Files
Qwen3-0.6B-SFT-Eval-Smoke/train_results.json

9 lines
243 B
JSON
Raw Normal View History

{
"epoch": 0.008849557522123894,
"total_flos": 21649856397312.0,
"train_loss": 1.0641121864318848,
"train_runtime": 10.9027,
"train_samples": 900,
"train_samples_per_second": 0.734,
"train_steps_per_second": 0.092
}