opus-em-deberta-3-large-v2 / trainer_state.json
Kerem P
End of training
a718b54
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"eval_steps": 500,
"global_step": 84,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.36,
"learning_rate": 7.996491320395433e-06,
"loss": 459.6986,
"step": 10
},
{
"epoch": 0.71,
"learning_rate": 7.582847040957651e-06,
"loss": 429.225,
"step": 20
},
{
"epoch": 1.0,
"eval_f1": 0.0,
"eval_loss": 329.4008483886719,
"eval_runtime": 2.9528,
"eval_samples_per_second": 59.604,
"eval_steps_per_second": 1.016,
"step": 28
},
{
"epoch": 1.07,
"learning_rate": 6.549695958994759e-06,
"loss": 361.3969,
"step": 30
},
{
"epoch": 1.43,
"learning_rate": 5.0756792824610625e-06,
"loss": 347.9836,
"step": 40
},
{
"epoch": 1.79,
"learning_rate": 3.415667885750354e-06,
"loss": 312.0525,
"step": 50
},
{
"epoch": 2.0,
"eval_f1": 0.0,
"eval_loss": 266.168701171875,
"eval_runtime": 2.5812,
"eval_samples_per_second": 68.187,
"eval_steps_per_second": 1.162,
"step": 56
},
{
"epoch": 2.14,
"learning_rate": 1.8566928200840125e-06,
"loss": 324.8967,
"step": 60
},
{
"epoch": 2.5,
"learning_rate": 6.683150371596022e-07,
"loss": 286.3452,
"step": 70
},
{
"epoch": 2.86,
"learning_rate": 5.601585171798051e-08,
"loss": 280.638,
"step": 80
},
{
"epoch": 3.0,
"eval_f1": 0.0,
"eval_loss": 253.1309051513672,
"eval_runtime": 2.6388,
"eval_samples_per_second": 66.696,
"eval_steps_per_second": 1.137,
"step": 84
},
{
"epoch": 3.0,
"step": 84,
"total_flos": 2697390194688.0,
"train_loss": 348.7352382114955,
"train_runtime": 184.625,
"train_samples_per_second": 14.251,
"train_steps_per_second": 0.455
}
],
"logging_steps": 10,
"max_steps": 84,
"num_input_tokens_seen": 0,
"num_train_epochs": 3,
"save_steps": 500,
"total_flos": 2697390194688.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}