime": 26.9494,
|
|
"eval_biology_samples_per_second": 18.553,
|
|
"eval_biology_steps_per_second": 9.277,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"eval_chemistry_entropy": 1.9263038992881776,
|
|
"eval_chemistry_loss": 1.9608160257339478,
|
|
"eval_chemistry_mean_token_accuracy": 0.5982974277734756,
|
|
"eval_chemistry_num_tokens": 2752258.0,
|
|
"eval_chemistry_runtime": 34.4173,
|
|
"eval_chemistry_samples_per_second": 14.528,
|
|
"eval_chemistry_steps_per_second": 7.264,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"eval_math_entropy": 1.0257348996400832,
|
|
"eval_math_loss": 1.029754638671875,
|
|
"eval_math_mean_token_accuracy": 0.7436505243778229,
|
|
"eval_math_num_tokens": 2752258.0,
|
|
"eval_math_runtime": 34.9579,
|
|
"eval_math_samples_per_second": 14.303,
|
|
"eval_math_steps_per_second": 7.151,
|
|
"step": 200
|
|
},
|
|
{
|
|
"epoch": 0.16,
|
|
"eval_physics_entropy": 1.5005105144977569,
|
|
"eval_physics_loss": 1.5203113555908203,
|
|
"eval_physics_mean_token_accuracy": 0.6669992462396622,
|
|
"eval_physics_num_tokens": 2752258.0,
|
|
"eval_physics_runtime": 40.5453,
|
|
"eval_physics_samples_per_second": 12.332,
|
|
"eval_physics_steps_per_second": 6.166,
|
|
"step": 200
|
|
},
|
|
{
|
|
"entropy": 1.0170325668528677,
|
|
"epoch": 0.168,
|
|
"grad_norm": 0.80859375,
|
|
"learning_rate": 1.3933333333333334e-05,
|
|
"loss": 1.0032,
|
|
"mean_token_accuracy": 0.7469347029924392,
|
|
"num_tokens": 2889249.0,
|
|
"step": 210
|
|
},
|
|
{
|
|
"entropy": 0.9709451261907815,
|
|
"epoch": 0.176,
|
|
"grad_norm": 0.7734375,
|
|
"learning_rate": 1.46e-05,
|
|
"loss": 0.9299,
|
|
"mean_token_accuracy": 0.7556545663625002,
|
|
"num_tokens": 3030305.0,
|
|
"step": 220
|
|
},
|
|
{
|
|
"entropy": 1.0022083457559348,
|
|
"epoch": 0.184,
|
|
"grad_norm": 0.90234375,
|
|
"learning_rate": 1.5266666666666667e-05,
|
|
"loss": 0.9815,
|
|
"mean_token_accuracy": 0.7520246226340532,
|
|
"num_tokens": 3167096.0,
|
|
"step": 230
|
|
},
|
|
{
|
|
"entropy": 0.9930278725922108,
|
|
"epoch": 0.192,
|
|
"grad_norm": 0.85546875,
|
|
"learning_rate": 1.5933333333333336e-05,
|
|
"loss": 0.9637,
|
|
"mean_token_accuracy": 0.7547307416796685,
|
|
"num_tokens": 3306293.0,
|
|
"step": 240
|
|
},
|
|
{
|
|
"entropy": 0.9768950846046209,
|
|
"epoch": 0.2,
|
|
"grad_norm": 0.7734375,
|
|
"learning_rate": 1.66e-05,
|
|
"loss": 0.9373,
|
|
"mean_token_accuracy": 0.7594478782266378,
|
|
"num_tokens": 3441895.0,
|
|
"step": 250
|
|
},
|
|
{
|
|
"entropy": 0.9833385257050395,
|
|
"epoch": 0.208,
|
|
"grad_norm": 0.8359375,
|
|
"learning_rate": 1.726666666666667e-05,
|
|
"loss": 0.9384,
|
|
"mean_token_accuracy": 0.7573206882923842,
|
|
"num_tokens": 3578349.0,
|
|
"step": 260
|
|
},
|
|
{
|
|
"entropy": 0.9361576547846198,
|
|
"epoch": 0.216,
|
|
"grad_norm": 0.74609375,
|
|
"learning_rate": 1.7933333333333333e-05,
|
|
"loss": 0.9004,
|
|
"mean_token_accuracy": 0.7651511203497648,
|
|
"num_tokens": 3718484.0,
|
|
"step": 270
|
|
},
|
|
{
|
|
"entropy": 0.9637048680335283,
|
|
"epoch": 0.224,
|
|
"grad_norm": 0.9140625,
|
|
"learning_rate": 1.86e-05,
|
|
"loss": 0.9349,
|
|
"mean_token_accuracy": 0.7605184197425843,
|
|
"num_tokens": 3851369.0,
|
|
"step": 280
|
|
},
|
|
{
|
|
"entropy": 0.9367688067257405,
|
|
"epoch": 0.232,
|
|
"grad_norm": 0.74609375,
|
|
"learning_rate": 1.926666666666667e-05,
|
|
"loss": 0.8962,
|
|
"mean_token_accuracy": 0.7659803122282028,
|
|
"num_tokens": 3992977.0,
|
|
"step": 290
|
|
},
|
|
{
|
|
"entropy": 0.9229113388806581,
|
|
"epoch": 0.24,
|
|
"grad_norm": 0.7890625,
|
|
"learning_rate": 1.9933333333333334e-05,
|
|
"loss": 0.9024,
|
|
"mean_token_accuracy": 0.77176120467484,
|
|
"num_tokens": 4133723.0,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.24,
|
|
"eval_biology_entropy": 3.018038348674774,
|
|
"eval_biology_loss": 3.2969858646392822,
|
|
"eval_biology_mean_token_accuracy": 0.43762873542308806,
|
|
"eval_biology_num_tokens": 4133723.0,
|
|
"eval_biology_runtime": 27.1593,
|
|
"eval_biology_samples_per_second": 18.41,
|
|
"eval_biology_steps_per_second": 9.205,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.24,
|
|
"eval_chemistry_entropy": 1.778470279932022,
|
|
"eval_chemistry_loss": 1.8064701557159424,
|
|
"eval_chemistry_mean_token_accuracy": 0.6220316555500031,
|
|
"eval_chemistry_num_tokens": 4133723.0,
|
|
"eval_chemistry_runtime": 34.6422,
|
|
"eval_chemistry_samples_per_second": 14.433,
|
|
"eval_chemistry_steps_per_second": 7.217,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.24,
|
|
"eval_math_entropy": 0.9541566734313964,
|
|
"eval_math_loss": 0.9326547980308533,
|
|
"eval_math_mean_token_accuracy": 0.7629979259967804,
|
|
"eval_math_num_tokens": 4133723.0,
|
|
"eval_math_runtime": 35.1904,
|
|
"eval_math_samples_per_second": 14.208,
|
|
"eval_math_steps_per_second": 7.104,
|
|
"step": 300
|
|
},
|
|
{
|
|
"epoch": 0.24,
|
|
"eval_physics_entropy": 1.387200428724289,
|
|
"eval_physics_loss": 1.3932181596755981,
|
|
"eval_physics_mean_token_accuracy": 0.6886798795461655,
|
|
"eval_physics_num_tokens": 4133723.0,
|
|
"eval_physics_runtime": 40.7999,
|
|
"eval_physics_samples_per_second": 12.255,
|
|
"eval_physics_steps_per_second": 6.127,
|
|
"step": 300
|
|
},
|
|
{
|
|
"entropy": 0.9078424336388707,
|
|
"epoch": 0.248,
|
|
"grad_norm": 0.75,
|
|
"learning_rate": 1.9933333333333334e-05,
|
|
"loss": 0.8542,
|
|
"mean_token_accuracy": 0.7745326109230518,
|
|
"num_tokens": 4273949.0,
|
|
"step": 310
|
|
},
|
|
{
|
|
"entropy": 0.9262684918940067,
|
|
"epoch": 0.256,
|
|
"grad_norm": 0.84765625,
|
|
"learning_rate": 1.985925925925926e-05,
|
|
"loss": 0.8761,
|
|
"mean_token_accuracy": 0.7701829686760903,
|
|
"num_tokens": 4414102.0,
|
|
"step": 320
|
|
},
|
|
{
|
|
"entropy": 0.9303955681622028,
|
|
"epoch": 0.264,
|
|
"grad_norm": 0.78515625,
|
|
"learning_rate": 1.9785185185185187e-05,
|
|
"loss": 0.8765,
|
|
"mean_token_accuracy": 0.7714239094406367,
|
|
"num_tokens": 4548402.0,
|
|
"step": 330
|
|
},
|
|
{
|
|
"entropy": 0.9207985673099757,
|
|
"epoch": 0.272,
|
|
"grad_norm": 0.7109375,
|
|
"learning_rate": 1.971111111111111e-05,
|
|
"loss": 0.8728,
|
|
"mean_token_accuracy": 0.7721738632768392,
|
|
"num_tokens": 4685739.0,
|
|
"step": 340
|
|
},
|
|
{
|
|
"entropy": 0.9172879874706268,
|
|
"epoch": 0.28,
|
|
"grad_norm": 0.80078125,
|
|
"learning_rate": 1.963703703703704e-05,
|
|
"loss": 0.869,
|
|
"mean_token_accuracy": 0.7719811659306288,
|
|
"num_tokens": 4827264.0,
|
|
"step": 350
|
|
},
|
|
{
|
|
"entropy": 0.8906141923740506,
|
|
"epoch": 0.288,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 1.9562962962962964e-05,
|
|
"loss": 0.8676,
|
|
"mean_token_accuracy": 0.779513431340456,
|
|
"num_tokens": 4963900.0,
|
|
"step": 360
|
|
},
|
|
{
|
|
"entropy": 0.9126916229724884,
|
|
"epoch": 0.296,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.948888888888889e-05,
|
|
"loss": 0.8727,
|
|
"mean_token_accuracy": 0.7722058985382318,
|
|
"num_tokens": 5106308.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"entropy": 0.9109785668551922,
|
|
"epoch": 0.304,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 1.9414814814814817e-05,
|
|
"loss": 0.8562,
|
|
"mean_token_accuracy": 0.7752636637538671,
|
|
"num_tokens": 5246290.0,
|
|
"step": 380
|
|
},
|
|
{
|
|
"entropy": 0.9183575309813022,
|
|
"epoch": 0.312,
|
|
"grad_norm": 0.71875,
|
|
"learning_rate": 1.9340740740740743e-05,
|
|
"loss": 0.8621,
|
|
"mean_token_accuracy": 0.7713671930134296,
|
|
"num_tokens": 5389824.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"entropy": 0.8591190708801151,
|
|
"epoch": 0.32,
|
|
"grad_norm": 0.82421875,
|
|
"learning_rate": 1.926666666666667e-05,
|
|
"loss": 0.8065,
|
|
"mean_token_accuracy": 0.7856244947761297,
|
|
"num_tokens": 5531902.0,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"eval_biology_entropy": 2.974372909069061,
|
|
"eval_biology_loss": 3.2137467861175537,
|
|
"eval_biology_mean_token_accuracy": 0.44984077060222627,
|
|
"eval_biology_num_tokens": 5531902.0,
|
|
"eval_biology_runtime": 26.9787,
|
|
"eval_biology_samples_per_second": 18.533,
|
|
"eval_biology_steps_per_second": 9.267,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"eval_chemistry_entropy": 1.7318408770561218,
|
|
"eval_chemistry_loss": 1.746923804283142,
|
|
"eval_chemistry_mean_token_accuracy": 0.632696654677391,
|
|
"eval_chemistry_num_tokens": 5531902.0,
|
|
"eval_chemistry_runtime": 34.5046,
|
|
"eval_chemistry_samples_per_second": 14.491,
|
|
"eval_chemistry_steps_per_second": 7.245,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"eval_math_entropy": 0.9170646283626557,
|
|
"eval_math_loss": 0.8942311406135559,
|
|
"eval_math_mean_token_accuracy": 0.7709064717292786,
|
|
"eval_math_num_tokens": 5531902.0,
|
|
"eval_math_runtime": 35.0734,
|
|
"eval_math_samples_per_second": 14.256,
|
|
"eval_math_steps_per_second": 7.128,
|
|
"step": 400
|
|
},
|
|
{
|
|
"epoch": 0.32,
|
|
"eval_physics_entropy": 1.349606383562088,
|
|
"eval_physics_loss": 1.343300223350525,
|
|
"eval_physics_mean_token_accuracy": 0.6978710038661957,
|
|
"eval_physics_num_tokens": 5531902.0,
|
|
"eval_physics_runtime": 40.7692,
|
|
"eval_physics_samples_per_second": 12.264,
|
|
"eval_physics_steps_per_second": 6.132,
|
|
"step": 400
|
|
},
|
|
{
|
|
"entropy": 0.9086290217936039,
|
|
"epoch": 0.328,
|
|
"grad_norm": 0.796875,
|
|
"learning_rate": 1.9192592592592593e-05,
|
|
"loss": 0.847,
|
|
"mean_token_accuracy": 0.7745547294616699,
|
|
"num_tokens": 5669047.0,
|
|
"step": 410
|
|
},
|
|
{
|
|
"entropy": 0.9156009530648589,
|
|
"epoch": 0.336,
|
|
"grad_norm": 0.7578125,
|
|
"learning_rate": 1.911851851851852e-05,
|
|
"loss": 0.8769,
|
|
"mean_token_accuracy": 0.7692474499344826,
|
|
"num_tokens": 5806889.0,
|
|
"step": 420
|
|
},
|
|
{
|
|
"entropy": 0.8934111742302775,
|
|
"epoch": 0.344,
|
|
"grad_norm": 0.67578125,
|
|
"learning_rate": 1.9044444444444446e-05,
|
|
"loss": 0.8462,
|
|
"mean_token_accuracy": 0.7796071924269199,
|
|
"num_tokens": 5947019.0,
|
|
"step": 430
|
|
},
|
|
{
|
|
"entropy": 0.8856402298435568,
|
|
"epoch": 0.352,
|
|
"grad_norm": 0.7578125,
|
|
"learning_rate": 1.8970370370370372e-05,
|
|
"loss": 0.8362,
|
|
"mean_token_accuracy": 0.7777645040303469,
|
|
"num_tokens": 6083443.0,
|
|
"step": 440
|
|
},
|
|
{
|
|
"entropy": 0.8964365001767873,
|
|
"epoch": 0.36,
|
|
"grad_norm": 0.7109375,
|
|
"learning_rate": 1.8896296296296295e-05,
|
|
"loss": 0.8406,
|
|
"mean_token_accuracy": 0.776063571870327,
|
|
"num_tokens": 6215195.0,
|
|
"step": 450
|
|
},
|
|
{
|
|
"entropy": 0.8810398764908314,
|
|
"epoch": 0.368,
|
|
"grad_norm": 0.69921875,
|
|
"learning_rate": 1.8822222222222225e-05,
|
|
"loss": 0.8247,
|
|
"mean_token_accuracy": 0.7789717759937048,
|
|
"num_tokens": 6357305.0,
|
|
"step": 460
|
|
},
|
|
{
|
|
"entropy": 0.8900227572768926,
|
|
"epoch": 0.376,
|
|
"grad_norm": 0.6875,
|
|
"learning_rate": 1.874814814814815e-05,
|
|
"loss": 0.8424,
|
|
"mean_token_accuracy": 0.7792754102498293,
|
|
"num_tokens": 6498815.0,
|
|
"step": 470
|
|
},
|
|
{
|
|
"entropy": 0.8727197473868727,
|
|
"epoch": 0.384,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.8674074074074075e-05,
|
|
"loss": 0.8209,
|
|
"mean_token_accuracy": 0.7820516306906938,
|
|
"num_tokens": 6635429.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"entropy": 0.8815433232113719,
|
|
"epoch": 0.392,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 1.86e-05,
|
|
"loss": 0.8274,
|
|
"mean_token_accuracy": 0.7792554065585137,
|
|
"num_tokens": 6774408.0,
|
|
"step": 490
|
|
},
|
|
{
|
|
"entropy": 0.9402453400194645,
|
|
"epoch": 0.4,
|
|
"grad_norm": 0.78515625,
|
|
"learning_rate": 1.8525925925925928e-05,
|
|
"loss": 0.9268,
|
|
"mean_token_accuracy": 0.7666260946542025,
|
|
"num_tokens": 6914489.0,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"eval_biology_entropy": 2.984537806510925,
|
|
"eval_biology_loss": 3.180291175842285,
|
|
"eval_biology_mean_token_accuracy": 0.4551255764961243,
|
|
"eval_biology_num_tokens": 6914489.0,
|
|
"eval_biology_runtime": 26.9382,
|
|
"eval_biology_samples_per_second": 18.561,
|
|
"eval_biology_steps_per_second": 9.281,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"eval_chemistry_entropy": 1.735370470046997,
|
|
"eval_chemistry_loss": 1.7186591625213623,
|
|
"eval_chemistry_mean_token_accuracy": 0.6382190730571747,
|
|
"eval_chemistry_num_tokens": 6914489.0,
|
|
"eval_chemistry_runtime": 34.4276,
|
|
"eval_chemistry_samples_per_second": 14.523,
|
|
"eval_chemistry_steps_per_second": 7.262,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"eval_math_entropy": 0.9168894953727722,
|
|
"eval_math_loss": 0.8749539852142334,
|
|
"eval_math_mean_token_accuracy": 0.7746825203895569,
|
|
"eval_math_num_tokens": 6914489.0,
|
|
"eval_math_runtime": 34.9026,
|
|
"eval_math_samples_per_second": 14.326,
|
|
"eval_math_steps_per_second": 7.163,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.4,
|
|
"eval_physics_entropy": 1.34889164853096,
|
|
"eval_physics_loss": 1.3155016899108887,
|
|
"eval_physics_mean_token_accuracy": 0.7029694278240204,
|
|
"eval_physics_num_tokens": 6914489.0,
|
|
"eval_physics_runtime": 40.5298,
|
|
"eval_physics_samples_per_second": 12.337,
|
|
"eval_physics_steps_per_second": 6.168,
|
|
"step": 500
|
|
},
|
|
{
|
|
"entropy": 0.8826847270131111,
|
|
"epoch": 0.408,
|
|
"grad_norm": 0.6328125,
|
|
"learning_rate": 1.8451851851851855e-05,
|
|
"loss": 0.8191,
|
|
"mean_token_accuracy": 0.7815356869250536,
|
|
"num_tokens": 7053065.0,
|
|
"step": 510
|
|
},
|
|
{
|
|
"entropy": 0.888327713869512,
|
|
"epoch": 0.416,
|
|
"grad_norm": 0.67578125,
|
|
"learning_rate": 1.8377777777777778e-05,
|
|
"loss": 0.8458,
|
|
"mean_token_accuracy": 0.7794966764748097,
|
|
"num_tokens": 7187783.0,
|
|
"step": 520
|
|
},
|
|
{
|
|
"entropy": 0.8952456833794713,
|
|
"epoch": 0.424,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.8303703703703704e-05,
|
|
"loss": 0.8535,
|
|
"mean_token_accuracy": 0.7754339158535004,
|
|
"num_tokens": 7320375.0,
|
|
"step": 530
|
|
},
|
|
{
|
|
"entropy": 0.8508788656443358,
|
|
"epoch": 0.432,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.822962962962963e-05,
|
|
"loss": 0.7973,
|
|
"mean_token_accuracy": 0.7881547212600708,
|
|
"num_tokens": 7454401.0,
|
|
"step": 540
|
|
},
|
|
{
|
|
"entropy": 0.8842971453443169,
|
|
"epoch": 0.44,
|
|
"grad_norm": 0.71484375,
|
|
"learning_rate": 1.8155555555555557e-05,
|
|
"loss": 0.8373,
|
|
"mean_token_accuracy": 0.7774394288659096,
|
|
"num_tokens": 7593764.0,
|
|
"step": 550
|
|
},
|
|
{
|
|
"entropy": 0.8545080715790391,
|
|
"epoch": 0.448,
|
|
"grad_norm": 0.75,
|
|
"learning_rate": 1.8081481481481484e-05,
|
|
"loss": 0.8067,
|
|
"mean_token_accuracy": 0.7855051398277283,
|
|
"num_tokens": 7732710.0,
|
|
"step": 560
|
|
},
|
|
{
|
|
"entropy": 0.8548783052712678,
|
|
"epoch": 0.456,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 1.800740740740741e-05,
|
|
"loss": 0.8214,
|
|
"mean_token_accuracy": 0.7835484184324741,
|
|
"num_tokens": 7873721.0,
|
|
"step": 570
|
|
},
|
|
{
|
|
"entropy": 0.8870714059099555,
|
|
"epoch": 0.464,
|
|
"grad_norm": 0.5859375,
|
|
"learning_rate": 1.7933333333333333e-05,
|
|
"loss": 0.8303,
|
|
"mean_token_accuracy": 0.7804129023104907,
|
|
"num_tokens": 8012478.0,
|
|
"step": 580
|
|
},
|
|
{
|
|
"entropy": 0.8966930942609906,
|
|
"epoch": 0.472,
|
|
"grad_norm": 0.64453125,
|
|
"learning_rate": 1.785925925925926e-05,
|
|
"loss": 0.8442,
|
|
"mean_token_accuracy": 0.7768670186400414,
|
|
"num_tokens": 8153429.0,
|
|
"step": 590
|
|
},
|
|
{
|
|
"entropy": 0.8990984376519918,
|
|
"epoch": 0.48,
|
|
"grad_norm": 0.7265625,
|
|
"learning_rate": 1.7785185185185186e-05,
|
|
"loss": 0.8363,
|
|
"mean_token_accuracy": 0.7791902955621481,
|
|
"num_tokens": 8288086.0,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.48,
|
|
"eval_biology_entropy": 3.026648554325104,
|
|
"eval_biology_loss": 3.1774065494537354,
|
|
"eval_biology_mean_token_accuracy": 0.45753026235103605,
|
|
"eval_biology_num_tokens": 8288086.0,
|
|
"eval_biology_runtime": 26.9914,
|
|
"eval_biology_samples_per_second": 18.524,
|
|
"eval_biology_steps_per_second": 9.262,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.48,
|
|
"eval_chemistry_entropy": 1.7211642112731933,
|
|
"eval_chemistry_loss": 1.7037526369094849,
|
|
"eval_chemistry_mean_token_accuracy": 0.6409431176185608,
|
|
"eval_chemistry_num_tokens": 8288086.0,
|
|
"eval_chemistry_runtime": 34.4519,
|
|
"eval_chemistry_samples_per_second": 14.513,
|
|
"eval_chemistry_steps_per_second": 7.256,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.48,
|
|
"eval_math_entropy": 0.8927879617214203,
|
|
"eval_math_loss": 0.863281786441803,
|
|
"eval_math_mean_token_accuracy": 0.7770743372440339,
|
|
"eval_math_num_tokens": 8288086.0,
|
|
"eval_math_runtime": 34.9704,
|
|
"eval_math_samples_per_second": 14.298,
|
|
"eval_math_steps_per_second": 7.149,
|
|
"step": 600
|
|
},
|
|
{
|
|
"epoch": 0.48,
|
|
"eval_physics_entropy": 1.3300827517509461,
|
|
"eval_physics_loss": 1.3015602827072144,
|
|
"eval_physics_mean_token_accuracy": 0.7059740360975265,
|
|
"eval_physics_num_tokens": 8288086.0,
|
|
"eval_physics_runtime": 40.5426,
|
|
"eval_physics_samples_per_second": 12.333,
|
|
"eval_physics_steps_per_second": 6.166,
|
|
"step": 600
|
|
},
|
|
{
|
|
"entropy": 0.8605974804610014,
|
|
"epoch": 0.488,
|
|
"grad_norm": 0.65234375,
|
|
"learning_rate": 1.7711111111111113e-05,
|
|
"loss": 0.8123,
|
|
"mean_token_accuracy": 0.7824984133243561,
|
|
"num_tokens": 8424958.0,
|
|
"step": 610
|
|
},
|
|
{
|
|
"entropy": 0.8556231519207358,
|
|
"epoch": 0.496,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 1.763703703703704e-05,
|
|
"loss": 0.8073,
|
|
"mean_token_accuracy": 0.7834980070590973,
|
|
"num_tokens": 8562656.0,
|
|
"step": 620
|
|
},
|
|
{
|
|
"entropy": 0.8707238981500268,
|
|
"epoch": 0.504,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 1.7562962962962962e-05,
|
|
"loss": 0.8323,
|
|
"mean_token_accuracy": 0.7818343795835971,
|
|
"num_tokens": 8706097.0,
|
|
"step": 630
|
|
},
|
|
{
|
|
"entropy": 0.8535389142110944,
|
|
"epoch": 0.512,
|
|
"grad_norm": 0.6953125,
|
|
"learning_rate": 1.7488888888888892e-05,
|
|
"loss": 0.7829,
|
|
"mean_token_accuracy": 0.7894029818475247,
|
|
"num_tokens": 8844762.0,
|
|
"step": 640
|
|
},
|
|
{
|
|
"entropy": 0.8502879558131099,
|
|
"epoch": 0.52,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 1.7414814814814815e-05,
|
|
"loss": 0.8016,
|
|
"mean_token_accuracy": 0.789192034304142,
|
|
"num_tokens": 8983268.0,
|
|
"step": 650
|
|
},
|
|
{
|
|
"entropy": 0.841147867590189,
|
|
"epoch": 0.528,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 1.7340740740740742e-05,
|
|
"loss": 0.7936,
|
|
"mean_token_accuracy": 0.7880374405533075,
|
|
"num_tokens": 9132725.0,
|
|
"step": 660
|
|
},
|
|
{
|
|
"entropy": 0.8636940447613597,
|
|
"epoch": 0.536,
|
|
"grad_norm": 0.796875,
|
|
"learning_rate": 1.726666666666667e-05,
|
|
"loss": 0.841,
|
|
"mean_token_accuracy": 0.7832605082541704,
|
|
"num_tokens": 9271490.0,
|
|
"step": 670
|
|
},
|
|
{
|
|
"entropy": 0.8671527728438377,
|
|
"epoch": 0.544,
|
|
"grad_norm": 0.78125,
|
|
"learning_rate": 1.7192592592592595e-05,
|
|
"loss": 0.8081,
|
|
"mean_token_accuracy": 0.7864414434880018,
|
|
"num_tokens": 9412327.0,
|
|
"step": 680
|
|
},
|
|
{
|
|
"entropy": 0.9130071861669421,
|
|
"epoch": 0.552,
|
|
"grad_norm": 0.85546875,
|
|
"learning_rate": 1.711851851851852e-05,
|
|
"loss": 0.8653,
|
|
"mean_token_accuracy": 0.7723087813705206,
|
|
"num_tokens": 9541196.0,
|
|
"step": 690
|
|
},
|
|
{
|
|
"entropy": 0.8585997836664319,
|
|
"epoch": 0.56,
|
|
"grad_norm": 0.7265625,
|
|
"learning_rate": 1.7044444444444445e-05,
|
|
"loss": 0.794,
|
|
"mean_token_accuracy": 0.7848813243210315,
|
|
"num_tokens": 9676595.0,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"eval_biology_entropy": 2.9588126630783083,
|
|
"eval_biology_loss": 3.134213447570801,
|
|
"eval_biology_mean_token_accuracy": 0.46100987315177916,
|
|
"eval_biology_num_tokens": 9676595.0,
|
|
"eval_biology_runtime": 27.0044,
|
|
"eval_biology_samples_per_second": 18.516,
|
|
"eval_biology_steps_per_second": 9.258,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"eval_chemistry_entropy": 1.6786610894203187,
|
|
"eval_chemistry_loss": 1.6839625835418701,
|
|
"eval_chemistry_mean_token_accuracy": 0.6436767561435699,
|
|
"eval_chemistry_num_tokens": 9676595.0,
|
|
"eval_chemistry_runtime": 34.4136,
|
|
"eval_chemistry_samples_per_second": 14.529,
|
|
"eval_chemistry_steps_per_second": 7.265,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"eval_math_entropy": 0.8769606277942658,
|
|
"eval_math_loss": 0.8541995882987976,
|
|
"eval_math_mean_token_accuracy": 0.7790154383182526,
|
|
"eval_math_num_tokens": 9676595.0,
|
|
"eval_math_runtime": 34.9482,
|
|
"eval_math_samples_per_second": 14.307,
|
|
"eval_math_steps_per_second": 7.153,
|
|
"step": 700
|
|
},
|
|
{
|
|
"epoch": 0.56,
|
|
"eval_physics_entropy": 1.302037154197693,
|
|
"eval_physics_loss": 1.289175271987915,
|
|
"eval_physics_mean_token_accuracy": 0.7078491554260254,
|
|
"eval_physics_num_tokens": 9676595.0,
|
|
"eval_physics_runtime": 40.5408,
|
|
"eval_physics_samples_per_second": 12.333,
|
|
"eval_physics_steps_per_second": 6.167,
|
|
"step": 700
|
|
},
|
|
{
|
|
"entropy": 0.8488044740632177,
|
|
"epoch": 0.568,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.697037037037037e-05,
|
|
"loss": 0.8027,
|
|
"mean_token_accuracy": 0.7859589237719774,
|
|
"num_tokens": 9812856.0,
|
|
"step": 710
|
|
},
|
|
{
|
|
"entropy": 0.8703772563487291,
|
|
"epoch": 0.576,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 1.6896296296296298e-05,
|
|
"loss": 0.8196,
|
|
"mean_token_accuracy": 0.7833330363035202,
|
|
"num_tokens": 9953227.0,
|
|
"step": 720
|
|
},
|
|
{
|
|
"entropy": 0.8656591676175595,
|
|
"epoch": 0.584,
|
|
"grad_norm": 0.7109375,
|
|
"learning_rate": 1.6822222222222224e-05,
|
|
"loss": 0.8193,
|
|
"mean_token_accuracy": 0.7825856961309909,
|
|
"num_tokens": 10089276.0,
|
|
"step": 730
|
|
},
|
|
{
|
|
"entropy": 0.8539155850186944,
|
|
"epoch": 0.592,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 1.6748148148148147e-05,
|
|
"loss": 0.8051,
|
|
"mean_token_accuracy": 0.7846140246838331,
|
|
"num_tokens": 10234220.0,
|
|
"step": 740
|
|
},
|
|
{
|
|
"entropy": 0.8651883415877819,
|
|
"epoch": 0.6,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 1.6674074074074077e-05,
|
|
"loss": 0.8268,
|
|
"mean_token_accuracy": 0.7814744919538498,
|
|
"num_tokens": 10370667.0,
|
|
"step": 750
|
|
},
|
|
{
|
|
"entropy": 0.863803020492196,
|
|
"epoch": 0.608,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 1.66e-05,
|
|
"loss": 0.8229,
|
|
"mean_token_accuracy": 0.783318106085062,
|
|
"num_tokens": 10515760.0,
|
|
"step": 760
|
|
},
|
|
{
|
|
"entropy": 0.8657625541090965,
|
|
"epoch": 0.616,
|
|
"grad_norm": 0.609375,
|
|
"learning_rate": 1.6525925925925927e-05,
|
|
"loss": 0.8201,
|
|
"mean_token_accuracy": 0.781252521276474,
|
|
"num_tokens": 10660578.0,
|
|
"step": 770
|
|
},
|
|
{
|
|
"entropy": 0.8430647728964686,
|
|
"epoch": 0.624,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 1.6451851851851853e-05,
|
|
"loss": 0.786,
|
|
"mean_token_accuracy": 0.7895114913582801,
|
|
"num_tokens": 10797566.0,
|
|
"step": 780
|
|
},
|
|
{
|
|
"entropy": 0.8240785686299205,
|
|
"epoch": 0.632,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 1.637777777777778e-05,
|
|
"loss": 0.7805,
|
|
"mean_token_accuracy": 0.7922219831496478,
|
|
"num_tokens": 10938456.0,
|
|
"step": 790
|
|
},
|
|
{
|
|
"entropy": 0.8485349236056209,
|
|
"epoch": 0.64,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.6303703703703706e-05,
|
|
"loss": 0.7989,
|
|
"mean_token_accuracy": 0.7866497144103051,
|
|
"num_tokens": 11079348.0,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.64,
|
|
"eval_biology_entropy": 2.932964876174927,
|
|
"eval_biology_loss": 3.1179301738739014,
|
|
"eval_biology_mean_token_accuracy": 0.46290264546871185,
|
|
"eval_biology_num_tokens": 11079348.0,
|
|
"eval_biology_runtime": 27.0006,
|
|
"eval_biology_samples_per_second": 18.518,
|
|
"eval_biology_steps_per_second": 9.259,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.64,
|
|
"eval_chemistry_entropy": 1.6625523691177369,
|
|
"eval_chemistry_loss": 1.6735429763793945,
|
|
"eval_chemistry_mean_token_accuracy": 0.645549393773079,
|
|
"eval_chemistry_num_tokens": 11079348.0,
|
|
"eval_chemistry_runtime": 34.504,
|
|
"eval_chemistry_samples_per_second": 14.491,
|
|
"eval_chemistry_steps_per_second": 7.246,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.64,
|
|
"eval_math_entropy": 0.8715307763814926,
|
|
"eval_math_loss": 0.8475318551063538,
|
|
"eval_math_mean_token_accuracy": 0.780082686662674,
|
|
"eval_math_num_tokens": 11079348.0,
|
|
"eval_math_runtime": 34.9793,
|
|
"eval_math_samples_per_second": 14.294,
|
|
"eval_math_steps_per_second": 7.147,
|
|
"step": 800
|
|
},
|
|
{
|
|
"epoch": 0.64,
|
|
"eval_physics_entropy": 1.292870044708252,
|
|
"eval_physics_loss": 1.2783890962600708,
|
|
"eval_physics_mean_token_accuracy": 0.7100210604667664,
|
|
"eval_physics_num_tokens": 11079348.0,
|
|
"eval_physics_runtime": 40.5422,
|
|
"eval_physics_samples_per_second": 12.333,
|
|
"eval_physics_steps_per_second": 6.166,
|
|
"step": 800
|
|
},
|
|
{
|
|
"entropy": 0.8196414019912481,
|
|
"epoch": 0.648,
|
|
"grad_norm": 0.6484375,
|
|
"learning_rate": 1.622962962962963e-05,
|
|
"loss": 0.7785,
|
|
"mean_token_accuracy": 0.7903594326227903,
|
|
"num_tokens": 11221757.0,
|
|
"step": 810
|
|
},
|
|
{
|
|
"entropy": 0.8491896335035563,
|
|
"epoch": 0.656,
|
|
"grad_norm": 0.66796875,
|
|
"learning_rate": 1.6155555555555556e-05,
|
|
"loss": 0.8188,
|
|
"mean_token_accuracy": 0.7861174423247576,
|
|
"num_tokens": 11360239.0,
|
|
"step": 820
|
|
},
|
|
{
|
|
"entropy": 0.8492498191073536,
|
|
"epoch": 0.664,
|
|
"grad_norm": 0.83984375,
|
|
"learning_rate": 1.6081481481481482e-05,
|
|
"loss": 0.7993,
|
|
"mean_token_accuracy": 0.7853016339242458,
|
|
"num_tokens": 11498677.0,
|
|
"step": 830
|
|
},
|
|
{
|
|
"entropy": 0.8525434667244554,
|
|
"epoch": 0.672,
|
|
"grad_norm": 0.69140625,
|
|
"learning_rate": 1.600740740740741e-05,
|
|
"loss": 0.8167,
|
|
"mean_token_accuracy": 0.7853171724826098,
|
|
"num_tokens": 11644674.0,
|
|
"step": 840
|
|
},
|
|
{
|
|
"entropy": 0.8380661699920893,
|
|
"epoch": 0.68,
|
|
"grad_norm": 0.73046875,
|
|
"learning_rate": 1.5933333333333336e-05,
|
|
"loss": 0.7928,
|
|
"mean_token_accuracy": 0.788429806381464,
|
|
"num_tokens": 11783080.0,
|
|
"step": 850
|
|
},
|
|
{
|
|
"entropy": 0.8381451951339841,
|
|
"epoch": 0.688,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 1.5859259259259262e-05,
|
|
"loss": 0.7921,
|
|
"mean_token_accuracy": 0.7877590708434582,
|
|
"num_tokens": 11917713.0,
|
|
"step": 860
|
|
},
|
|
{
|
|
"entropy": 0.8567008957266807,
|
|
"epoch": 0.696,
|
|
"grad_norm": 0.6640625,
|
|
"learning_rate": 1.5785185185185185e-05,
|
|
"loss": 0.7966,
|
|
"mean_token_accuracy": 0.7848114792257548,
|
|
"num_tokens": 12058193.0,
|
|
"step": 870
|
|
},
|
|
{
|
|
"entropy": 0.8353428209200502,
|
|
"epoch": 0.704,
|
|
"grad_norm": 0.65234375,
|
|
"learning_rate": 1.571111111111111e-05,
|
|
"loss": 0.7868,
|
|
"mean_token_accuracy": 0.7881882593035698,
|
|
"num_tokens": 12199067.0,
|
|
"step": 880
|
|
},
|
|
{
|
|
"entropy": 0.8710033852607012,
|
|
"epoch": 0.712,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.5637037037037038e-05,
|
|
"loss": 0.8088,
|
|
"mean_token_accuracy": 0.7816738404333592,
|
|
"num_tokens": 12335740.0,
|
|
"step": 890
|
|
},
|
|
{
|
|
"entropy": 0.8482802568003536,
|
|
"epoch": 0.72,
|
|
"grad_norm": 0.8671875,
|
|
"learning_rate": 1.5562962962962965e-05,
|
|
"loss": 0.8002,
|
|
"mean_token_accuracy": 0.7863489333540201,
|
|
"num_tokens": 12469017.0,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"eval_biology_entropy": 2.9198746318817137,
|
|
"eval_biology_loss": 3.095550775527954,
|
|
"eval_biology_mean_token_accuracy": 0.4650931693315506,
|
|
"eval_biology_num_tokens": 12469017.0,
|
|
"eval_biology_runtime": 26.9459,
|
|
"eval_biology_samples_per_second": 18.556,
|
|
"eval_biology_steps_per_second": 9.278,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"eval_chemistry_entropy": 1.654184314250946,
|
|
"eval_chemistry_loss": 1.6625065803527832,
|
|
"eval_chemistry_mean_token_accuracy": 0.6471669510602951,
|
|
"eval_chemistry_num_tokens": 12469017.0,
|
|
"eval_chemistry_runtime": 34.4235,
|
|
"eval_chemistry_samples_per_second": 14.525,
|
|
"eval_chemistry_steps_per_second": 7.262,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"eval_math_entropy": 0.8690718128681183,
|
|
"eval_math_loss": 0.8421436548233032,
|
|
"eval_math_mean_token_accuracy": 0.7811065847873688,
|
|
"eval_math_num_tokens": 12469017.0,
|
|
"eval_math_runtime": 34.9224,
|
|
"eval_math_samples_per_second": 14.317,
|
|
"eval_math_steps_per_second": 7.159,
|
|
"step": 900
|
|
},
|
|
{
|
|
"epoch": 0.72,
|
|
"eval_physics_entropy": 1.2865873792171478,
|
|
"eval_physics_loss": 1.2697285413742065,
|
|
"eval_physics_mean_token_accuracy": 0.7113078062534333,
|
|
"eval_physics_num_tokens": 12469017.0,
|
|
"eval_physics_runtime": 40.5051,
|
|
"eval_physics_samples_per_second": 12.344,
|
|
"eval_physics_steps_per_second": 6.172,
|
|
"step": 900
|
|
},
|
|
{
|
|
"entropy": 0.8559863271191717,
|
|
"epoch": 0.728,
|
|
"grad_norm": 0.859375,
|
|
"learning_rate": 1.548888888888889e-05,
|
|
"loss": 0.8087,
|
|
"mean_token_accuracy": 0.7872065585106611,
|
|
"num_tokens": 12604524.0,
|
|
"step": 910
|
|
},
|
|
{
|
|
"entropy": 0.8351603880524635,
|
|
"epoch": 0.736,
|
|
"grad_norm": 0.69921875,
|
|
"learning_rate": 1.5414814814814814e-05,
|
|
"loss": 0.7815,
|
|
"mean_token_accuracy": 0.7907860331237316,
|
|
"num_tokens": 12746145.0,
|
|
"step": 920
|
|
},
|
|
{
|
|
"entropy": 0.8385633688420058,
|
|
"epoch": 0.744,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.5340740740740744e-05,
|
|
"loss": 0.7849,
|
|
"mean_token_accuracy": 0.7884344782680273,
|
|
"num_tokens": 12882140.0,
|
|
"step": 930
|
|
},
|
|
{
|
|
"entropy": 0.8265942361205816,
|
|
"epoch": 0.752,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.5266666666666667e-05,
|
|
"loss": 0.7683,
|
|
"mean_token_accuracy": 0.7920810837298632,
|
|
"num_tokens": 13019785.0,
|
|
"step": 940
|
|
},
|
|
{
|
|
"entropy": 0.801921004243195,
|
|
"epoch": 0.76,
|
|
"grad_norm": 0.64453125,
|
|
"learning_rate": 1.5192592592592594e-05,
|
|
"loss": 0.7567,
|
|
"mean_token_accuracy": 0.7981402266770601,
|
|
"num_tokens": 13167975.0,
|
|
"step": 950
|
|
},
|
|
{
|
|
"entropy": 0.8200628351420164,
|
|
"epoch": 0.768,
|
|
"grad_norm": 0.6015625,
|
|
"learning_rate": 1.5118518518518519e-05,
|
|
"loss": 0.7712,
|
|
"mean_token_accuracy": 0.7937319375574589,
|
|
"num_tokens": 13307071.0,
|
|
"step": 960
|
|
},
|
|
{
|
|
"entropy": 0.844639228284359,
|
|
"epoch": 0.776,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.5044444444444445e-05,
|
|
"loss": 0.8066,
|
|
"mean_token_accuracy": 0.7887587692588568,
|
|
"num_tokens": 13444206.0,
|
|
"step": 970
|
|
},
|
|
{
|
|
"entropy": 0.8560901025310159,
|
|
"epoch": 0.784,
|
|
"grad_norm": 0.7109375,
|
|
"learning_rate": 1.497037037037037e-05,
|
|
"loss": 0.8159,
|
|
"mean_token_accuracy": 0.7839560698717832,
|
|
"num_tokens": 13586074.0,
|
|
"step": 980
|
|
},
|
|
{
|
|
"entropy": 0.8473082648590207,
|
|
"epoch": 0.792,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 1.4896296296296298e-05,
|
|
"loss": 0.8055,
|
|
"mean_token_accuracy": 0.785613476485014,
|
|
"num_tokens": 13727642.0,
|
|
"step": 990
|
|
},
|
|
{
|
|
"entropy": 0.8641716981306672,
|
|
"epoch": 0.8,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 1.4822222222222225e-05,
|
|
"loss": 0.8161,
|
|
"mean_token_accuracy": 0.7843089539557695,
|
|
"num_tokens": 13869252.0,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"eval_biology_entropy": 2.927406606197357,
|
|
"eval_biology_loss": 3.0816211700439453,
|
|
"eval_biology_mean_token_accuracy": 0.46697849237918854,
|
|
"eval_biology_num_tokens": 13869252.0,
|
|
"eval_biology_runtime": 26.9151,
|
|
"eval_biology_samples_per_second": 18.577,
|
|
"eval_biology_steps_per_second": 9.288,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"eval_chemistry_entropy": 1.6538746666908264,
|
|
"eval_chemistry_loss": 1.6552778482437134,
|
|
"eval_chemistry_mean_token_accuracy": 0.6485219527482986,
|
|
"eval_chemistry_num_tokens": 13869252.0,
|
|
"eval_chemistry_runtime": 34.4246,
|
|
"eval_chemistry_samples_per_second": 14.525,
|
|
"eval_chemistry_steps_per_second": 7.262,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"eval_math_entropy": 0.8677517844438553,
|
|
"eval_math_loss": 0.8380001187324524,
|
|
"eval_math_mean_token_accuracy": 0.7821021761894226,
|
|
"eval_math_num_tokens": 13869252.0,
|
|
"eval_math_runtime": 34.92,
|
|
"eval_math_samples_per_second": 14.318,
|
|
"eval_math_steps_per_second": 7.159,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"epoch": 0.8,
|
|
"eval_physics_entropy": 1.2842487235069275,
|
|
"eval_physics_loss": 1.2633119821548462,
|
|
"eval_physics_mean_token_accuracy": 0.7128965456485749,
|
|
"eval_physics_num_tokens": 13869252.0,
|
|
"eval_physics_runtime": 40.4838,
|
|
"eval_physics_samples_per_second": 12.351,
|
|
"eval_physics_steps_per_second": 6.175,
|
|
"step": 1000
|
|
},
|
|
{
|
|
"entropy": 0.826583449728787,
|
|
"epoch": 0.808,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.474814814814815e-05,
|
|
"loss": 0.7747,
|
|
"mean_token_accuracy": 0.7898658849298954,
|
|
"num_tokens": 14008931.0,
|
|
"step": 1010
|
|
},
|
|
{
|
|
"entropy": 0.8473596401512623,
|
|
"epoch": 0.816,
|
|
"grad_norm": 0.69140625,
|
|
"learning_rate": 1.4674074074074076e-05,
|
|
"loss": 0.7943,
|
|
"mean_token_accuracy": 0.7863069783896208,
|
|
"num_tokens": 14151374.0,
|
|
"step": 1020
|
|
},
|
|
{
|
|
"entropy": 0.8361653342843056,
|
|
"epoch": 0.824,
|
|
"grad_norm": 0.71484375,
|
|
"learning_rate": 1.46e-05,
|
|
"loss": 0.7867,
|
|
"mean_token_accuracy": 0.7892423421144485,
|
|
"num_tokens": 14291862.0,
|
|
"step": 1030
|
|
},
|
|
{
|
|
"entropy": 0.832403457723558,
|
|
"epoch": 0.832,
|
|
"grad_norm": 0.7578125,
|
|
"learning_rate": 1.4525925925925927e-05,
|
|
"loss": 0.7757,
|
|
"mean_token_accuracy": 0.792064693197608,
|
|
"num_tokens": 14429535.0,
|
|
"step": 1040
|
|
},
|
|
{
|
|
"entropy": 0.8536718141287565,
|
|
"epoch": 0.84,
|
|
"grad_norm": 0.8359375,
|
|
"learning_rate": 1.4451851851851852e-05,
|
|
"loss": 0.8129,
|
|
"mean_token_accuracy": 0.7852799411863088,
|
|
"num_tokens": 14565836.0,
|
|
"step": 1050
|
|
},
|
|
{
|
|
"entropy": 0.8711680036038161,
|
|
"epoch": 0.848,
|
|
"grad_norm": 0.64453125,
|
|
"learning_rate": 1.4377777777777779e-05,
|
|
"loss": 0.822,
|
|
"mean_token_accuracy": 0.7821739461272955,
|
|
"num_tokens": 14703912.0,
|
|
"step": 1060
|
|
},
|
|
{
|
|
"entropy": 0.8762293307110667,
|
|
"epoch": 0.856,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 1.4303703703703703e-05,
|
|
"loss": 0.8157,
|
|
"mean_token_accuracy": 0.7823412474244833,
|
|
"num_tokens": 14843747.0,
|
|
"step": 1070
|
|
},
|
|
{
|
|
"entropy": 0.8367005018517375,
|
|
"epoch": 0.864,
|
|
"grad_norm": 0.67578125,
|
|
"learning_rate": 1.4229629629629632e-05,
|
|
"loss": 0.7889,
|
|
"mean_token_accuracy": 0.7886287022382021,
|
|
"num_tokens": 14981626.0,
|
|
"step": 1080
|
|
},
|
|
{
|
|
"entropy": 0.8624969441443682,
|
|
"epoch": 0.872,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.4155555555555556e-05,
|
|
"loss": 0.8014,
|
|
"mean_token_accuracy": 0.7859421294182539,
|
|
"num_tokens": 15119687.0,
|
|
"step": 1090
|
|
},
|
|
{
|
|
"entropy": 0.8763546861708165,
|
|
"epoch": 0.88,
|
|
"grad_norm": 0.88671875,
|
|
"learning_rate": 1.4081481481481483e-05,
|
|
"loss": 0.8178,
|
|
"mean_token_accuracy": 0.778026782348752,
|
|
"num_tokens": 15256318.0,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 0.88,
|
|
"eval_biology_entropy": 2.896754152297974,
|
|
"eval_biology_loss": 3.063400983810425,
|
|
"eval_biology_mean_token_accuracy": 0.46827907478809355,
|
|
"eval_biology_num_tokens": 15256318.0,
|
|
"eval_biology_runtime": 27.0374,
|
|
"eval_biology_samples_per_second": 18.493,
|
|
"eval_biology_steps_per_second": 9.246,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 0.88,
|
|
"eval_chemistry_entropy": 1.638319587945938,
|
|
"eval_chemistry_loss": 1.6441051959991455,
|
|
"eval_chemistry_mean_token_accuracy": 0.6500709130764007,
|
|
"eval_chemistry_num_tokens": 15256318.0,
|
|
"eval_chemistry_runtime": 34.5158,
|
|
"eval_chemistry_samples_per_second": 14.486,
|
|
"eval_chemistry_steps_per_second": 7.243,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 0.88,
|
|
"eval_math_entropy": 0.862964648604393,
|
|
"eval_math_loss": 0.8344796299934387,
|
|
"eval_math_mean_token_accuracy": 0.7827803816795349,
|
|
"eval_math_num_tokens": 15256318.0,
|
|
"eval_math_runtime": 34.9946,
|
|
"eval_math_samples_per_second": 14.288,
|
|
"eval_math_steps_per_second": 7.144,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"epoch": 0.88,
|
|
"eval_physics_entropy": 1.2776771280765533,
|
|
"eval_physics_loss": 1.2576273679733276,
|
|
"eval_physics_mean_token_accuracy": 0.7135676975250245,
|
|
"eval_physics_num_tokens": 15256318.0,
|
|
"eval_physics_runtime": 40.5631,
|
|
"eval_physics_samples_per_second": 12.326,
|
|
"eval_physics_steps_per_second": 6.163,
|
|
"step": 1100
|
|
},
|
|
{
|
|
"entropy": 0.8450109241530299,
|
|
"epoch": 0.888,
|
|
"grad_norm": 0.703125,
|
|
"learning_rate": 1.400740740740741e-05,
|
|
"loss": 0.7945,
|
|
"mean_token_accuracy": 0.7892760831862689,
|
|
"num_tokens": 15390095.0,
|
|
"step": 1110
|
|
},
|
|
{
|
|
"entropy": 0.8675164245069027,
|
|
"epoch": 0.896,
|
|
"grad_norm": 0.71484375,
|
|
"learning_rate": 1.3933333333333334e-05,
|
|
"loss": 0.8145,
|
|
"mean_token_accuracy": 0.7832493595778942,
|
|
"num_tokens": 15527679.0,
|
|
"step": 1120
|
|
},
|
|
{
|
|
"entropy": 0.8140978695824742,
|
|
"epoch": 0.904,
|
|
"grad_norm": 0.66796875,
|
|
"learning_rate": 1.385925925925926e-05,
|
|
"loss": 0.7663,
|
|
"mean_token_accuracy": 0.7915640484541655,
|
|
"num_tokens": 15667231.0,
|
|
"step": 1130
|
|
},
|
|
{
|
|
"entropy": 0.8478411814197898,
|
|
"epoch": 0.912,
|
|
"grad_norm": 0.6796875,
|
|
"learning_rate": 1.3785185185185186e-05,
|
|
"loss": 0.7968,
|
|
"mean_token_accuracy": 0.7854505475610495,
|
|
"num_tokens": 15808515.0,
|
|
"step": 1140
|
|
},
|
|
{
|
|
"entropy": 0.8326777325943112,
|
|
"epoch": 0.92,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.3711111111111112e-05,
|
|
"loss": 0.7785,
|
|
"mean_token_accuracy": 0.7895378533750772,
|
|
"num_tokens": 15944670.0,
|
|
"step": 1150
|
|
},
|
|
{
|
|
"entropy": 0.8248712478205562,
|
|
"epoch": 0.928,
|
|
"grad_norm": 0.6015625,
|
|
"learning_rate": 1.3637037037037037e-05,
|
|
"loss": 0.7752,
|
|
"mean_token_accuracy": 0.7912526894360781,
|
|
"num_tokens": 16081416.0,
|
|
"step": 1160
|
|
},
|
|
{
|
|
"entropy": 0.8408484322950244,
|
|
"epoch": 0.936,
|
|
"grad_norm": 0.6796875,
|
|
"learning_rate": 1.3562962962962965e-05,
|
|
"loss": 0.7991,
|
|
"mean_token_accuracy": 0.7868186816573143,
|
|
"num_tokens": 16218228.0,
|
|
"step": 1170
|
|
},
|
|
{
|
|
"entropy": 0.8680851750075818,
|
|
"epoch": 0.944,
|
|
"grad_norm": 0.82421875,
|
|
"learning_rate": 1.3488888888888888e-05,
|
|
"loss": 0.827,
|
|
"mean_token_accuracy": 0.7814383570104837,
|
|
"num_tokens": 16356073.0,
|
|
"step": 1180
|
|
},
|
|
{
|
|
"entropy": 0.8465319074690342,
|
|
"epoch": 0.952,
|
|
"grad_norm": 0.72265625,
|
|
"learning_rate": 1.3414814814814817e-05,
|
|
"loss": 0.7988,
|
|
"mean_token_accuracy": 0.7877550296485424,
|
|
"num_tokens": 16492413.0,
|
|
"step": 1190
|
|
},
|
|
{
|
|
"entropy": 0.8646586099639535,
|
|
"epoch": 0.96,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 1.3340740740740741e-05,
|
|
"loss": 0.8061,
|
|
"mean_token_accuracy": 0.7851971492171288,
|
|
"num_tokens": 16632796.0,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"eval_biology_entropy": 2.908616391181946,
|
|
"eval_biology_loss": 3.0608139038085938,
|
|
"eval_biology_mean_token_accuracy": 0.4685148627758026,
|
|
"eval_biology_num_tokens": 16632796.0,
|
|
"eval_biology_runtime": 26.9626,
|
|
"eval_biology_samples_per_second": 18.544,
|
|
"eval_biology_steps_per_second": 9.272,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"eval_chemistry_entropy": 1.6397918918132781,
|
|
"eval_chemistry_loss": 1.640189290046692,
|
|
"eval_chemistry_mean_token_accuracy": 0.6506748970746994,
|
|
"eval_chemistry_num_tokens": 16632796.0,
|
|
"eval_chemistry_runtime": 34.4834,
|
|
"eval_chemistry_samples_per_second": 14.5,
|
|
"eval_chemistry_steps_per_second": 7.25,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"eval_math_entropy": 0.8619930381774903,
|
|
"eval_math_loss": 0.8314722776412964,
|
|
"eval_math_mean_token_accuracy": 0.783630312204361,
|
|
"eval_math_num_tokens": 16632796.0,
|
|
"eval_math_runtime": 35.3561,
|
|
"eval_math_samples_per_second": 14.142,
|
|
"eval_math_steps_per_second": 7.071,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"epoch": 0.96,
|
|
"eval_physics_entropy": 1.2789247527122498,
|
|
"eval_physics_loss": 1.2552392482757568,
|
|
"eval_physics_mean_token_accuracy": 0.7142456164360046,
|
|
"eval_physics_num_tokens": 16632796.0,
|
|
"eval_physics_runtime": 40.5336,
|
|
"eval_physics_samples_per_second": 12.335,
|
|
"eval_physics_steps_per_second": 6.168,
|
|
"step": 1200
|
|
},
|
|
{
|
|
"entropy": 0.8285112351179122,
|
|
"epoch": 0.968,
|
|
"grad_norm": 0.75390625,
|
|
"learning_rate": 1.3266666666666668e-05,
|
|
"loss": 0.776,
|
|
"mean_token_accuracy": 0.7918854866176843,
|
|
"num_tokens": 16773370.0,
|
|
"step": 1210
|
|
},
|
|
{
|
|
"entropy": 0.8358210870996118,
|
|
"epoch": 0.976,
|
|
"grad_norm": 0.6875,
|
|
"learning_rate": 1.3192592592592594e-05,
|
|
"loss": 0.7953,
|
|
"mean_token_accuracy": 0.7874758914113045,
|
|
"num_tokens": 16908991.0,
|
|
"step": 1220
|
|
},
|
|
{
|
|
"entropy": 0.8504814989864826,
|
|
"epoch": 0.984,
|
|
"grad_norm": 0.71875,
|
|
"learning_rate": 1.311851851851852e-05,
|
|
"loss": 0.8009,
|
|
"mean_token_accuracy": 0.7851403091102839,
|
|
"num_tokens": 17048278.0,
|
|
"step": 1230
|
|
},
|
|
{
|
|
"entropy": 0.8459016673266888,
|
|
"epoch": 0.992,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.3044444444444446e-05,
|
|
"loss": 0.7909,
|
|
"mean_token_accuracy": 0.7847922585904599,
|
|
"num_tokens": 17189644.0,
|
|
"step": 1240
|
|
},
|
|
{
|
|
"entropy": 0.848821627907455,
|
|
"epoch": 1.0,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 1.297037037037037e-05,
|
|
"loss": 0.8135,
|
|
"mean_token_accuracy": 0.7836502768099308,
|
|
"num_tokens": 17329320.0,
|
|
"step": 1250
|
|
},
|
|
{
|
|
"entropy": 0.8311884285882115,
|
|
"epoch": 1.008,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 1.2896296296296299e-05,
|
|
"loss": 0.7746,
|
|
"mean_token_accuracy": 0.7915325038135052,
|
|
"num_tokens": 17471675.0,
|
|
"step": 1260
|
|
},
|
|
{
|
|
"entropy": 0.8471982046961785,
|
|
"epoch": 1.016,
|
|
"grad_norm": 0.77734375,
|
|
"learning_rate": 1.2822222222222222e-05,
|
|
"loss": 0.7847,
|
|
"mean_token_accuracy": 0.7903359431773425,
|
|
"num_tokens": 17611177.0,
|
|
"step": 1270
|
|
},
|
|
{
|
|
"entropy": 0.8569712620228529,
|
|
"epoch": 1.024,
|
|
"grad_norm": 0.6484375,
|
|
"learning_rate": 1.274814814814815e-05,
|
|
"loss": 0.793,
|
|
"mean_token_accuracy": 0.7870935752987862,
|
|
"num_tokens": 17744248.0,
|
|
"step": 1280
|
|
},
|
|
{
|
|
"entropy": 0.8451526958495379,
|
|
"epoch": 1.032,
|
|
"grad_norm": 0.6875,
|
|
"learning_rate": 1.2674074074074075e-05,
|
|
"loss": 0.7979,
|
|
"mean_token_accuracy": 0.7867367029190063,
|
|
"num_tokens": 17881760.0,
|
|
"step": 1290
|
|
},
|
|
{
|
|
"entropy": 0.8108762178570033,
|
|
"epoch": 1.04,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 1.2600000000000001e-05,
|
|
"loss": 0.7625,
|
|
"mean_token_accuracy": 0.7972131386399269,
|
|
"num_tokens": 18014485.0,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"eval_biology_entropy": 2.861083526134491,
|
|
"eval_biology_loss": 3.054857015609741,
|
|
"eval_biology_mean_token_accuracy": 0.469379000544548,
|
|
"eval_biology_num_tokens": 18014485.0,
|
|
"eval_biology_runtime": 26.9793,
|
|
"eval_biology_samples_per_second": 18.533,
|
|
"eval_biology_steps_per_second": 9.266,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"eval_chemistry_entropy": 1.614394629240036,
|
|
"eval_chemistry_loss": 1.6372945308685303,
|
|
"eval_chemistry_mean_token_accuracy": 0.6513354105949402,
|
|
"eval_chemistry_num_tokens": 18014485.0,
|
|
"eval_chemistry_runtime": 34.4802,
|
|
"eval_chemistry_samples_per_second": 14.501,
|
|
"eval_chemistry_steps_per_second": 7.251,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"eval_math_entropy": 0.8521397960186005,
|
|
"eval_math_loss": 0.8296064734458923,
|
|
"eval_math_mean_token_accuracy": 0.7838501288890839,
|
|
"eval_math_num_tokens": 18014485.0,
|
|
"eval_math_runtime": 34.9703,
|
|
"eval_math_samples_per_second": 14.298,
|
|
"eval_math_steps_per_second": 7.149,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"epoch": 1.04,
|
|
"eval_physics_entropy": 1.2607037408351898,
|
|
"eval_physics_loss": 1.2520027160644531,
|
|
"eval_physics_mean_token_accuracy": 0.7147621221542358,
|
|
"eval_physics_num_tokens": 18014485.0,
|
|
"eval_physics_runtime": 40.5706,
|
|
"eval_physics_samples_per_second": 12.324,
|
|
"eval_physics_steps_per_second": 6.162,
|
|
"step": 1300
|
|
},
|
|
{
|
|
"entropy": 0.8444511197507382,
|
|
"epoch": 1.048,
|
|
"grad_norm": 0.71875,
|
|
"learning_rate": 1.2525925925925928e-05,
|
|
"loss": 0.797,
|
|
"mean_token_accuracy": 0.7875816393643618,
|
|
"num_tokens": 18156077.0,
|
|
"step": 1310
|
|
},
|
|
{
|
|
"entropy": 0.8128441723063589,
|
|
"epoch": 1.056,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.2451851851851853e-05,
|
|
"loss": 0.7764,
|
|
"mean_token_accuracy": 0.7942416854202747,
|
|
"num_tokens": 18294650.0,
|
|
"step": 1320
|
|
},
|
|
{
|
|
"entropy": 0.8420778449624777,
|
|
"epoch": 1.064,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.237777777777778e-05,
|
|
"loss": 0.7963,
|
|
"mean_token_accuracy": 0.7890487045049668,
|
|
"num_tokens": 18435862.0,
|
|
"step": 1330
|
|
},
|
|
{
|
|
"entropy": 0.8193851249292493,
|
|
"epoch": 1.072,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.2303703703703704e-05,
|
|
"loss": 0.7754,
|
|
"mean_token_accuracy": 0.7921654149889946,
|
|
"num_tokens": 18568936.0,
|
|
"step": 1340
|
|
},
|
|
{
|
|
"entropy": 0.841069664247334,
|
|
"epoch": 1.08,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.222962962962963e-05,
|
|
"loss": 0.7797,
|
|
"mean_token_accuracy": 0.7857704412192106,
|
|
"num_tokens": 18712013.0,
|
|
"step": 1350
|
|
},
|
|
{
|
|
"entropy": 0.8188380904495716,
|
|
"epoch": 1.088,
|
|
"grad_norm": 0.59765625,
|
|
"learning_rate": 1.2155555555555555e-05,
|
|
"loss": 0.7684,
|
|
"mean_token_accuracy": 0.7954085689038038,
|
|
"num_tokens": 18852497.0,
|
|
"step": 1360
|
|
},
|
|
{
|
|
"entropy": 0.8550961302593351,
|
|
"epoch": 1.096,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.2081481481481484e-05,
|
|
"loss": 0.7868,
|
|
"mean_token_accuracy": 0.7869294688105584,
|
|
"num_tokens": 18997208.0,
|
|
"step": 1370
|
|
},
|
|
{
|
|
"entropy": 0.8430958347395062,
|
|
"epoch": 1.104,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.2007407407407408e-05,
|
|
"loss": 0.7933,
|
|
"mean_token_accuracy": 0.787052009999752,
|
|
"num_tokens": 19136485.0,
|
|
"step": 1380
|
|
},
|
|
{
|
|
"entropy": 0.818677469342947,
|
|
"epoch": 1.112,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 1.1933333333333335e-05,
|
|
"loss": 0.7707,
|
|
"mean_token_accuracy": 0.7958467341959476,
|
|
"num_tokens": 19272237.0,
|
|
"step": 1390
|
|
},
|
|
{
|
|
"entropy": 0.8190919890999794,
|
|
"epoch": 1.12,
|
|
"grad_norm": 0.62890625,
|
|
"learning_rate": 1.185925925925926e-05,
|
|
"loss": 0.7635,
|
|
"mean_token_accuracy": 0.793834337219596,
|
|
"num_tokens": 19415146.0,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.12,
|
|
"eval_biology_entropy": 2.8546162362098695,
|
|
"eval_biology_loss": 3.046806573867798,
|
|
"eval_biology_mean_token_accuracy": 0.47026353430747986,
|
|
"eval_biology_num_tokens": 19415146.0,
|
|
"eval_biology_runtime": 26.9248,
|
|
"eval_biology_samples_per_second": 18.57,
|
|
"eval_biology_steps_per_second": 9.285,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.12,
|
|
"eval_chemistry_entropy": 1.6122739651203155,
|
|
"eval_chemistry_loss": 1.6334017515182495,
|
|
"eval_chemistry_mean_token_accuracy": 0.6514999225139618,
|
|
"eval_chemistry_num_tokens": 19415146.0,
|
|
"eval_chemistry_runtime": 34.4486,
|
|
"eval_chemistry_samples_per_second": 14.514,
|
|
"eval_chemistry_steps_per_second": 7.257,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.12,
|
|
"eval_math_entropy": 0.8521307862997055,
|
|
"eval_math_loss": 0.827960193157196,
|
|
"eval_math_mean_token_accuracy": 0.7839585332870483,
|
|
"eval_math_num_tokens": 19415146.0,
|
|
"eval_math_runtime": 34.9467,
|
|
"eval_math_samples_per_second": 14.308,
|
|
"eval_math_steps_per_second": 7.154,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"epoch": 1.12,
|
|
"eval_physics_entropy": 1.2608284051418304,
|
|
"eval_physics_loss": 1.2490959167480469,
|
|
"eval_physics_mean_token_accuracy": 0.7151757082939147,
|
|
"eval_physics_num_tokens": 19415146.0,
|
|
"eval_physics_runtime": 40.537,
|
|
"eval_physics_samples_per_second": 12.334,
|
|
"eval_physics_steps_per_second": 6.167,
|
|
"step": 1400
|
|
},
|
|
{
|
|
"entropy": 0.8294604141265154,
|
|
"epoch": 1.1280000000000001,
|
|
"grad_norm": 0.7734375,
|
|
"learning_rate": 1.1785185185185186e-05,
|
|
"loss": 0.7807,
|
|
"mean_token_accuracy": 0.7920530281960965,
|
|
"num_tokens": 19550300.0,
|
|
"step": 1410
|
|
},
|
|
{
|
|
"entropy": 0.8140227854251861,
|
|
"epoch": 1.1360000000000001,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 1.1711111111111113e-05,
|
|
"loss": 0.7705,
|
|
"mean_token_accuracy": 0.7926124900579452,
|
|
"num_tokens": 19687362.0,
|
|
"step": 1420
|
|
},
|
|
{
|
|
"entropy": 0.8327476343140006,
|
|
"epoch": 1.144,
|
|
"grad_norm": 0.6328125,
|
|
"learning_rate": 1.1637037037037037e-05,
|
|
"loss": 0.7873,
|
|
"mean_token_accuracy": 0.7886258989572525,
|
|
"num_tokens": 19825647.0,
|
|
"step": 1430
|
|
},
|
|
{
|
|
"entropy": 0.8256687164306641,
|
|
"epoch": 1.152,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 1.1562962962962964e-05,
|
|
"loss": 0.7785,
|
|
"mean_token_accuracy": 0.7936491280794143,
|
|
"num_tokens": 19968311.0,
|
|
"step": 1440
|
|
},
|
|
{
|
|
"entropy": 0.8199941594153642,
|
|
"epoch": 1.16,
|
|
"grad_norm": 0.64453125,
|
|
"learning_rate": 1.1488888888888889e-05,
|
|
"loss": 0.7707,
|
|
"mean_token_accuracy": 0.7947168540209532,
|
|
"num_tokens": 20115101.0,
|
|
"step": 1450
|
|
},
|
|
{
|
|
"entropy": 0.8253829684108496,
|
|
"epoch": 1.168,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.1414814814814817e-05,
|
|
"loss": 0.7826,
|
|
"mean_token_accuracy": 0.7904348153620958,
|
|
"num_tokens": 20254700.0,
|
|
"step": 1460
|
|
},
|
|
{
|
|
"entropy": 0.8298774238675832,
|
|
"epoch": 1.176,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.1340740740740742e-05,
|
|
"loss": 0.7714,
|
|
"mean_token_accuracy": 0.7898557730019092,
|
|
"num_tokens": 20394054.0,
|
|
"step": 1470
|
|
},
|
|
{
|
|
"entropy": 0.8195257894694805,
|
|
"epoch": 1.184,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 1.1266666666666668e-05,
|
|
"loss": 0.7762,
|
|
"mean_token_accuracy": 0.7923125520348548,
|
|
"num_tokens": 20535739.0,
|
|
"step": 1480
|
|
},
|
|
{
|
|
"entropy": 0.8128214536234737,
|
|
"epoch": 1.192,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 1.1192592592592593e-05,
|
|
"loss": 0.7665,
|
|
"mean_token_accuracy": 0.7968613497912884,
|
|
"num_tokens": 20679391.0,
|
|
"step": 1490
|
|
},
|
|
{
|
|
"entropy": 0.8249955477192998,
|
|
"epoch": 1.2,
|
|
"grad_norm": 0.6640625,
|
|
"learning_rate": 1.111851851851852e-05,
|
|
"loss": 0.7804,
|
|
"mean_token_accuracy": 0.789951428771019,
|
|
"num_tokens": 20822188.0,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"eval_biology_entropy": 2.8529665331840515,
|
|
"eval_biology_loss": 3.043259620666504,
|
|
"eval_biology_mean_token_accuracy": 0.4708596212863922,
|
|
"eval_biology_num_tokens": 20822188.0,
|
|
"eval_biology_runtime": 26.932,
|
|
"eval_biology_samples_per_second": 18.565,
|
|
"eval_biology_steps_per_second": 9.283,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"eval_chemistry_entropy": 1.6120401091575622,
|
|
"eval_chemistry_loss": 1.6317973136901855,
|
|
"eval_chemistry_mean_token_accuracy": 0.6518439574241638,
|
|
"eval_chemistry_num_tokens": 20822188.0,
|
|
"eval_chemistry_runtime": 34.4788,
|
|
"eval_chemistry_samples_per_second": 14.502,
|
|
"eval_chemistry_steps_per_second": 7.251,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"eval_math_entropy": 0.8526870446205139,
|
|
"eval_math_loss": 0.8266505599021912,
|
|
"eval_math_mean_token_accuracy": 0.7845923402309418,
|
|
"eval_math_num_tokens": 20822188.0,
|
|
"eval_math_runtime": 34.9055,
|
|
"eval_math_samples_per_second": 14.324,
|
|
"eval_math_steps_per_second": 7.162,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"epoch": 1.2,
|
|
"eval_physics_entropy": 1.2596520464420318,
|
|
"eval_physics_loss": 1.2472213506698608,
|
|
"eval_physics_mean_token_accuracy": 0.715502037525177,
|
|
"eval_physics_num_tokens": 20822188.0,
|
|
"eval_physics_runtime": 40.5487,
|
|
"eval_physics_samples_per_second": 12.331,
|
|
"eval_physics_steps_per_second": 6.165,
|
|
"step": 1500
|
|
},
|
|
{
|
|
"entropy": 0.8487134363502264,
|
|
"epoch": 1.208,
|
|
"grad_norm": 0.6484375,
|
|
"learning_rate": 1.1044444444444444e-05,
|
|
"loss": 0.7823,
|
|
"mean_token_accuracy": 0.788084252178669,
|
|
"num_tokens": 20960481.0,
|
|
"step": 1510
|
|
},
|
|
{
|
|
"entropy": 0.8314280182123184,
|
|
"epoch": 1.216,
|
|
"grad_norm": 0.66796875,
|
|
"learning_rate": 1.0970370370370371e-05,
|
|
"loss": 0.7697,
|
|
"mean_token_accuracy": 0.7901332624256611,
|
|
"num_tokens": 21095959.0,
|
|
"step": 1520
|
|
},
|
|
{
|
|
"entropy": 0.8034210726618767,
|
|
"epoch": 1.224,
|
|
"grad_norm": 0.65234375,
|
|
"learning_rate": 1.0896296296296298e-05,
|
|
"loss": 0.7583,
|
|
"mean_token_accuracy": 0.7961955599486827,
|
|
"num_tokens": 21234422.0,
|
|
"step": 1530
|
|
},
|
|
{
|
|
"entropy": 0.8255144201219082,
|
|
"epoch": 1.232,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 1.0822222222222222e-05,
|
|
"loss": 0.7632,
|
|
"mean_token_accuracy": 0.7938656162470579,
|
|
"num_tokens": 21370052.0,
|
|
"step": 1540
|
|
},
|
|
{
|
|
"entropy": 0.8085383821278811,
|
|
"epoch": 1.24,
|
|
"grad_norm": 0.6640625,
|
|
"learning_rate": 1.074814814814815e-05,
|
|
"loss": 0.7552,
|
|
"mean_token_accuracy": 0.79493248462677,
|
|
"num_tokens": 21511515.0,
|
|
"step": 1550
|
|
},
|
|
{
|
|
"entropy": 0.824030945263803,
|
|
"epoch": 1.248,
|
|
"grad_norm": 0.6171875,
|
|
"learning_rate": 1.0674074074074074e-05,
|
|
"loss": 0.77,
|
|
"mean_token_accuracy": 0.792547807842493,
|
|
"num_tokens": 21648728.0,
|
|
"step": 1560
|
|
},
|
|
{
|
|
"entropy": 0.8271475210785866,
|
|
"epoch": 1.256,
|
|
"grad_norm": 0.6796875,
|
|
"learning_rate": 1.0600000000000002e-05,
|
|
"loss": 0.7797,
|
|
"mean_token_accuracy": 0.7907101519405841,
|
|
"num_tokens": 21789454.0,
|
|
"step": 1570
|
|
},
|
|
{
|
|
"entropy": 0.8608066206797957,
|
|
"epoch": 1.264,
|
|
"grad_norm": 0.6953125,
|
|
"learning_rate": 1.0525925925925927e-05,
|
|
"loss": 0.7857,
|
|
"mean_token_accuracy": 0.7850348219275475,
|
|
"num_tokens": 21927034.0,
|
|
"step": 1580
|
|
},
|
|
{
|
|
"entropy": 0.8113653404638171,
|
|
"epoch": 1.272,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.0451851851851853e-05,
|
|
"loss": 0.7717,
|
|
"mean_token_accuracy": 0.794109757617116,
|
|
"num_tokens": 22057647.0,
|
|
"step": 1590
|
|
},
|
|
{
|
|
"entropy": 0.8274726118892431,
|
|
"epoch": 1.28,
|
|
"grad_norm": 0.62890625,
|
|
"learning_rate": 1.0377777777777778e-05,
|
|
"loss": 0.7794,
|
|
"mean_token_accuracy": 0.790321846306324,
|
|
"num_tokens": 22199229.0,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.28,
|
|
"eval_biology_entropy": 2.8388440871238707,
|
|
"eval_biology_loss": 3.0385401248931885,
|
|
"eval_biology_mean_token_accuracy": 0.47107076036930084,
|
|
"eval_biology_num_tokens": 22199229.0,
|
|
"eval_biology_runtime": 26.9767,
|
|
"eval_biology_samples_per_second": 18.535,
|
|
"eval_biology_steps_per_second": 9.267,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.28,
|
|
"eval_chemistry_entropy": 1.6038966157436372,
|
|
"eval_chemistry_loss": 1.629550576210022,
|
|
"eval_chemistry_mean_token_accuracy": 0.6523506959676743,
|
|
"eval_chemistry_num_tokens": 22199229.0,
|
|
"eval_chemistry_runtime": 34.5012,
|
|
"eval_chemistry_samples_per_second": 14.492,
|
|
"eval_chemistry_steps_per_second": 7.246,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.28,
|
|
"eval_math_entropy": 0.8486473338603974,
|
|
"eval_math_loss": 0.8255683183670044,
|
|
"eval_math_mean_token_accuracy": 0.7848021316528321,
|
|
"eval_math_num_tokens": 22199229.0,
|
|
"eval_math_runtime": 34.9828,
|
|
"eval_math_samples_per_second": 14.293,
|
|
"eval_math_steps_per_second": 7.146,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"epoch": 1.28,
|
|
"eval_physics_entropy": 1.2539866659641266,
|
|
"eval_physics_loss": 1.2459049224853516,
|
|
"eval_physics_mean_token_accuracy": 0.7157720692157745,
|
|
"eval_physics_num_tokens": 22199229.0,
|
|
"eval_physics_runtime": 40.5655,
|
|
"eval_physics_samples_per_second": 12.326,
|
|
"eval_physics_steps_per_second": 6.163,
|
|
"step": 1600
|
|
},
|
|
{
|
|
"entropy": 0.8169244319200516,
|
|
"epoch": 1.288,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.0303703703703705e-05,
|
|
"loss": 0.7704,
|
|
"mean_token_accuracy": 0.7924644846469164,
|
|
"num_tokens": 22335695.0,
|
|
"step": 1610
|
|
},
|
|
{
|
|
"entropy": 0.8122199434787035,
|
|
"epoch": 1.296,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 1.0229629629629631e-05,
|
|
"loss": 0.7639,
|
|
"mean_token_accuracy": 0.7955255527049303,
|
|
"num_tokens": 22470502.0,
|
|
"step": 1620
|
|
},
|
|
{
|
|
"entropy": 0.8375904524698854,
|
|
"epoch": 1.304,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 1.0155555555555556e-05,
|
|
"loss": 0.7728,
|
|
"mean_token_accuracy": 0.7911740705370903,
|
|
"num_tokens": 22610780.0,
|
|
"step": 1630
|
|
},
|
|
{
|
|
"entropy": 0.8467503670603037,
|
|
"epoch": 1.312,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 1.0081481481481484e-05,
|
|
"loss": 0.8097,
|
|
"mean_token_accuracy": 0.7875058270990849,
|
|
"num_tokens": 22749763.0,
|
|
"step": 1640
|
|
},
|
|
{
|
|
"entropy": 0.8325416158884764,
|
|
"epoch": 1.32,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 1.0007407407407407e-05,
|
|
"loss": 0.7816,
|
|
"mean_token_accuracy": 0.7904935259371996,
|
|
"num_tokens": 22886557.0,
|
|
"step": 1650
|
|
},
|
|
{
|
|
"entropy": 0.8405437588691711,
|
|
"epoch": 1.328,
|
|
"grad_norm": 0.73828125,
|
|
"learning_rate": 9.933333333333334e-06,
|
|
"loss": 0.7953,
|
|
"mean_token_accuracy": 0.7865632444620132,
|
|
"num_tokens": 23020866.0,
|
|
"step": 1660
|
|
},
|
|
{
|
|
"entropy": 0.8233300801366568,
|
|
"epoch": 1.336,
|
|
"grad_norm": 0.75390625,
|
|
"learning_rate": 9.85925925925926e-06,
|
|
"loss": 0.7806,
|
|
"mean_token_accuracy": 0.7902509056031704,
|
|
"num_tokens": 23155782.0,
|
|
"step": 1670
|
|
},
|
|
{
|
|
"entropy": 0.7958553103730083,
|
|
"epoch": 1.3439999999999999,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 9.785185185185187e-06,
|
|
"loss": 0.7631,
|
|
"mean_token_accuracy": 0.797454484179616,
|
|
"num_tokens": 23291068.0,
|
|
"step": 1680
|
|
},
|
|
{
|
|
"entropy": 0.8144848184660077,
|
|
"epoch": 1.3519999999999999,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 9.711111111111111e-06,
|
|
"loss": 0.7765,
|
|
"mean_token_accuracy": 0.7929423958063125,
|
|
"num_tokens": 23430048.0,
|
|
"step": 1690
|
|
},
|
|
{
|
|
"entropy": 0.8195401934906841,
|
|
"epoch": 1.3599999999999999,
|
|
"grad_norm": 0.69921875,
|
|
"learning_rate": 9.637037037037038e-06,
|
|
"loss": 0.7735,
|
|
"mean_token_accuracy": 0.7930282630026341,
|
|
"num_tokens": 23571779.0,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 1.3599999999999999,
|
|
"eval_biology_entropy": 2.8502994742393493,
|
|
"eval_biology_loss": 3.0442328453063965,
|
|
"eval_biology_mean_token_accuracy": 0.47077508652210237,
|
|
"eval_biology_num_tokens": 23571779.0,
|
|
"eval_biology_runtime": 27.0342,
|
|
"eval_biology_samples_per_second": 18.495,
|
|
"eval_biology_steps_per_second": 9.248,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 1.3599999999999999,
|
|
"eval_chemistry_entropy": 1.604833221435547,
|
|
"eval_chemistry_loss": 1.6297847032546997,
|
|
"eval_chemistry_mean_token_accuracy": 0.65207393181324,
|
|
"eval_chemistry_num_tokens": 23571779.0,
|
|
"eval_chemistry_runtime": 34.6203,
|
|
"eval_chemistry_samples_per_second": 14.442,
|
|
"eval_chemistry_steps_per_second": 7.221,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 1.3599999999999999,
|
|
"eval_math_entropy": 0.8477920205593109,
|
|
"eval_math_loss": 0.8247544169425964,
|
|
"eval_math_mean_token_accuracy": 0.7847949225902557,
|
|
"eval_math_num_tokens": 23571779.0,
|
|
"eval_math_runtime": 35.0722,
|
|
"eval_math_samples_per_second": 14.256,
|
|
"eval_math_steps_per_second": 7.128,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"epoch": 1.3599999999999999,
|
|
"eval_physics_entropy": 1.2535700430870056,
|
|
"eval_physics_loss": 1.2455135583877563,
|
|
"eval_physics_mean_token_accuracy": 0.7158266956806183,
|
|
"eval_physics_num_tokens": 23571779.0,
|
|
"eval_physics_runtime": 40.6362,
|
|
"eval_physics_samples_per_second": 12.304,
|
|
"eval_physics_steps_per_second": 6.152,
|
|
"step": 1700
|
|
},
|
|
{
|
|
"entropy": 0.8383084131404758,
|
|
"epoch": 1.3679999999999999,
|
|
"grad_norm": 0.74609375,
|
|
"learning_rate": 9.562962962962965e-06,
|
|
"loss": 0.7956,
|
|
"mean_token_accuracy": 0.7880340579897165,
|
|
"num_tokens": 23707772.0,
|
|
"step": 1710
|
|
},
|
|
{
|
|
"entropy": 0.8829018753021955,
|
|
"epoch": 1.376,
|
|
"grad_norm": 0.62890625,
|
|
"learning_rate": 9.48888888888889e-06,
|
|
"loss": 0.8363,
|
|
"mean_token_accuracy": 0.7783376723527908,
|
|
"num_tokens": 23844985.0,
|
|
"step": 1720
|
|
},
|
|
{
|
|
"entropy": 0.8262542556971312,
|
|
"epoch": 1.384,
|
|
"grad_norm": 0.66796875,
|
|
"learning_rate": 9.414814814814816e-06,
|
|
"loss": 0.7657,
|
|
"mean_token_accuracy": 0.7913823142647743,
|
|
"num_tokens": 23981628.0,
|
|
"step": 1730
|
|
},
|
|
{
|
|
"entropy": 0.8316655196249485,
|
|
"epoch": 1.392,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 9.34074074074074e-06,
|
|
"loss": 0.7787,
|
|
"mean_token_accuracy": 0.7895013779401779,
|
|
"num_tokens": 24119732.0,
|
|
"step": 1740
|
|
},
|
|
{
|
|
"entropy": 0.7973068812862039,
|
|
"epoch": 1.4,
|
|
"grad_norm": 0.71484375,
|
|
"learning_rate": 9.266666666666667e-06,
|
|
"loss": 0.7521,
|
|
"mean_token_accuracy": 0.7974414359778166,
|
|
"num_tokens": 24262401.0,
|
|
"step": 1750
|
|
},
|
|
{
|
|
"entropy": 0.8182015813887119,
|
|
"epoch": 1.408,
|
|
"grad_norm": 0.76171875,
|
|
"learning_rate": 9.192592592592594e-06,
|
|
"loss": 0.7642,
|
|
"mean_token_accuracy": 0.7919386733323336,
|
|
"num_tokens": 24401537.0,
|
|
"step": 1760
|
|
},
|
|
{
|
|
"entropy": 0.8352840909734368,
|
|
"epoch": 1.416,
|
|
"grad_norm": 0.71875,
|
|
"learning_rate": 9.118518518518518e-06,
|
|
"loss": 0.7858,
|
|
"mean_token_accuracy": 0.7885801542550326,
|
|
"num_tokens": 24540281.0,
|
|
"step": 1770
|
|
},
|
|
{
|
|
"entropy": 0.7927129395306111,
|
|
"epoch": 1.424,
|
|
"grad_norm": 0.60546875,
|
|
"learning_rate": 9.044444444444445e-06,
|
|
"loss": 0.742,
|
|
"mean_token_accuracy": 0.7989379655569792,
|
|
"num_tokens": 24682390.0,
|
|
"step": 1780
|
|
},
|
|
{
|
|
"entropy": 0.8145547799766064,
|
|
"epoch": 1.432,
|
|
"grad_norm": 0.640625,
|
|
"learning_rate": 8.970370370370372e-06,
|
|
"loss": 0.7624,
|
|
"mean_token_accuracy": 0.7933211088180542,
|
|
"num_tokens": 24820884.0,
|
|
"step": 1790
|
|
},
|
|
{
|
|
"entropy": 0.8423224657773971,
|
|
"epoch": 1.44,
|
|
"grad_norm": 0.671875,
|
|
"learning_rate": 8.896296296296298e-06,
|
|
"loss": 0.7958,
|
|
"mean_token_accuracy": 0.7877123914659023,
|
|
"num_tokens": 24957390.0,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"eval_biology_entropy": 2.8344242753982543,
|
|
"eval_biology_loss": 3.0405893325805664,
|
|
"eval_biology_mean_token_accuracy": 0.47124050533771517,
|
|
"eval_biology_num_tokens": 24957390.0,
|
|
"eval_biology_runtime": 27.0536,
|
|
"eval_biology_samples_per_second": 18.482,
|
|
"eval_biology_steps_per_second": 9.241,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"eval_chemistry_entropy": 1.5973036425113678,
|
|
"eval_chemistry_loss": 1.6277400255203247,
|
|
"eval_chemistry_mean_token_accuracy": 0.6527108114957809,
|
|
"eval_chemistry_num_tokens": 24957390.0,
|
|
"eval_chemistry_runtime": 34.5577,
|
|
"eval_chemistry_samples_per_second": 14.469,
|
|
"eval_chemistry_steps_per_second": 7.234,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"eval_math_entropy": 0.8444431309700012,
|
|
"eval_math_loss": 0.823908805847168,
|
|
"eval_math_mean_token_accuracy": 0.7849345812797547,
|
|
"eval_math_num_tokens": 24957390.0,
|
|
"eval_math_runtime": 35.0263,
|
|
"eval_math_samples_per_second": 14.275,
|
|
"eval_math_steps_per_second": 7.137,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"epoch": 1.44,
|
|
"eval_physics_entropy": 1.2485797266960144,
|
|
"eval_physics_loss": 1.2440050840377808,
|
|
"eval_physics_mean_token_accuracy": 0.7161458423137664,
|
|
"eval_physics_num_tokens": 24957390.0,
|
|
"eval_physics_runtime": 40.7709,
|
|
"eval_physics_samples_per_second": 12.264,
|
|
"eval_physics_steps_per_second": 6.132,
|
|
"step": 1800
|
|
},
|
|
{
|
|
"entropy": 0.8316453367471695,
|
|
"epoch": 1.448,
|
|
"grad_norm": 0.703125,
|
|
"learning_rate": 8.822222222222223e-06,
|
|
"loss": 0.7816,
|
|
"mean_token_accuracy": 0.7880163863301277,
|
|
"num_tokens": 25091961.0,
|
|
"step": 1810
|
|
},
|
|
{
|
|
"entropy": 0.8137075098231434,
|
|
"epoch": 1.456,
|
|
"grad_norm": 0.67578125,
|
|
"learning_rate": 8.74814814814815e-06,
|
|
"loss": 0.7661,
|
|
"mean_token_accuracy": 0.7935271102935075,
|
|
"num_tokens": 25229390.0,
|
|
"step": 1820
|
|
},
|
|
{
|
|
"entropy": 0.8185559125617147,
|
|
"epoch": 1.464,
|
|
"grad_norm": 0.6015625,
|
|
"learning_rate": 8.674074074074074e-06,
|
|
"loss": 0.7898,
|
|
"mean_token_accuracy": 0.790428926050663,
|
|
"num_tokens": 25365072.0,
|
|
"step": 1830
|
|
},
|
|
{
|
|
"entropy": 0.8206291127949953,
|
|
"epoch": 1.472,
|
|
"grad_norm": 0.5703125,
|
|
"learning_rate": 8.6e-06,
|
|
"loss": 0.7714,
|
|
"mean_token_accuracy": 0.7919356789439916,
|
|
"num_tokens": 25499263.0,
|
|
"step": 1840
|
|
},
|
|
{
|
|
"entropy": 0.833819380402565,
|
|
"epoch": 1.48,
|
|
"grad_norm": 0.69921875,
|
|
"learning_rate": 8.525925925925927e-06,
|
|
"loss": 0.7875,
|
|
"mean_token_accuracy": 0.7891976829618216,
|
|
"num_tokens": 25630695.0,
|
|
"step": 1850
|
|
},
|
|
{
|
|
"entropy": 0.860048278607428,
|
|
"epoch": 1.488,
|
|
"grad_norm": 0.765625,
|
|
"learning_rate": 8.451851851851852e-06,
|
|
"loss": 0.8185,
|
|
"mean_token_accuracy": 0.7816774118691683,
|
|
"num_tokens": 25766470.0,
|
|
"step": 1860
|
|
},
|
|
{
|
|
"entropy": 0.8099161019548774,
|
|
"epoch": 1.496,
|
|
"grad_norm": 0.66015625,
|
|
"learning_rate": 8.377777777777779e-06,
|
|
"loss": 0.7587,
|
|
"mean_token_accuracy": 0.7959996823221445,
|
|
"num_tokens": 25909007.0,
|
|
"step": 1870
|
|
},
|
|
{
|
|
"entropy": 0.8094545688480139,
|
|
"epoch": 1.504,
|
|
"grad_norm": 0.734375,
|
|
"learning_rate": 8.303703703703705e-06,
|
|
"loss": 0.7604,
|
|
"mean_token_accuracy": 0.7959994535893202,
|
|
"num_tokens": 26051351.0,
|
|
"step": 1880
|
|
},
|
|
{
|
|
"entropy": 0.8493411887437106,
|
|
"epoch": 1.512,
|
|
"grad_norm": 0.68359375,
|
|
"learning_rate": 8.229629629629632e-06,
|
|
"loss": 0.8068,
|
|
"mean_token_accuracy": 0.7882944524288178,
|
|
"num_tokens": 26188237.0,
|
|
"step": 1890
|
|
},
|
|
{
|
|
"entropy": 0.8235599439591169,
|
|
"epoch": 1.52,
|
|
"grad_norm": 0.62890625,
|
|
"learning_rate": 8.155555555555556e-06,
|
|
"loss": 0.7783,
|
|
"mean_token_accuracy": 0.79057194031775,
|
|
"num_tokens": 26331956.0,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 1.52,
|
|
"eval_biology_entropy": 2.8468204393386842,
|
|
"eval_biology_loss": 3.039438009262085,
|
|
"eval_biology_mean_token_accuracy": 0.4713715660572052,
|
|
"eval_biology_num_tokens": 26331956.0,
|
|
"eval_biology_runtime": 26.9649,
|
|
"eval_biology_samples_per_second": 18.543,
|
|
"eval_biology_steps_per_second": 9.271,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 1.52,
|
|
"eval_chemistry_entropy": 1.6046333646774291,
|
|
"eval_chemistry_loss": 1.6270935535430908,
|
|
"eval_chemistry_mean_token_accuracy": 0.6528140946626663,
|
|
"eval_chemistry_num_tokens": 26331956.0,
|
|
"eval_chemistry_runtime": 34.4714,
|
|
"eval_chemistry_samples_per_second": 14.505,
|
|
"eval_chemistry_steps_per_second": 7.252,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 1.52,
|
|
"eval_math_entropy": 0.8492136422395706,
|
|
"eval_math_loss": 0.8234105110168457,
|
|
"eval_math_mean_token_accuracy": 0.7850139417648315,
|
|
"eval_math_num_tokens": 26331956.0,
|
|
"eval_math_runtime": 34.9296,
|
|
"eval_math_samples_per_second": 14.314,
|
|
"eval_math_steps_per_second": 7.157,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"epoch": 1.52,
|
|
"eval_physics_entropy": 1.254153406381607,
|
|
"eval_physics_loss": 1.2435050010681152,
|
|
"eval_physics_mean_token_accuracy": 0.7162581722736359,
|
|
"eval_physics_num_tokens": 26331956.0,
|
|
"eval_physics_runtime": 40.5372,
|
|
"eval_physics_samples_per_second": 12.334,
|
|
"eval_physics_steps_per_second": 6.167,
|
|
"step": 1900
|
|
},
|
|
{
|
|
"entropy": 0.8410981273278594,
|
|
"epoch": 1.528,
|
|
"grad_norm": 0.70703125,
|
|
"learning_rate": 8.081481481481483e-06,
|
|
"loss": 0.7966,
|
|
"mean_token_accuracy": 0.7865708086639642,
|
|
"num_tokens": 26469780.0,
|
|
"step": 1910
|
|
},
|
|
{
|
|
"entropy": 0.8120917391031981,
|
|
"epoch": 1.536,
|
|
"grad_norm": 0.6796875,
|
|
"learning_rate": 8.007407407407408e-06,
|
|
"loss": 0.7679,
|
|
"mean_token_accuracy": 0.7950867023319006,
|
|
"num_tokens": 26613031.0,
|
|
"step": 1920
|
|
},
|
|
{
|
|
"entropy": 0.813549479842186,
|
|
"epoch": 1.544,
|
|
"grad_norm": 0.63671875,
|
|
"learning_rate": 7.933333333333334e-06,
|
|
"loss": 0.7604,
|
|
"mean_token_accuracy": 0.7951221365481616,
|
|
"num_tokens": 26748931.0,
|
|
"step": 1930
|
|
},
|
|
{
|
|
"entropy": 0.8368369644507766,
|
|
"epoch": 1.552,
|
|
"grad_norm": 0.65625,
|
|
"learning_rate": 7.859259259259259e-06,
|
|
"loss": 0.7908,
|
|
"mean_token_accuracy": 0.7891507402062417,
|
|
"num_tokens": 26892482.0,
|
|
"step": 1940
|
|
},
|
|
{
|
|
"entropy": 0.8099740859121084,
|
|
"epoch": 1.56,
|
|
"grad_norm": 0.7890625,
|
|
"learning_rate": 7.785185185185185e-06,
|
|
"loss": 0.7465,
|
|
"mean_token_accuracy": 0.7952026624232531,
|
|
"num_tokens": 27033385.0,
|
|
"step": 1950
|
|
},
|
|
{
|
|
"entropy": 0.8125249484553934,
|
|
"epoch": 1.568,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 7.711111111111112e-06,
|
|
"loss": 0.7716,
|
|
"mean_token_accuracy": 0.795727027580142,
|
|
"num_tokens": 27177253.0,
|
|
"step": 1960
|
|
},
|
|
{
|
|
"entropy": 0.8376018533483147,
|
|
"epoch": 1.576,
|
|
"grad_norm": 0.7109375,
|
|
"learning_rate": 7.637037037037037e-06,
|
|
"loss": 0.7807,
|
|
"mean_token_accuracy": 0.7907413702458144,
|
|
"num_tokens": 27314645.0,
|
|
"step": 1970
|
|
},
|
|
{
|
|
"entropy": 0.8189013006165624,
|
|
"epoch": 1.584,
|
|
"grad_norm": 0.87109375,
|
|
"learning_rate": 7.562962962962963e-06,
|
|
"loss": 0.771,
|
|
"mean_token_accuracy": 0.7916807536035776,
|
|
"num_tokens": 27456524.0,
|
|
"step": 1980
|
|
},
|
|
{
|
|
"entropy": 0.8082204090431333,
|
|
"epoch": 1.592,
|
|
"grad_norm": 0.6328125,
|
|
"learning_rate": 7.48888888888889e-06,
|
|
"loss": 0.7578,
|
|
"mean_token_accuracy": 0.7939547438174486,
|
|
"num_tokens": 27592021.0,
|
|
"step": 1990
|
|
},
|
|
{
|
|
"entropy": 0.8296483233571053,
|
|
"epoch": 1.6,
|
|
"grad_norm": 0.62109375,
|
|
"learning_rate": 7.4148148148148155e-06,
|
|
"loss": 0.7837,
|
|
"mean_token_accuracy": 0.7896194510161877,
|
|
"num_tokens": 27734080.0,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"eval_biology_entropy": 2.848469418525696,
|
|
"eval_biology_loss": 3.041050672531128,
|
|
"eval_biology_mean_token_accuracy": 0.4710944994688034,
|
|
"eval_biology_num_tokens": 27734080.0,
|
|
"eval_biology_runtime": 26.9667,
|
|
"eval_biology_samples_per_second": 18.541,
|
|
"eval_biology_steps_per_second": 9.271,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"eval_chemistry_entropy": 1.6052903940677643,
|
|
"eval_chemistry_loss": 1.6265658140182495,
|
|
"eval_chemistry_mean_token_accuracy": 0.6528225523233414,
|
|
"eval_chemistry_num_tokens": 27734080.0,
|
|
"eval_chemistry_runtime": 34.5552,
|
|
"eval_chemistry_samples_per_second": 14.47,
|
|
"eval_chemistry_steps_per_second": 7.235,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"eval_math_entropy": 0.8491179572343827,
|
|
"eval_math_loss": 0.8228975534439087,
|
|
"eval_math_mean_token_accuracy": 0.7851616134643554,
|
|
"eval_math_num_tokens": 27734080.0,
|
|
"eval_math_runtime": 35.0391,
|
|
"eval_math_samples_per_second": 14.27,
|
|
"eval_math_steps_per_second": 7.135,
|
|
"step": 2000
|
|
},
|
|
{
|
|
"epoch": 1.6,
|
|
"eval_physics_entropy": 1.2549104006290437,
|
|
"eval_physics_loss": 1.2431191205978394,
|
|
"eval_physics_mean_token_accuracy": 0.7163957259654998,
|
|
"eval_physics_num_tokens": 27734080.0,
|
|
"eval_physics_runtime": 40.6487,
|
|
"eval_physics_samples_per_second": 12.301,
|
|
"eval_physics_steps_per_second": 6.15,
|
|
"step": 2000
|
|
}
|
|
],
|
|
"logging_steps": 10,
|
|
"max_steps": 3000,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 3,
|
|
"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": 1.4559841566518477e+18,
|
|
"train_batch_size": 2,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|