gpt2-tdk / all_results.json
TrLOX's picture
Update from
ff265b2
{
"epoch": 10.0,
"train_loss": 2.2214493739870194,
"train_runtime": 13454.7366,
"train_samples": 51053,
"train_samples_per_second": 37.944,
"train_steps_per_second": 2.372
}