deepmath / train_results.json
sedrickkeh's picture
End of training
b84657d verified
{
"epoch": 4.981412639405205,
"total_flos": 2.069268270658722e+19,
"train_loss": 0.35086224974684455,
"train_runtime": 52973.0269,
"train_samples_per_second": 9.749,
"train_steps_per_second": 0.019
}