distilbert_rand_100_v1 / train_results.json
Hartunka's picture
End of training
3cdb891 verified
{
"epoch": 25.0,
"total_flos": 7.69437063436032e+17,
"train_loss": 8.341223693386434,
"train_runtime": 17177.2378,
"train_samples": 228639,
"train_samples_per_second": 332.765,
"train_steps_per_second": 3.467
}