Files
Qwen2.5-3B-Instruct_Short_CoT/all_results.json

12 lines
344 B
JSON
Raw Permalink Normal View History

{
"epoch": 2.0,
"eval_loss": 0.13600705564022064,
"eval_runtime": 3.0073,
"eval_samples_per_second": 17.957,
"eval_steps_per_second": 4.655,
"total_flos": 12178659213312.0,
"train_loss": 0.07465152751130169,
"train_runtime": 886.8941,
"train_samples_per_second": 12.017,
"train_steps_per_second": 0.753
}