Llama-3.1-8B-Instruct-SFT-100 / all_results.json
chchen's picture
End of training
0b5db6a verified
raw
history blame contribute delete
356 Bytes
{
"epoch": 8.88888888888889,
"eval_loss": 1.191369891166687,
"eval_runtime": 0.2544,
"eval_samples_per_second": 39.311,
"eval_steps_per_second": 19.656,
"total_flos": 4594288692953088.0,
"train_loss": 1.2649920654296876,
"train_runtime": 69.6409,
"train_samples_per_second": 12.923,
"train_steps_per_second": 0.718
}