Llama-3.1-8B-Instruct-SFT-200 / train_results.json
chchen's picture
End of training
f046cc5 verified
raw
history blame contribute delete
221 Bytes
{
"epoch": 9.777777777777779,
"total_flos": 1.002088129757184e+16,
"train_loss": 0.8184320038015193,
"train_runtime": 152.2376,
"train_samples_per_second": 11.824,
"train_steps_per_second": 0.723
}