gpt_16_5_5.6e-5_lp5_nb10 / train_results.json
54data's picture
End of training
9471140
{
"epoch": 5.0,
"train_loss": 2.6959276412548974,
"train_runtime": 6590.4176,
"train_samples": 42367,
"train_samples_per_second": 32.143,
"train_steps_per_second": 2.009
}