arc_cot_256 / train_results.json
brettbbb's picture
End of training
8608d8e
raw
history blame contribute delete
168 Bytes
{
"epoch": 20.0,
"train_loss": 0.21532799505366712,
"train_runtime": 2152.3369,
"train_samples_per_second": 2.379,
"train_steps_per_second": 0.595
}