|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 20.0, |
|
"eval_steps": 500, |
|
"global_step": 2440, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 4.303664207458496, |
|
"learning_rate": 4.75e-05, |
|
"loss": 0.5535, |
|
"step": 122 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.7293233082706767, |
|
"eval_f1": 0.6372727272727272, |
|
"eval_loss": 0.49923330545425415, |
|
"eval_precision": 0.6645702306079665, |
|
"eval_recall": 0.6284779050736498, |
|
"eval_runtime": 5.2819, |
|
"eval_samples_per_second": 75.541, |
|
"eval_steps_per_second": 9.466, |
|
"step": 122 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 4.080423831939697, |
|
"learning_rate": 4.5e-05, |
|
"loss": 0.444, |
|
"step": 244 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.8170426065162907, |
|
"eval_f1": 0.7960536910871955, |
|
"eval_loss": 0.4052737355232239, |
|
"eval_precision": 0.7846938775510204, |
|
"eval_recall": 0.8255591925804692, |
|
"eval_runtime": 5.5544, |
|
"eval_samples_per_second": 71.835, |
|
"eval_steps_per_second": 9.002, |
|
"step": 244 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 4.110426425933838, |
|
"learning_rate": 4.25e-05, |
|
"loss": 0.3464, |
|
"step": 366 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.8421052631578947, |
|
"eval_f1": 0.7905197629940748, |
|
"eval_loss": 0.3424628674983978, |
|
"eval_precision": 0.8345238095238094, |
|
"eval_recall": 0.7682760501909438, |
|
"eval_runtime": 5.6127, |
|
"eval_samples_per_second": 71.089, |
|
"eval_steps_per_second": 8.908, |
|
"step": 366 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 1.6437464952468872, |
|
"learning_rate": 4e-05, |
|
"loss": 0.2852, |
|
"step": 488 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.8721804511278195, |
|
"eval_f1": 0.8469505178365937, |
|
"eval_loss": 0.3135569393634796, |
|
"eval_precision": 0.844489247311828, |
|
"eval_recall": 0.8495635570103655, |
|
"eval_runtime": 5.5998, |
|
"eval_samples_per_second": 71.253, |
|
"eval_steps_per_second": 8.929, |
|
"step": 488 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 3.6121034622192383, |
|
"learning_rate": 3.7500000000000003e-05, |
|
"loss": 0.2608, |
|
"step": 610 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.8721804511278195, |
|
"eval_f1": 0.8469505178365937, |
|
"eval_loss": 0.3060314953327179, |
|
"eval_precision": 0.844489247311828, |
|
"eval_recall": 0.8495635570103655, |
|
"eval_runtime": 5.6686, |
|
"eval_samples_per_second": 70.388, |
|
"eval_steps_per_second": 8.821, |
|
"step": 610 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 2.0154571533203125, |
|
"learning_rate": 3.5e-05, |
|
"loss": 0.2415, |
|
"step": 732 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.8646616541353384, |
|
"eval_f1": 0.8447157518450185, |
|
"eval_loss": 0.3100413978099823, |
|
"eval_precision": 0.8325401217487549, |
|
"eval_recall": 0.864248045099109, |
|
"eval_runtime": 5.5541, |
|
"eval_samples_per_second": 71.838, |
|
"eval_steps_per_second": 9.002, |
|
"step": 732 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 0.5025161504745483, |
|
"learning_rate": 3.2500000000000004e-05, |
|
"loss": 0.2329, |
|
"step": 854 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.8847117794486216, |
|
"eval_f1": 0.8642214594306682, |
|
"eval_loss": 0.28597915172576904, |
|
"eval_precision": 0.8566755442334414, |
|
"eval_recall": 0.8734315330060011, |
|
"eval_runtime": 5.5629, |
|
"eval_samples_per_second": 71.725, |
|
"eval_steps_per_second": 8.988, |
|
"step": 854 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 2.4278862476348877, |
|
"learning_rate": 3e-05, |
|
"loss": 0.199, |
|
"step": 976 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.8872180451127819, |
|
"eval_f1": 0.8622036668943447, |
|
"eval_loss": 0.2878971993923187, |
|
"eval_precision": 0.8671602787456446, |
|
"eval_recall": 0.8577014002545917, |
|
"eval_runtime": 5.6077, |
|
"eval_samples_per_second": 71.152, |
|
"eval_steps_per_second": 8.916, |
|
"step": 976 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"grad_norm": 8.504670143127441, |
|
"learning_rate": 2.7500000000000004e-05, |
|
"loss": 0.1939, |
|
"step": 1098 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_accuracy": 0.8897243107769424, |
|
"eval_f1": 0.8676337535436396, |
|
"eval_loss": 0.28258949518203735, |
|
"eval_precision": 0.8658613445378152, |
|
"eval_recall": 0.8694762684124386, |
|
"eval_runtime": 5.5938, |
|
"eval_samples_per_second": 71.329, |
|
"eval_steps_per_second": 8.939, |
|
"step": 1098 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"grad_norm": 2.462061882019043, |
|
"learning_rate": 2.5e-05, |
|
"loss": 0.1806, |
|
"step": 1220 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_accuracy": 0.8796992481203008, |
|
"eval_f1": 0.8439374185136896, |
|
"eval_loss": 0.2981988787651062, |
|
"eval_precision": 0.8794955044955045, |
|
"eval_recall": 0.822376795781051, |
|
"eval_runtime": 5.5576, |
|
"eval_samples_per_second": 71.793, |
|
"eval_steps_per_second": 8.997, |
|
"step": 1220 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"grad_norm": 1.0077548027038574, |
|
"learning_rate": 2.25e-05, |
|
"loss": 0.1674, |
|
"step": 1342 |
|
}, |
|
{ |
|
"epoch": 11.0, |
|
"eval_accuracy": 0.8947368421052632, |
|
"eval_f1": 0.8730223677032187, |
|
"eval_loss": 0.2734816372394562, |
|
"eval_precision": 0.8730223677032187, |
|
"eval_recall": 0.8730223677032187, |
|
"eval_runtime": 5.5924, |
|
"eval_samples_per_second": 71.346, |
|
"eval_steps_per_second": 8.941, |
|
"step": 1342 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"grad_norm": 3.9673709869384766, |
|
"learning_rate": 2e-05, |
|
"loss": 0.1553, |
|
"step": 1464 |
|
}, |
|
{ |
|
"epoch": 12.0, |
|
"eval_accuracy": 0.8947368421052632, |
|
"eval_f1": 0.8717238211879976, |
|
"eval_loss": 0.2753015458583832, |
|
"eval_precision": 0.8757194133300328, |
|
"eval_recall": 0.8680214584469903, |
|
"eval_runtime": 5.5661, |
|
"eval_samples_per_second": 71.684, |
|
"eval_steps_per_second": 8.983, |
|
"step": 1464 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"grad_norm": 3.968949794769287, |
|
"learning_rate": 1.75e-05, |
|
"loss": 0.1431, |
|
"step": 1586 |
|
}, |
|
{ |
|
"epoch": 13.0, |
|
"eval_accuracy": 0.8922305764411027, |
|
"eval_f1": 0.8661961395983623, |
|
"eval_loss": 0.2937251627445221, |
|
"eval_precision": 0.8784532165625604, |
|
"eval_recall": 0.8562465902891435, |
|
"eval_runtime": 5.5699, |
|
"eval_samples_per_second": 71.635, |
|
"eval_steps_per_second": 8.977, |
|
"step": 1586 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"grad_norm": 8.566404342651367, |
|
"learning_rate": 1.5e-05, |
|
"loss": 0.1417, |
|
"step": 1708 |
|
}, |
|
{ |
|
"epoch": 14.0, |
|
"eval_accuracy": 0.9072681704260651, |
|
"eval_f1": 0.8910359080340997, |
|
"eval_loss": 0.29110613465309143, |
|
"eval_precision": 0.8822647601476015, |
|
"eval_recall": 0.9018912529550827, |
|
"eval_runtime": 5.5858, |
|
"eval_samples_per_second": 71.432, |
|
"eval_steps_per_second": 8.951, |
|
"step": 1708 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"grad_norm": 0.1758796125650406, |
|
"learning_rate": 1.25e-05, |
|
"loss": 0.1236, |
|
"step": 1830 |
|
}, |
|
{ |
|
"epoch": 15.0, |
|
"eval_accuracy": 0.9022556390977443, |
|
"eval_f1": 0.8817957385392532, |
|
"eval_loss": 0.2955999970436096, |
|
"eval_precision": 0.8827677592299257, |
|
"eval_recall": 0.8808419712675032, |
|
"eval_runtime": 5.6052, |
|
"eval_samples_per_second": 71.184, |
|
"eval_steps_per_second": 8.92, |
|
"step": 1830 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"grad_norm": 0.7015694975852966, |
|
"learning_rate": 1e-05, |
|
"loss": 0.1304, |
|
"step": 1952 |
|
}, |
|
{ |
|
"epoch": 16.0, |
|
"eval_accuracy": 0.9022556390977443, |
|
"eval_f1": 0.884617951284618, |
|
"eval_loss": 0.3010990023612976, |
|
"eval_precision": 0.8772893772893773, |
|
"eval_recall": 0.8933442444080741, |
|
"eval_runtime": 5.6368, |
|
"eval_samples_per_second": 70.785, |
|
"eval_steps_per_second": 8.87, |
|
"step": 1952 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"grad_norm": 0.19915825128555298, |
|
"learning_rate": 7.5e-06, |
|
"loss": 0.1164, |
|
"step": 2074 |
|
}, |
|
{ |
|
"epoch": 17.0, |
|
"eval_accuracy": 0.899749373433584, |
|
"eval_f1": 0.879667048676036, |
|
"eval_loss": 0.29428762197494507, |
|
"eval_precision": 0.8778361344537815, |
|
"eval_recall": 0.8815693762502272, |
|
"eval_runtime": 5.5793, |
|
"eval_samples_per_second": 71.514, |
|
"eval_steps_per_second": 8.962, |
|
"step": 2074 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"grad_norm": 9.03445816040039, |
|
"learning_rate": 5e-06, |
|
"loss": 0.1144, |
|
"step": 2196 |
|
}, |
|
{ |
|
"epoch": 18.0, |
|
"eval_accuracy": 0.8972431077694235, |
|
"eval_f1": 0.8775533117267087, |
|
"eval_loss": 0.2937219738960266, |
|
"eval_precision": 0.873246730188791, |
|
"eval_recall": 0.8822967812329514, |
|
"eval_runtime": 5.5648, |
|
"eval_samples_per_second": 71.7, |
|
"eval_steps_per_second": 8.985, |
|
"step": 2196 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"grad_norm": 5.45957612991333, |
|
"learning_rate": 2.5e-06, |
|
"loss": 0.1198, |
|
"step": 2318 |
|
}, |
|
{ |
|
"epoch": 19.0, |
|
"eval_accuracy": 0.8972431077694235, |
|
"eval_f1": 0.8737897035111135, |
|
"eval_loss": 0.29848915338516235, |
|
"eval_precision": 0.8812047813777917, |
|
"eval_recall": 0.8672940534642661, |
|
"eval_runtime": 5.5642, |
|
"eval_samples_per_second": 71.709, |
|
"eval_steps_per_second": 8.986, |
|
"step": 2318 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"grad_norm": 3.5000033378601074, |
|
"learning_rate": 0.0, |
|
"loss": 0.1104, |
|
"step": 2440 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"eval_accuracy": 0.9022556390977443, |
|
"eval_f1": 0.8811928811928812, |
|
"eval_loss": 0.29284632205963135, |
|
"eval_precision": 0.8842105263157894, |
|
"eval_recall": 0.878341516639389, |
|
"eval_runtime": 5.5868, |
|
"eval_samples_per_second": 71.418, |
|
"eval_steps_per_second": 8.95, |
|
"step": 2440 |
|
}, |
|
{ |
|
"epoch": 20.0, |
|
"step": 2440, |
|
"total_flos": 8551203605328000.0, |
|
"train_loss": 0.2130153269064231, |
|
"train_runtime": 2280.4037, |
|
"train_samples_per_second": 31.907, |
|
"train_steps_per_second": 1.07 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 2440, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 20, |
|
"save_steps": 500, |
|
"total_flos": 8551203605328000.0, |
|
"train_batch_size": 30, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|