hBERTv1_new_pretrain_w_init__sst2 / train_results.json
gokuls's picture
End of training
a3a117d
raw
history blame contribute delete
199 Bytes
{
"epoch": 6.0,
"train_loss": 0.18471441202567845,
"train_runtime": 2261.4914,
"train_samples": 67349,
"train_samples_per_second": 1489.039,
"train_steps_per_second": 11.652
}