python-bytes-distilgpt2 / all_results.json
arvkevi's picture
End of training
b3b0af8
raw
history blame
440 Bytes
{
"epoch": 3.0,
"eval_accuracy": 0.39690118338084135,
"eval_loss": 3.037240505218506,
"eval_runtime": 31.7078,
"eval_samples": 1816,
"eval_samples_per_second": 57.273,
"eval_steps_per_second": 14.318,
"perplexity": 20.84763484705914,
"train_loss": 3.0642820732502996,
"train_runtime": 1709.0811,
"train_samples": 10112,
"train_samples_per_second": 17.75,
"train_steps_per_second": 4.437
}