pretrain-20250131 / train_results.json
furmaniak's picture
End of training
f496eee verified
raw
history blame contribute delete
220 Bytes
{
"epoch": 0.9986282578875172,
"total_flos": 1921476294868992.0,
"train_loss": 1.1405032803287436,
"train_runtime": 59872.6859,
"train_samples_per_second": 0.584,
"train_steps_per_second": 0.005
}