Mixsmol-4x400M-v0.1-epoch3 / train_results.json
qnguyen3's picture
End of training
40787b0 verified
{
"epoch": 3.0,
"train_loss": 1.6263817286927247,
"train_runtime": 649257.4466,
"train_samples_per_second": 60.135,
"train_steps_per_second": 0.059
}