Files
Qwen3-4B-Thinking-2507-SFT-tr5/train_results.json

9 lines
247 B
JSON
Raw Normal View History

{
"epoch": 4.6268656716417915,
"total_flos": 829282868854784.0,
"train_loss": 0.02679153286641644,
"train_runtime": 1845.6941,
"train_samples": 4281,
"train_samples_per_second": 10.749,
"train_steps_per_second": 0.336
}