qwen7B_models/qwen_7B_d2_iter3_model/train_results.json

{
    "epoch": 1.9428571428571428,
    "total_flos": 5795075391488.0,
    "train_loss": 0.984823696753558,
    "train_runtime": 96.5413,
    "train_samples_per_second": 5.759,
    "train_steps_per_second": 0.352
}