Qwen2.5-3B-Knowledge-R1-GRPO / all_results.json
hzy's picture
Model save
4b48b2b verified
{
"total_flos": 0.0,
"train_loss": 0.08651691437249102,
"train_runtime": 12893.4863,
"train_samples": 15000,
"train_samples_per_second": 1.163,
"train_steps_per_second": 0.166
}