Files
qwen3-8b-finetuned-train/all_results.json

8 lines
209 B
JSON
Raw Normal View History

{
"epoch": 3.0,
"total_flos": 4.717303430077481e+17,
"train_loss": 0.15186350211219263,
"train_runtime": 38487.8681,
"train_samples_per_second": 1.115,
"train_steps_per_second": 0.139
}