llama-160m-qqp / train_results.json
Cheng98's picture
End of training
a1c373a
raw
history blame contribute delete
197 Bytes
{
"epoch": 4.0,
"train_loss": 0.5902769758531864,
"train_runtime": 11488.0766,
"train_samples": 363846,
"train_samples_per_second": 126.686,
"train_steps_per_second": 0.99
}