BertjeWDialDataALLQonly128 / all_results.json
Jeska's picture
End of training
f887713
raw
history blame
400 Bytes
{
"epoch": 12.0,
"eval_loss": 2.036412477493286,
"eval_runtime": 10.0166,
"eval_samples": 2933,
"eval_samples_per_second": 292.814,
"eval_steps_per_second": 36.639,
"perplexity": 7.663068396450195,
"train_loss": 1.9038402734019955,
"train_runtime": 9444.1472,
"train_samples": 55736,
"train_samples_per_second": 70.82,
"train_steps_per_second": 1.107
}