HBERTv1_48_L8_H64_A2 / train_results.json
gokuls's picture
End of training
2d07213
raw
history blame contribute delete
No virus
202 Bytes
{
"epoch": 11.78,
"train_loss": 6.314137552510395,
"train_runtime": 197999.2232,
"train_samples": 5858758,
"train_samples_per_second": 2958.98,
"train_steps_per_second": 23.863
}