weight_decay_05 / all_results.json
sedrickkeh's picture
End of training
39df1d7 verified
raw
history blame contribute delete
222 Bytes
{
"epoch": 2.9878213802435725,
"total_flos": 7.10230108051551e+18,
"train_loss": 0.4706156641460847,
"train_runtime": 60487.3598,
"train_samples_per_second": 2.345,
"train_steps_per_second": 0.005
}