b1_code_top_16 / train_results.json
neginr's picture
End of training
aecaae7 verified
{
"epoch": 5.0,
"total_flos": 2.1632960746246636e+18,
"train_loss": 0.4484364619380549,
"train_runtime": 24245.7929,
"train_samples_per_second": 6.517,
"train_steps_per_second": 0.051
}