236 lines
6.1 KiB
JSON
236 lines
6.1 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 1.0,
|
|
"eval_steps": 99999999,
|
|
"global_step": 3189,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"epoch": 0.0003136639874534405,
|
|
"grad_norm": 174.2938690185547,
|
|
"learning_rate": 5e-05,
|
|
"loss": 175.3938,
|
|
"step": 1
|
|
},
|
|
{
|
|
"epoch": 0.0003136639874534405,
|
|
"eval_accuracy": 0.038034240719819076,
|
|
"eval_loss": 10.384196281433105,
|
|
"eval_runtime": 194.5696,
|
|
"eval_samples_per_second": 29.83,
|
|
"eval_steps_per_second": 3.731,
|
|
"step": 1
|
|
},
|
|
{
|
|
"epoch": 0.000627327974906881,
|
|
"grad_norm": 83.63700103759766,
|
|
"learning_rate": 4.99843211037943e-05,
|
|
"loss": 164.5011,
|
|
"step": 2
|
|
},
|
|
{
|
|
"epoch": 0.000627327974906881,
|
|
"eval_accuracy": 0.038057314435118395,
|
|
"eval_loss": 10.165145874023438,
|
|
"eval_runtime": 259.579,
|
|
"eval_samples_per_second": 22.359,
|
|
"eval_steps_per_second": 2.797,
|
|
"step": 2
|
|
},
|
|
{
|
|
"epoch": 0.001254655949813762,
|
|
"grad_norm": 43.825782775878906,
|
|
"learning_rate": 4.9952963311382884e-05,
|
|
"loss": 158.7612,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.001254655949813762,
|
|
"eval_accuracy": 0.038057314435118395,
|
|
"eval_loss": 10.000017166137695,
|
|
"eval_runtime": 258.7681,
|
|
"eval_samples_per_second": 22.429,
|
|
"eval_steps_per_second": 2.806,
|
|
"step": 4
|
|
},
|
|
{
|
|
"epoch": 0.002509311899627524,
|
|
"grad_norm": 35.26978302001953,
|
|
"learning_rate": 4.9890247726560055e-05,
|
|
"loss": 155.0138,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.002509311899627524,
|
|
"eval_accuracy": 0.0483111387771133,
|
|
"eval_loss": 9.72864055633545,
|
|
"eval_runtime": 309.1294,
|
|
"eval_samples_per_second": 18.775,
|
|
"eval_steps_per_second": 2.349,
|
|
"step": 8
|
|
},
|
|
{
|
|
"epoch": 0.005018623799255048,
|
|
"grad_norm": 40.16830825805664,
|
|
"learning_rate": 4.97648165569144e-05,
|
|
"loss": 149.2228,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.005018623799255048,
|
|
"eval_accuracy": 0.033622108459261925,
|
|
"eval_loss": 11.213191986083984,
|
|
"eval_runtime": 112.2816,
|
|
"eval_samples_per_second": 51.691,
|
|
"eval_steps_per_second": 6.466,
|
|
"step": 16
|
|
},
|
|
{
|
|
"epoch": 0.010037247598510097,
|
|
"grad_norm": 37.20323944091797,
|
|
"learning_rate": 4.951395421762308e-05,
|
|
"loss": 145.2656,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.010037247598510097,
|
|
"eval_accuracy": 0.06755377523034978,
|
|
"eval_loss": 8.518936157226562,
|
|
"eval_runtime": 113.0533,
|
|
"eval_samples_per_second": 51.339,
|
|
"eval_steps_per_second": 6.422,
|
|
"step": 32
|
|
},
|
|
{
|
|
"epoch": 0.020074495197020193,
|
|
"grad_norm": 24.089946746826172,
|
|
"learning_rate": 4.9012229539040455e-05,
|
|
"loss": 115.6545,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.020074495197020193,
|
|
"eval_accuracy": 0.0989353754076637,
|
|
"eval_loss": 7.334339618682861,
|
|
"eval_runtime": 118.0378,
|
|
"eval_samples_per_second": 49.171,
|
|
"eval_steps_per_second": 6.151,
|
|
"step": 64
|
|
},
|
|
{
|
|
"epoch": 0.040148990394040386,
|
|
"grad_norm": 13.746461868286133,
|
|
"learning_rate": 4.80087801818752e-05,
|
|
"loss": 93.3162,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.040148990394040386,
|
|
"eval_accuracy": 0.12674374971789434,
|
|
"eval_loss": 6.555325508117676,
|
|
"eval_runtime": 111.1576,
|
|
"eval_samples_per_second": 52.214,
|
|
"eval_steps_per_second": 6.531,
|
|
"step": 128
|
|
},
|
|
{
|
|
"epoch": 0.08029798078808077,
|
|
"grad_norm": 13.765669822692871,
|
|
"learning_rate": 4.600188146754469e-05,
|
|
"loss": 77.7084,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 0.08029798078808077,
|
|
"eval_accuracy": 0.16966187070231,
|
|
"eval_loss": 5.74245548248291,
|
|
"eval_runtime": 112.901,
|
|
"eval_samples_per_second": 51.408,
|
|
"eval_steps_per_second": 6.43,
|
|
"step": 256
|
|
},
|
|
{
|
|
"epoch": 0.16059596157616154,
|
|
"grad_norm": 10.164519309997559,
|
|
"learning_rate": 4.1988084038883666e-05,
|
|
"loss": 71.3208,
|
|
"step": 512
|
|
},
|
|
{
|
|
"epoch": 0.16059596157616154,
|
|
"eval_accuracy": 0.1885688435453892,
|
|
"eval_loss": 5.244011878967285,
|
|
"eval_runtime": 116.5825,
|
|
"eval_samples_per_second": 49.784,
|
|
"eval_steps_per_second": 6.227,
|
|
"step": 512
|
|
},
|
|
{
|
|
"epoch": 0.3211919231523231,
|
|
"grad_norm": 6.641696453094482,
|
|
"learning_rate": 3.3960489181561625e-05,
|
|
"loss": 71.3612,
|
|
"step": 1024
|
|
},
|
|
{
|
|
"epoch": 0.3211919231523231,
|
|
"eval_accuracy": 0.18997667702120694,
|
|
"eval_loss": 4.9998674392700195,
|
|
"eval_runtime": 112.5118,
|
|
"eval_samples_per_second": 51.586,
|
|
"eval_steps_per_second": 6.453,
|
|
"step": 1024
|
|
},
|
|
{
|
|
"epoch": 0.6423838463046462,
|
|
"grad_norm": 12.357915878295898,
|
|
"learning_rate": 1.790529946691753e-05,
|
|
"loss": 80.1318,
|
|
"step": 2048
|
|
},
|
|
{
|
|
"epoch": 0.6423838463046462,
|
|
"eval_accuracy": 0.20517383433948205,
|
|
"eval_loss": 4.625439643859863,
|
|
"eval_runtime": 116.1969,
|
|
"eval_samples_per_second": 49.95,
|
|
"eval_steps_per_second": 6.248,
|
|
"step": 2048
|
|
},
|
|
{
|
|
"epoch": 1.0,
|
|
"step": 3189,
|
|
"total_flos": 2.132534919168e+17,
|
|
"train_loss": 81.87549413786894,
|
|
"train_runtime": 16625.0874,
|
|
"train_samples_per_second": 24.546,
|
|
"train_steps_per_second": 0.192
|
|
}
|
|
],
|
|
"logging_steps": 99999999,
|
|
"max_steps": 3189,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 1,
|
|
"save_steps": 99999999,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": true
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 2.132534919168e+17,
|
|
"train_batch_size": 8,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|