distilbert_rand_20_v1 / train_results.json
Hartunka's picture
End of training
d7ca044 verified
{
"epoch": 25.0,
"total_flos": 7.68356804521728e+17,
"train_loss": 7.646758914822596,
"train_runtime": 16728.5175,
"train_samples": 228639,
"train_samples_per_second": 341.69,
"train_steps_per_second": 3.56
}