roberta-large-neg-tags / trainer_state.json
spoiled's picture
End of training
5e1b599
{
"best_metric": 0.001565825194120407,
"best_model_checkpoint": "./roberta-large-neg-tags/checkpoint-3752",
"epoch": 4.0,
"global_step": 3752,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.53,
"learning_rate": 4.333688699360342e-05,
"loss": 0.0143,
"step": 500
},
{
"epoch": 1.0,
"eval_accuracy": 0.9994915007990701,
"eval_f1": 0.0,
"eval_loss": 0.0031818156130611897,
"eval_precision": 0.0,
"eval_recall": 0.0,
"eval_runtime": 38.1925,
"eval_samples_per_second": 78.314,
"eval_steps_per_second": 4.896,
"step": 938
},
{
"epoch": 1.07,
"learning_rate": 3.6673773987206824e-05,
"loss": 0.0044,
"step": 1000
},
{
"epoch": 1.6,
"learning_rate": 3.0010660980810235e-05,
"loss": 0.0033,
"step": 1500
},
{
"epoch": 2.0,
"eval_accuracy": 0.9996367862850501,
"eval_f1": 0.0,
"eval_loss": 0.0016651672776788473,
"eval_precision": 0.0,
"eval_recall": 0.0,
"eval_runtime": 38.5549,
"eval_samples_per_second": 77.578,
"eval_steps_per_second": 4.85,
"step": 1876
},
{
"epoch": 2.13,
"learning_rate": 2.3347547974413646e-05,
"loss": 0.0026,
"step": 2000
},
{
"epoch": 2.67,
"learning_rate": 1.668443496801706e-05,
"loss": 0.0039,
"step": 2500
},
{
"epoch": 3.0,
"eval_accuracy": 0.9996731076565452,
"eval_f1": 0.0,
"eval_loss": 0.001775014097802341,
"eval_precision": 0.0,
"eval_recall": 0.0,
"eval_runtime": 38.5702,
"eval_samples_per_second": 77.547,
"eval_steps_per_second": 4.848,
"step": 2814
},
{
"epoch": 3.2,
"learning_rate": 1.002132196162047e-05,
"loss": 0.0016,
"step": 3000
},
{
"epoch": 3.73,
"learning_rate": 3.358208955223881e-06,
"loss": 0.0012,
"step": 3500
},
{
"epoch": 4.0,
"eval_accuracy": 0.9996731076565452,
"eval_f1": 0.0,
"eval_loss": 0.001565825194120407,
"eval_precision": 0.0,
"eval_recall": 0.0,
"eval_runtime": 38.2316,
"eval_samples_per_second": 78.234,
"eval_steps_per_second": 4.891,
"step": 3752
},
{
"epoch": 4.0,
"step": 3752,
"total_flos": 4997923744417344.0,
"train_loss": 0.004218238860622906,
"train_runtime": 2309.4182,
"train_samples_per_second": 51.961,
"train_steps_per_second": 1.625
}
],
"max_steps": 3752,
"num_train_epochs": 4,
"total_flos": 4997923744417344.0,
"trial_name": null,
"trial_params": null
}