python-bytes-distilgpt2 / train_results.json
arvkevi's picture
End of training
b3b0af8
{
"epoch": 3.0,
"train_loss": 3.0642820732502996,
"train_runtime": 1709.0811,
"train_samples": 10112,
"train_samples_per_second": 17.75,
"train_steps_per_second": 4.437
}