|
{ |
|
"best_metric": 54.244334343127505, |
|
"best_model_checkpoint": "/root/turkic_qa/ru_kaz_models/orig_kaz_roberta_base_model/checkpoint-4896", |
|
"epoch": 10.0, |
|
"eval_steps": 500, |
|
"global_step": 5440, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"step": 544, |
|
"train_exact_match": 5.794205794205794, |
|
"train_f1": 11.972302998983034, |
|
"train_runtime": 8.7896, |
|
"train_samples_per_second": 114.226, |
|
"train_steps_per_second": 4.096 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 20.404644012451172, |
|
"learning_rate": 5e-06, |
|
"loss": 4.7402, |
|
"step": 544 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_exact_match": 6.0625, |
|
"eval_f1": 13.131087101850751, |
|
"eval_runtime": 28.5581, |
|
"eval_samples_per_second": 113.278, |
|
"eval_steps_per_second": 4.062, |
|
"step": 544 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"step": 1088, |
|
"train_exact_match": 26.073926073926074, |
|
"train_f1": 40.5622076827893, |
|
"train_runtime": 8.9356, |
|
"train_samples_per_second": 113.255, |
|
"train_steps_per_second": 4.141 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 46.1266975402832, |
|
"learning_rate": 1e-05, |
|
"loss": 3.5753, |
|
"step": 1088 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_exact_match": 20.1875, |
|
"eval_f1": 34.90862745331215, |
|
"eval_runtime": 28.4283, |
|
"eval_samples_per_second": 113.795, |
|
"eval_steps_per_second": 4.08, |
|
"step": 1088 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"step": 1632, |
|
"train_exact_match": 38.86113886113886, |
|
"train_f1": 55.19725821876116, |
|
"train_runtime": 8.8752, |
|
"train_samples_per_second": 113.687, |
|
"train_steps_per_second": 4.169 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 62.19130325317383, |
|
"learning_rate": 8.750000000000001e-06, |
|
"loss": 2.6856, |
|
"step": 1632 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_exact_match": 29.4375, |
|
"eval_f1": 45.76631499899302, |
|
"eval_runtime": 28.556, |
|
"eval_samples_per_second": 113.286, |
|
"eval_steps_per_second": 4.062, |
|
"step": 1632 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"step": 2176, |
|
"train_exact_match": 50.04995004995005, |
|
"train_f1": 66.18847663797447, |
|
"train_runtime": 8.7061, |
|
"train_samples_per_second": 115.551, |
|
"train_steps_per_second": 4.135 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 62.81818389892578, |
|
"learning_rate": 7.500000000000001e-06, |
|
"loss": 2.1937, |
|
"step": 2176 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_exact_match": 32.8125, |
|
"eval_f1": 50.18599134235767, |
|
"eval_runtime": 28.5462, |
|
"eval_samples_per_second": 113.325, |
|
"eval_steps_per_second": 4.064, |
|
"step": 2176 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"step": 2720, |
|
"train_exact_match": 54.645354645354644, |
|
"train_f1": 69.97887368732482, |
|
"train_runtime": 8.7487, |
|
"train_samples_per_second": 115.217, |
|
"train_steps_per_second": 4.115 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 91.22914123535156, |
|
"learning_rate": 6.25e-06, |
|
"loss": 1.9021, |
|
"step": 2720 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_exact_match": 34.6875, |
|
"eval_f1": 52.40841569306178, |
|
"eval_runtime": 28.459, |
|
"eval_samples_per_second": 113.672, |
|
"eval_steps_per_second": 4.076, |
|
"step": 2720 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"step": 3264, |
|
"train_exact_match": 59.54045954045954, |
|
"train_f1": 73.49991951106799, |
|
"train_runtime": 8.7586, |
|
"train_samples_per_second": 114.858, |
|
"train_steps_per_second": 4.11 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 47.82566833496094, |
|
"learning_rate": 5e-06, |
|
"loss": 1.699, |
|
"step": 3264 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_exact_match": 35.4375, |
|
"eval_f1": 53.35301692015627, |
|
"eval_runtime": 28.4753, |
|
"eval_samples_per_second": 113.607, |
|
"eval_steps_per_second": 4.074, |
|
"step": 3264 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"step": 3808, |
|
"train_exact_match": 63.536463536463536, |
|
"train_f1": 76.54817371445398, |
|
"train_runtime": 8.9511, |
|
"train_samples_per_second": 112.724, |
|
"train_steps_per_second": 4.134 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 53.39065933227539, |
|
"learning_rate": 3.7500000000000005e-06, |
|
"loss": 1.535, |
|
"step": 3808 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_exact_match": 36.0, |
|
"eval_f1": 53.62658383188149, |
|
"eval_runtime": 28.5193, |
|
"eval_samples_per_second": 113.432, |
|
"eval_steps_per_second": 4.067, |
|
"step": 3808 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"step": 4352, |
|
"train_exact_match": 65.33466533466533, |
|
"train_f1": 78.68948965537953, |
|
"train_runtime": 8.9571, |
|
"train_samples_per_second": 112.983, |
|
"train_steps_per_second": 4.131 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 31.080413818359375, |
|
"learning_rate": 2.5e-06, |
|
"loss": 1.4091, |
|
"step": 4352 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_exact_match": 35.84375, |
|
"eval_f1": 53.836711882732146, |
|
"eval_runtime": 28.5603, |
|
"eval_samples_per_second": 113.269, |
|
"eval_steps_per_second": 4.062, |
|
"step": 4352 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"step": 4896, |
|
"train_exact_match": 69.33066933066934, |
|
"train_f1": 81.65414813135658, |
|
"train_runtime": 8.9499, |
|
"train_samples_per_second": 113.632, |
|
"train_steps_per_second": 4.134 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"grad_norm": 35.275394439697266, |
|
"learning_rate": 1.25e-06, |
|
"loss": 1.3273, |
|
"step": 4896 |
|
}, |
|
{ |
|
"epoch": 9.0, |
|
"eval_exact_match": 36.5, |
|
"eval_f1": 54.244334343127505, |
|
"eval_runtime": 28.5454, |
|
"eval_samples_per_second": 113.328, |
|
"eval_steps_per_second": 4.064, |
|
"step": 4896 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"step": 5440, |
|
"train_exact_match": 66.83316683316683, |
|
"train_f1": 80.29439955738826, |
|
"train_runtime": 8.8172, |
|
"train_samples_per_second": 114.322, |
|
"train_steps_per_second": 4.083 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"grad_norm": 82.89323425292969, |
|
"learning_rate": 0.0, |
|
"loss": 1.278, |
|
"step": 5440 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"eval_exact_match": 36.375, |
|
"eval_f1": 54.17151192944855, |
|
"eval_runtime": 28.6338, |
|
"eval_samples_per_second": 112.978, |
|
"eval_steps_per_second": 4.051, |
|
"step": 5440 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"step": 5440, |
|
"total_flos": 1.490131269139968e+16, |
|
"train_loss": 2.234532984565286, |
|
"train_runtime": 1951.3415, |
|
"train_samples_per_second": 77.931, |
|
"train_steps_per_second": 2.788 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 5440, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 10, |
|
"save_steps": 500, |
|
"total_flos": 1.490131269139968e+16, |
|
"train_batch_size": 28, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|