llama-3-8b-finetuned / train_results.json
fay-ong's picture
End of training
2883832 verified
raw
history blame
203 Bytes
{
"epoch": 1.0,
"total_flos": 7074583249158144.0,
"train_loss": 0.2772752811908722,
"train_runtime": 751.7766,
"train_samples_per_second": 0.665,
"train_steps_per_second": 0.166
}