ml815-model9 / train_results.json
adpretko's picture
End of training
d773cbd verified
{
"epoch": 1.0,
"total_flos": 3.151751154761728e+17,
"train_loss": 0.0763875040132251,
"train_runtime": 2448.9622,
"train_samples_per_second": 4.037,
"train_steps_per_second": 0.126
}