Files
Qwen2-5-Coder-32B-sft-kimi-800/all_results.json

8 lines
206 B
JSON
Raw Normal View History

{
"epoch": 10.0,
"total_flos": 502552426708992.0,
"train_loss": 0.13730106168985368,
"train_runtime": 20764.2529,
"train_samples_per_second": 0.384,
"train_steps_per_second": 0.012
}