Files
Qwen3-1.7B-SFT-100k/all_results.json

8 lines
203 B
JSON
Raw Normal View History

{
"epoch": 3.0,
"total_flos": 145115603533824.0,
"train_loss": 0.698847218818268,
"train_runtime": 3222.3497,
"train_samples_per_second": 29.579,
"train_steps_per_second": 0.463
}