qwen14B_models/qwen_14B_d1_iter2_model/train_results.json

{
    "epoch": 2.0,
    "total_flos": 93650261901312.0,
    "train_loss": 0.5521188734551153,
    "train_runtime": 1213.4234,
    "train_samples_per_second": 4.45,
    "train_steps_per_second": 0.279
}