Files
Qwen2.5-1.5B-SFT-Mixture-All/train_results.json

9 lines
241 B
JSON
Raw Permalink Normal View History

{
"epoch": 0.6451612903225806,
"total_flos": 542010736640.0,
"train_loss": 1.3059584617614746,
"train_runtime": 162.7126,
"train_samples": 1000,
"train_samples_per_second": 3.933,
"train_steps_per_second": 0.061
}