Files
Llama3.2-3B_Paper_Impact_ci…/train_results.json

8 lines
208 B
JSON
Raw Permalink Normal View History

{
"epoch": 1.0,
"total_flos": 2.50613257277014e+17,
"train_loss": 0.07390243510770447,
"train_runtime": 1155.3856,
"train_samples_per_second": 14.999,
"train_steps_per_second": 0.118
}