Files
Llama-3.1-8B-Instruct_SFT_M…/train_results.json

8 lines
221 B
JSON
Raw Normal View History

{
"total_flos": 2.6506254854358303e+19,
"train_loss": 0.37255040888851254,
"train_runtime": 23487.3778,
"train_samples": 86158,
"train_samples_per_second": 11.005,
"train_steps_per_second": 0.688
}