73 lines
1.8 KiB
JSON
73 lines
1.8 KiB
JSON
|
{
|
||
|
|
"best_global_step": 1268,
|
||
|
|
"best_metric": 78.18340762080186,
|
||
|
|
"best_model_checkpoint": "mbert-loraxs-qa-vanilla/run-9/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": 0.7886435331230284,
|
||
|
|
"grad_norm": 28.185476303100586,
|
||
|
|
"learning_rate": 1.950773981150722e-05,
|
||
|
|
"loss": 2.463543212890625,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0,
|
||
|
|
"eval_exact_match": 61.84,
|
||
|
|
"eval_f1": 76.07411778933292,
|
||
|
|
"eval_loss": 1.221575140953064,
|
||
|
|
"eval_runtime": 5.8101,
|
||
|
|
"eval_samples_per_second": 217.724,
|
||
|
|
"eval_steps_per_second": 13.769,
|
||
|
|
"step": 634
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.5772870662460567,
|
||
|
|
"grad_norm": 30.28702735900879,
|
||
|
|
"learning_rate": 1.2555587348389893e-05,
|
||
|
|
"loss": 1.1182257080078124,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 2.0,
|
||
|
|
"eval_exact_match": 64.4,
|
||
|
|
"eval_f1": 78.18340762080186,
|
||
|
|
"eval_loss": 1.1935663223266602,
|
||
|
|
"eval_runtime": 5.9873,
|
||
|
|
"eval_samples_per_second": 211.279,
|
||
|
|
"eval_steps_per_second": 13.362,
|
||
|
|
"step": 1268
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 500,
|
||
|
|
"max_steps": 1902,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 3,
|
||
|
|
"save_steps": 50,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"TrainerControl": {
|
||
|
|
"args": {
|
||
|
|
"should_epoch_stop": false,
|
||
|
|
"should_evaluate": false,
|
||
|
|
"should_log": false,
|
||
|
|
"should_save": true,
|
||
|
|
"should_training_stop": false
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 3555334320528384.0,
|
||
|
|
"train_batch_size": 16,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": {
|
||
|
|
"learning_rate": 2.3790265728787493e-05,
|
||
|
|
"num_train_epochs": 3,
|
||
|
|
"weight_decay": 0.03290502978450524
|
||
|
|
}
|
||
|
|
}
|