Files
Qwen1.5-MOE-sft-ESFT-law/train_results.json

8 lines
216 B
JSON
Raw Normal View History

{
"total_flos": 1.2005510253379584e+17,
"train_loss": 0.40030854754149914,
"train_runtime": 469.3197,
"train_samples": 927,
"train_samples_per_second": 7.901,
"train_steps_per_second": 0.247
}