distill_70b_infra_together / all_results.json
sedrickkeh's picture
End of training
ea62e93 verified
raw
history blame
218 Bytes
{
"epoch": 2.962025316455696,
"total_flos": 153686659301376.0,
"train_loss": 0.36335091522106755,
"train_runtime": 1907.3671,
"train_samples_per_second": 3.932,
"train_steps_per_second": 0.041
}