b2_science_length_gpt41nano / all_results.json
neginr's picture
End of training
5f7a54a verified
{
"epoch": 4.984802431610943,
"total_flos": 2.5001335726405714e+18,
"train_loss": 0.4134981281389066,
"train_runtime": 24880.2523,
"train_samples_per_second": 6.345,
"train_steps_per_second": 0.049
}