erberry's picture
Training in progress, epoch 4
4f75d1f verified
{
"best_metric": 0.0017974835230677051,
"best_model_checkpoint": "bert-base-multilingual-uncased-finetuned-keyword/run-1/checkpoint-136",
"epoch": 4.0,
"eval_steps": 500,
"global_step": 272,
"is_hyper_param_search": true,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"eval_accuracy": 0.0011983223487118035,
"eval_f1": 2.9424735388871782e-06,
"eval_loss": 6.737747669219971,
"eval_precision": 1.473045296511129e-06,
"eval_recall": 0.0011983223487118035,
"eval_runtime": 4.6104,
"eval_samples_per_second": 362.007,
"eval_steps_per_second": 22.775,
"step": 68
},
{
"epoch": 2.0,
"eval_accuracy": 0.0017974835230677051,
"eval_f1": 7.557349934848005e-05,
"eval_loss": 6.736907482147217,
"eval_precision": 4.0131326075164875e-05,
"eval_recall": 0.0017974835230677051,
"eval_runtime": 4.5841,
"eval_samples_per_second": 364.083,
"eval_steps_per_second": 22.905,
"step": 136
},
{
"epoch": 3.0,
"eval_accuracy": 0.0017974835230677051,
"eval_f1": 2.0759356980301492e-05,
"eval_loss": 6.735987663269043,
"eval_precision": 1.051324452567915e-05,
"eval_recall": 0.0017974835230677051,
"eval_runtime": 4.5567,
"eval_samples_per_second": 366.276,
"eval_steps_per_second": 23.043,
"step": 204
},
{
"epoch": 4.0,
"eval_accuracy": 0.0017974835230677051,
"eval_f1": 2.0035414206783107e-05,
"eval_loss": 6.7355804443359375,
"eval_precision": 1.0140345544131168e-05,
"eval_recall": 0.0017974835230677051,
"eval_runtime": 5.2816,
"eval_samples_per_second": 316.0,
"eval_steps_per_second": 19.88,
"step": 272
}
],
"logging_steps": 500,
"max_steps": 272,
"num_input_tokens_seen": 0,
"num_train_epochs": 4,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0,
"train_batch_size": 64,
"trial_name": null,
"trial_params": {
"learning_rate": 2.4066336582627224e-06,
"num_train_epochs": 4,
"per_device_train_batch_size": 64,
"seed": 32
}
}