squarerun_large_model / train_results.json
corranm's picture
End of training
536f287 verified
raw
history blame contribute delete
208 Bytes
{
"epoch": 25.0,
"total_flos": 3.163993239336653e+18,
"train_loss": 0.5988656284373478,
"train_runtime": 971.0069,
"train_samples_per_second": 11.895,
"train_steps_per_second": 0.747
}