HBERTv1_L12_H768_A12_ffn_2 / train_results.json
gokuls's picture
End of training
ec97d82
raw
history blame contribute delete
200 Bytes
{
"epoch": 2.2,
"train_loss": 4.86919160338339,
"train_runtime": 197999.0138,
"train_samples": 5858758,
"train_samples_per_second": 2958.983,
"train_steps_per_second": 61.646
}