distilbert_rand_50_v2 / train_results.json
Hartunka's picture
End of training
1768d44 verified
{
"epoch": 25.0,
"total_flos": 7.68761901614592e+17,
"train_loss": 8.045040057203932,
"train_runtime": 16960.8581,
"train_samples": 228639,
"train_samples_per_second": 337.01,
"train_steps_per_second": 3.511
}