Files
qwen2.5-3b-dpo-vanilla/train_results.json

9 lines
223 B
JSON
Raw Normal View History

{
"epoch": 0.9999813887699838,
"total_flos": 0.0,
"train_loss": 0.0,
"train_runtime": 37.9451,
"train_samples": 3438778,
"train_samples_per_second": 90625.095,
"train_steps_per_second": 707.997
}