Llama-3.1-8B-Instruct-SAA-600 / train_results.json
chchen's picture
End of training
680afe0 verified
raw
history blame contribute delete
220 Bytes
{
"epoch": 9.777777777777779,
"total_flos": 5.97337003547689e+16,
"train_loss": 0.41489303653890436,
"train_runtime": 721.1538,
"train_samples_per_second": 7.488,
"train_steps_per_second": 0.458
}