distilbert_rand_100_v1_rte / train_results.json
Hartunka's picture
End of training
fd2bd06 verified
{
"epoch": 6.0,
"total_flos": 989531467960320.0,
"train_loss": 0.5249984780947368,
"train_runtime": 22.3739,
"train_samples": 2490,
"train_samples_per_second": 5564.514,
"train_steps_per_second": 22.347
}