AML_A2_Q4 / train_results.json
heyitskim1912's picture
End of training
c8cf223
{
"epoch": 3.0,
"total_flos": 1.046216869705728e+19,
"train_loss": 0.17347003343370226,
"train_runtime": 5139.576,
"train_samples_per_second": 26.267,
"train_steps_per_second": 1.313
}