llama-3.1-8b-final_model / train_results.json
yuzhounie's picture
End of training
1912dd1 verified
{
"epoch": 2.9972020145495244,
"total_flos": 1.9585051653255987e+17,
"train_loss": 0.5026469534026787,
"train_runtime": 9007.4522,
"train_samples_per_second": 4.76,
"train_steps_per_second": 0.099
}