76 lines
1.9 KiB
JSON
76 lines
1.9 KiB
JSON
|
{
|
||
|
|
"best_global_step": 1268,
|
||
|
|
"best_metric": 10.893501835802274,
|
||
|
|
"best_model_checkpoint": "mbert-loraxs-qa-vanilla/run-0/checkpoint-1268",
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_steps": 500,
|
||
|
|
"global_step": 1268,
|
||
|
|
"is_hyper_param_search": true,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"grad_norm": 9.316121101379395,
|
||
|
|
"learning_rate": 1.2339740823845353e-05,
|
||
|
|
"loss": 5.004710296727129,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"eval_exact_match": 3.52,
|
||
|
|
"eval_f1": 8.929509943172043,
|
||
|
|
"eval_loss": 4.074629306793213,
|
||
|
|
"eval_runtime": 7.6506,
|
||
|
|
"eval_samples_per_second": 165.347,
|
||
|
|
"eval_steps_per_second": 10.457,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 8.535039901733398,
|
||
|
|
"learning_rate": 1.943266271471709e-08,
|
||
|
|
"loss": 4.068548064126578,
|
||
|
|
"step": 1268
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_exact_match": 5.04,
|
||
|
|
"eval_f1": 10.893501835802274,
|
||
|
|
"eval_loss": 3.852447986602783,
|
||
|
|
"eval_runtime": 7.6373,
|
||
|
|
"eval_samples_per_second": 165.635,
|
||
|
|
"eval_steps_per_second": 10.475,
|
||
|
|
"step": 1268
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 500,
|
||
|
|
"max_steps": 1268,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 2,
|
||
|
|
"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": 4485574781614080.0,
|
||
|
|
"train_batch_size": 16,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": {
|
||
|
|
"learning_rate": 2.21726681574922e-05,
|
||
|
|
"lora_alpha": 128,
|
||
|
|
"lora_dropout": 0.04085479876923811,
|
||
|
|
"lora_r": 64,
|
||
|
|
"num_train_epochs": 2,
|
||
|
|
"weight_decay": 0.013911443680370162
|
||
|
|
}
|
||
|
|
}
|