76 lines
1.9 KiB
JSON
76 lines
1.9 KiB
JSON
|
{
|
||
|
|
"best_global_step": 1268,
|
||
|
|
"best_metric": 66.75662514290244,
|
||
|
|
"best_model_checkpoint": "mbert-loraxs-qa-vanilla/run-1/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": 52.79657745361328,
|
||
|
|
"learning_rate": 1.949699397447388e-05,
|
||
|
|
"loss": 3.1322443146810923,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"eval_exact_match": 48.16,
|
||
|
|
"eval_f1": 62.48638687085724,
|
||
|
|
"eval_loss": 1.7752151489257812,
|
||
|
|
"eval_runtime": 7.6669,
|
||
|
|
"eval_samples_per_second": 164.996,
|
||
|
|
"eval_steps_per_second": 10.435,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 47.9760856628418,
|
||
|
|
"learning_rate": 3.0703927518856505e-08,
|
||
|
|
"loss": 1.7527718047614353,
|
||
|
|
"step": 1268
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_exact_match": 52.72,
|
||
|
|
"eval_f1": 66.75662514290244,
|
||
|
|
"eval_loss": 1.5720099210739136,
|
||
|
|
"eval_runtime": 7.7172,
|
||
|
|
"eval_samples_per_second": 163.919,
|
||
|
|
"eval_steps_per_second": 10.366,
|
||
|
|
"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": 5023145924278272.0,
|
||
|
|
"train_batch_size": 16,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": {
|
||
|
|
"learning_rate": 3.503318129901527e-05,
|
||
|
|
"lora_alpha": 512,
|
||
|
|
"lora_dropout": 0.0734622590514945,
|
||
|
|
"lora_r": 128,
|
||
|
|
"num_train_epochs": 2,
|
||
|
|
"weight_decay": 0.004905620950863932
|
||
|
|
}
|
||
|
|
}
|