Files
llama-3-8b-base-margin-dpo-…/train_results.json

9 lines
232 B
JSON
Raw Permalink Normal View History

{
"epoch": 0.9977324263038548,
"total_flos": 0.0,
"train_loss": 4.569111451235685,
"train_runtime": 1865.9225,
"train_samples": 42336,
"train_samples_per_second": 22.689,
"train_steps_per_second": 0.177
}