clip-roberta-finetuned / trainer_state.json
ddanshin's picture
End of training
043a8b5
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"global_step": 6933,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.22,
"learning_rate": 4.639405740660609e-05,
"loss": 0.8524,
"step": 500
},
{
"epoch": 0.43,
"learning_rate": 4.278811481321218e-05,
"loss": 0.5142,
"step": 1000
},
{
"epoch": 0.65,
"learning_rate": 3.9182172219818266e-05,
"loss": 0.433,
"step": 1500
},
{
"epoch": 0.87,
"learning_rate": 3.5576229626424354e-05,
"loss": 0.3839,
"step": 2000
},
{
"epoch": 1.08,
"learning_rate": 3.1970287033030435e-05,
"loss": 0.3263,
"step": 2500
},
{
"epoch": 1.3,
"learning_rate": 2.836434443963652e-05,
"loss": 0.2745,
"step": 3000
},
{
"epoch": 1.51,
"learning_rate": 2.475840184624261e-05,
"loss": 0.2612,
"step": 3500
},
{
"epoch": 1.73,
"learning_rate": 2.11524592528487e-05,
"loss": 0.2418,
"step": 4000
},
{
"epoch": 1.95,
"learning_rate": 1.7546516659454783e-05,
"loss": 0.2301,
"step": 4500
},
{
"epoch": 2.16,
"learning_rate": 1.394057406606087e-05,
"loss": 0.1851,
"step": 5000
},
{
"epoch": 2.38,
"learning_rate": 1.0334631472666955e-05,
"loss": 0.1674,
"step": 5500
},
{
"epoch": 2.6,
"learning_rate": 6.728688879273043e-06,
"loss": 0.1604,
"step": 6000
},
{
"epoch": 2.81,
"learning_rate": 3.122746285879129e-06,
"loss": 0.1545,
"step": 6500
},
{
"epoch": 3.0,
"step": 6933,
"total_flos": 2.3697287737344e+17,
"train_loss": 0.3111796811702326,
"train_runtime": 14391.5786,
"train_samples_per_second": 123.354,
"train_steps_per_second": 0.482
}
],
"max_steps": 6933,
"num_train_epochs": 3,
"total_flos": 2.3697287737344e+17,
"trial_name": null,
"trial_params": null
}