Files
qwen3-14b-fft-coding/all_results.json

8 lines
203 B
JSON
Raw Normal View History

{
"epoch": 4.0,
"total_flos": 152338919849984.0,
"train_loss": 0.44427714956567643,
"train_runtime": 18988.0523,
"train_samples_per_second": 6.32,
"train_steps_per_second": 0.05
}