d1_code_longest / train_results.json
neginr's picture
End of training
f75a4cf verified
{
"epoch": 5.0,
"total_flos": 5.625788126002676e+18,
"train_loss": 0.5128965462991584,
"train_runtime": 29595.0197,
"train_samples_per_second": 5.339,
"train_steps_per_second": 0.042
}