DeepSeek-R1-Distill-Qwen-7B-GSPO-Basic/train_results.json

{
    "total_flos": 0.0,
    "train_loss": 0.10312542560923918,
    "train_runtime": 114002.2359,
    "train_samples": 3053,
    "train_samples_per_second": 0.134,
    "train_steps_per_second": 0.017
}