2023-02-12 20:03:40 +01:00
|
|
|
{
|
|
|
|
"best_metric": null,
|
|
|
|
"best_model_checkpoint": null,
|
2023-02-13 00:49:10 +01:00
|
|
|
"epoch": 1.0,
|
|
|
|
"global_step": 2000,
|
2023-02-12 20:03:40 +01:00
|
|
|
"is_hyper_param_search": false,
|
|
|
|
"is_local_process_zero": true,
|
|
|
|
"is_world_process_zero": true,
|
|
|
|
"log_history": [
|
|
|
|
{
|
2023-02-13 00:49:10 +01:00
|
|
|
"epoch": 0.25,
|
|
|
|
"learning_rate": 3.7500000000000003e-05,
|
|
|
|
"loss": 0.9449,
|
2023-02-12 20:03:40 +01:00
|
|
|
"step": 500
|
|
|
|
},
|
|
|
|
{
|
2023-02-13 00:49:10 +01:00
|
|
|
"epoch": 0.5,
|
|
|
|
"learning_rate": 2.5e-05,
|
|
|
|
"loss": 0.3705,
|
2023-02-12 20:03:40 +01:00
|
|
|
"step": 1000
|
|
|
|
},
|
|
|
|
{
|
2023-02-13 00:49:10 +01:00
|
|
|
"epoch": 0.75,
|
|
|
|
"learning_rate": 1.25e-05,
|
|
|
|
"loss": 0.264,
|
|
|
|
"step": 1500
|
|
|
|
},
|
|
|
|
{
|
|
|
|
"epoch": 1.0,
|
|
|
|
"learning_rate": 0.0,
|
|
|
|
"loss": 0.2223,
|
|
|
|
"step": 2000
|
|
|
|
},
|
|
|
|
{
|
|
|
|
"epoch": 1.0,
|
|
|
|
"step": 2000,
|
|
|
|
"total_flos": 1204741472256000.0,
|
|
|
|
"train_loss": 0.4504347610473633,
|
|
|
|
"train_runtime": 524.6759,
|
|
|
|
"train_samples_per_second": 30.495,
|
|
|
|
"train_steps_per_second": 3.812
|
2023-02-12 20:03:40 +01:00
|
|
|
}
|
|
|
|
],
|
2023-02-13 00:49:10 +01:00
|
|
|
"max_steps": 2000,
|
|
|
|
"num_train_epochs": 1,
|
|
|
|
"total_flos": 1204741472256000.0,
|
2023-02-12 20:03:40 +01:00
|
|
|
"trial_name": null,
|
|
|
|
"trial_params": null
|
|
|
|
}
|