Files
Llama-3.1-8B-Instruct-GRPO-…/train_results.json

8 lines
202 B
JSON
Raw Permalink Normal View History

{
"total_flos": 0.0,
"train_loss": 2.980232949312267e-11,
"train_runtime": 14681.8696,
"train_samples": 7500,
"train_samples_per_second": 0.545,
"train_steps_per_second": 0.007
}