qwen2.5-0.5b-instruct_gsm8k_full-finetuning/all_results.json

{
    "epoch": 0.010036130068245684,
    "total_flos": 7693271040.0,
    "train_loss": 0.6475268173217773,
    "train_runtime": 260.2772,
    "train_samples": 7473,
    "train_samples_per_second": 0.287,
    "train_steps_per_second": 0.288
}