llama3_3bfull / train_results.json
moyixiao's picture
End of training
b81ee64 verified
{
"epoch": 1.996291718170581,
"total_flos": 8.154512631170335e+17,
"train_loss": 0.9267532895549689,
"train_runtime": 5596.5522,
"train_samples_per_second": 18.497,
"train_steps_per_second": 0.144
}