{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 33.333333333333336,
"eval_steps": 50,
"global_step": 1500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.1111111111111112,
"grad_norm": 2.6699063777923584,
"learning_rate": 2.443609022556391e-05,
"loss": 1.8217,
"step": 50
},
{
"epoch": 1.1111111111111112,
"eval_loss": 1.0424705743789673,
"eval_runtime": 4.0916,
"eval_samples_per_second": 43.993,
"eval_steps_per_second": 5.621,
"step": 50
},
{
"epoch": 2.2222222222222223,
"grad_norm": 0.44133517146110535,
"learning_rate": 2.380952380952381e-05,
"loss": 0.5439,
"step": 100
},
{
"epoch": 2.2222222222222223,
"eval_loss": 0.3811015784740448,
"eval_runtime": 4.0749,
"eval_samples_per_second": 44.173,
"eval_steps_per_second": 5.644,
"step": 100
},
{
"epoch": 3.3333333333333335,
"grad_norm": 0.540984034538269,
"learning_rate": 2.3182957393483708e-05,
"loss": 0.3496,
"step": 150
},
{
"epoch": 3.3333333333333335,
"eval_loss": 0.3363141715526581,
"eval_runtime": 4.089,
"eval_samples_per_second": 44.02,
"eval_steps_per_second": 5.625,
"step": 150
},
{
"epoch": 4.444444444444445,
"grad_norm": 0.4714152216911316,
"learning_rate": 2.255639097744361e-05,
"loss": 0.3178,
"step": 200
},
{
"epoch": 4.444444444444445,
"eval_loss": 0.3166368007659912,
"eval_runtime": 4.0511,
"eval_samples_per_second": 44.432,
"eval_steps_per_second": 5.677,
"step": 200
},
{
"epoch": 5.555555555555555,
"grad_norm": 0.5102563500404358,
"learning_rate": 2.1929824561403507e-05,
"loss": 0.2971,
"step": 250
},
{
"epoch": 5.555555555555555,
"eval_loss": 0.304388165473938,
"eval_runtime": 4.0541,
"eval_samples_per_second": 44.399,
"eval_steps_per_second": 5.673,
"step": 250
},
{
"epoch": 6.666666666666667,
"grad_norm": 0.8356263041496277,
"learning_rate": 2.130325814536341e-05,
"loss": 0.2825,
"step": 300
},
{
"epoch": 6.666666666666667,
"eval_loss": 0.29874512553215027,
"eval_runtime": 4.0506,
"eval_samples_per_second": 44.438,
"eval_steps_per_second": 5.678,
"step": 300
},
{
"epoch": 7.777777777777778,
"grad_norm": 0.694858968257904,
"learning_rate": 2.067669172932331e-05,
"loss": 0.2714,
"step": 350
},
{
"epoch": 7.777777777777778,
"eval_loss": 0.2942558825016022,
"eval_runtime": 4.074,
"eval_samples_per_second": 44.182,
"eval_steps_per_second": 5.646,
"step": 350
},
{
"epoch": 8.88888888888889,
"grad_norm": 0.7505309581756592,
"learning_rate": 2.0050125313283208e-05,
"loss": 0.2616,
"step": 400
},
{
"epoch": 8.88888888888889,
"eval_loss": 0.29128891229629517,
"eval_runtime": 4.0871,
"eval_samples_per_second": 44.041,
"eval_steps_per_second": 5.627,
"step": 400
},
{
"epoch": 10.0,
"grad_norm": 0.7603775858879089,
"learning_rate": 1.942355889724311e-05,
"loss": 0.2552,
"step": 450
},
{
"epoch": 10.0,
"eval_loss": 0.28786608576774597,
"eval_runtime": 4.052,
"eval_samples_per_second": 44.422,
"eval_steps_per_second": 5.676,
"step": 450
},
{
"epoch": 11.11111111111111,
"grad_norm": 0.831240177154541,
"learning_rate": 1.8796992481203007e-05,
"loss": 0.2472,
"step": 500
},
{
"epoch": 11.11111111111111,
"eval_loss": 0.28864821791648865,
"eval_runtime": 4.0371,
"eval_samples_per_second": 44.586,
"eval_steps_per_second": 5.697,
"step": 500
},
{
"epoch": 12.222222222222221,
"grad_norm": 0.7524340152740479,
"learning_rate": 1.8170426065162908e-05,
"loss": 0.2388,
"step": 550
},
{
"epoch": 12.222222222222221,
"eval_loss": 0.28881773352622986,
"eval_runtime": 4.0933,
"eval_samples_per_second": 43.974,
"eval_steps_per_second": 5.619,
"step": 550
},
{
"epoch": 13.333333333333334,
"grad_norm": 0.8262473940849304,
"learning_rate": 1.7543859649122806e-05,
"loss": 0.2309,
"step": 600
},
{
"epoch": 13.333333333333334,
"eval_loss": 0.2914719581604004,
"eval_runtime": 4.0773,
"eval_samples_per_second": 44.147,
"eval_steps_per_second": 5.641,
"step": 600
},
{
"epoch": 14.444444444444445,
"grad_norm": 0.8302382826805115,
"learning_rate": 1.6917293233082707e-05,
"loss": 0.2263,
"step": 650
},
{
"epoch": 14.444444444444445,
"eval_loss": 0.2899852395057678,
"eval_runtime": 4.0443,
"eval_samples_per_second": 44.507,
"eval_steps_per_second": 5.687,
"step": 650
},
{
"epoch": 15.555555555555555,
"grad_norm": 0.957399845123291,
"learning_rate": 1.6290726817042605e-05,
"loss": 0.2181,
"step": 700
},
{
"epoch": 15.555555555555555,
"eval_loss": 0.29406097531318665,
"eval_runtime": 4.0913,
"eval_samples_per_second": 43.996,
"eval_steps_per_second": 5.622,
"step": 700
},
{
"epoch": 16.666666666666668,
"grad_norm": 1.0171995162963867,
"learning_rate": 1.5664160401002506e-05,
"loss": 0.2115,
"step": 750
},
{
"epoch": 16.666666666666668,
"eval_loss": 0.2935585081577301,
"eval_runtime": 4.0514,
"eval_samples_per_second": 44.429,
"eval_steps_per_second": 5.677,
"step": 750
},
{
"epoch": 17.77777777777778,
"grad_norm": 0.9579535722732544,
"learning_rate": 1.5037593984962406e-05,
"loss": 0.2056,
"step": 800
},
{
"epoch": 17.77777777777778,
"eval_loss": 0.2986494302749634,
"eval_runtime": 4.0426,
"eval_samples_per_second": 44.526,
"eval_steps_per_second": 5.689,
"step": 800
},
{
"epoch": 18.88888888888889,
"grad_norm": 1.0283973217010498,
"learning_rate": 1.4411027568922305e-05,
"loss": 0.1983,
"step": 850
},
{
"epoch": 18.88888888888889,
"eval_loss": 0.30257824063301086,
"eval_runtime": 4.0596,
"eval_samples_per_second": 44.339,
"eval_steps_per_second": 5.666,
"step": 850
},
{
"epoch": 20.0,
"grad_norm": 1.068368673324585,
"learning_rate": 1.3784461152882205e-05,
"loss": 0.1915,
"step": 900
},
{
"epoch": 20.0,
"eval_loss": 0.30601659417152405,
"eval_runtime": 4.0465,
"eval_samples_per_second": 44.483,
"eval_steps_per_second": 5.684,
"step": 900
},
{
"epoch": 21.11111111111111,
"grad_norm": 0.987966001033783,
"learning_rate": 1.3157894736842106e-05,
"loss": 0.1834,
"step": 950
},
{
"epoch": 21.11111111111111,
"eval_loss": 0.3163508176803589,
"eval_runtime": 4.0462,
"eval_samples_per_second": 44.486,
"eval_steps_per_second": 5.684,
"step": 950
},
{
"epoch": 22.22222222222222,
"grad_norm": 1.046950101852417,
"learning_rate": 1.2531328320802006e-05,
"loss": 0.1773,
"step": 1000
},
{
"epoch": 22.22222222222222,
"eval_loss": 0.32225680351257324,
"eval_runtime": 4.0814,
"eval_samples_per_second": 44.102,
"eval_steps_per_second": 5.635,
"step": 1000
},
{
"epoch": 23.333333333333332,
"grad_norm": 1.2292746305465698,
"learning_rate": 1.1904761904761905e-05,
"loss": 0.1707,
"step": 1050
},
{
"epoch": 23.333333333333332,
"eval_loss": 0.3270200490951538,
"eval_runtime": 4.0696,
"eval_samples_per_second": 44.23,
"eval_steps_per_second": 5.652,
"step": 1050
},
{
"epoch": 24.444444444444443,
"grad_norm": 1.107651948928833,
"learning_rate": 1.1278195488721805e-05,
"loss": 0.1641,
"step": 1100
},
{
"epoch": 24.444444444444443,
"eval_loss": 0.3326013684272766,
"eval_runtime": 4.0592,
"eval_samples_per_second": 44.343,
"eval_steps_per_second": 5.666,
"step": 1100
},
{
"epoch": 25.555555555555557,
"grad_norm": 1.3203223943710327,
"learning_rate": 1.0651629072681704e-05,
"loss": 0.1575,
"step": 1150
},
{
"epoch": 25.555555555555557,
"eval_loss": 0.33997443318367004,
"eval_runtime": 4.0804,
"eval_samples_per_second": 44.113,
"eval_steps_per_second": 5.637,
"step": 1150
},
{
"epoch": 26.666666666666668,
"grad_norm": 1.2884823083877563,
"learning_rate": 1.0025062656641604e-05,
"loss": 0.1517,
"step": 1200
},
{
"epoch": 26.666666666666668,
"eval_loss": 0.3428906202316284,
"eval_runtime": 4.0803,
"eval_samples_per_second": 44.114,
"eval_steps_per_second": 5.637,
"step": 1200
},
{
"epoch": 27.77777777777778,
"grad_norm": 1.4833685159683228,
"learning_rate": 9.398496240601503e-06,
"loss": 0.1442,
"step": 1250
},
{
"epoch": 27.77777777777778,
"eval_loss": 0.35523876547813416,
"eval_runtime": 4.0449,
"eval_samples_per_second": 44.5,
"eval_steps_per_second": 5.686,
"step": 1250
},
{
"epoch": 28.88888888888889,
"grad_norm": 1.4259836673736572,
"learning_rate": 8.771929824561403e-06,
"loss": 0.1397,
"step": 1300
},
{
"epoch": 28.88888888888889,
"eval_loss": 0.36296555399894714,
"eval_runtime": 4.0386,
"eval_samples_per_second": 44.57,
"eval_steps_per_second": 5.695,
"step": 1300
},
{
"epoch": 30.0,
"grad_norm": 1.5832905769348145,
"learning_rate": 8.145363408521302e-06,
"loss": 0.1345,
"step": 1350
},
{
"epoch": 30.0,
"eval_loss": 0.3718196153640747,
"eval_runtime": 4.0526,
"eval_samples_per_second": 44.416,
"eval_steps_per_second": 5.675,
"step": 1350
},
{
"epoch": 31.11111111111111,
"grad_norm": 1.5873230695724487,
"learning_rate": 7.518796992481203e-06,
"loss": 0.1289,
"step": 1400
},
{
"epoch": 31.11111111111111,
"eval_loss": 0.38369277119636536,
"eval_runtime": 4.0471,
"eval_samples_per_second": 44.476,
"eval_steps_per_second": 5.683,
"step": 1400
},
{
"epoch": 32.22222222222222,
"grad_norm": 1.3492352962493896,
"learning_rate": 6.892230576441102e-06,
"loss": 0.1234,
"step": 1450
},
{
"epoch": 32.22222222222222,
"eval_loss": 0.39153799414634705,
"eval_runtime": 4.0507,
"eval_samples_per_second": 44.436,
"eval_steps_per_second": 5.678,
"step": 1450
},
{
"epoch": 33.333333333333336,
"grad_norm": 1.5531470775604248,
"learning_rate": 6.265664160401003e-06,
"loss": 0.1194,
"step": 1500
},
{
"epoch": 33.333333333333336,
"eval_loss": 0.3885114789009094,
"eval_runtime": 4.0759,
"eval_samples_per_second": 44.162,
"eval_steps_per_second": 5.643,
"step": 1500
}
],
"logging_steps": 50,
"max_steps": 2000,
"num_input_tokens_seen": 0,
"num_train_epochs": 45,
"save_steps": 25,
"total_flos": 3.996628758124954e+16,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}