d1_code_shortest / train_results.json
neginr's picture
End of training
a6a99ca verified
{
"epoch": 5.0,
"total_flos": 4.783951705282707e+18,
"train_loss": 0.49450628289326964,
"train_runtime": 29398.0301,
"train_samples_per_second": 5.375,
"train_steps_per_second": 0.042
}