arc_cot_128 / train_results.json
brettbbb's picture
End of training
52afb3f
raw
history blame contribute delete
168 Bytes
{
"epoch": 20.0,
"train_loss": 0.23844831050664653,
"train_runtime": 1109.6761,
"train_samples_per_second": 2.307,
"train_steps_per_second": 0.577
}