76 lines
1.9 KiB
JSON
76 lines
1.9 KiB
JSON
|
{
|
||
|
|
"best_global_step": 1268,
|
||
|
|
"best_metric": 74.2554549397151,
|
||
|
|
"best_model_checkpoint": "mbert-loraxs-qa-vanilla/run-12/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": 4.95363712310791,
|
||
|
|
"learning_rate": 0.00043918755006463925,
|
||
|
|
"loss": 2.7746150741817823,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"eval_exact_match": 57.12,
|
||
|
|
"eval_f1": 71.22068270961232,
|
||
|
|
"eval_loss": 1.3444347381591797,
|
||
|
|
"eval_runtime": 7.7895,
|
||
|
|
"eval_samples_per_second": 162.397,
|
||
|
|
"eval_steps_per_second": 10.27,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"grad_norm": 4.687035083770752,
|
||
|
|
"learning_rate": 0.00032943394078557684,
|
||
|
|
"loss": 1.3028373838599172,
|
||
|
|
"step": 1268
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_exact_match": 59.44,
|
||
|
|
"eval_f1": 74.2554549397151,
|
||
|
|
"eval_loss": 1.300154447555542,
|
||
|
|
"eval_runtime": 7.784,
|
||
|
|
"eval_samples_per_second": 162.512,
|
||
|
|
"eval_steps_per_second": 10.277,
|
||
|
|
"step": 1268
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 500,
|
||
|
|
"max_steps": 3170,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 5,
|
||
|
|
"save_steps": 500,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"TrainerControl": {
|
||
|
|
"args": {
|
||
|
|
"should_epoch_stop": false,
|
||
|
|
"should_evaluate": false,
|
||
|
|
"should_log": false,
|
||
|
|
"should_save": true,
|
||
|
|
"should_training_stop": false
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 6180991462324224.0,
|
||
|
|
"train_batch_size": 16,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": {
|
||
|
|
"learning_rate": 0.0004938912417557808,
|
||
|
|
"lora_alpha": 64,
|
||
|
|
"lora_dropout": 0.08091352388249862,
|
||
|
|
"lora_r": 256,
|
||
|
|
"num_train_epochs": 5,
|
||
|
|
"weight_decay": 0.04137309253112481
|
||
|
|
}
|
||
|
|
}
|