# Data Preparation Script
from datasets import load_dataset
import json

def format_dataset(dataset_id):
    dataset = load_dataset(dataset_id, split="train")
    # Add your custom formatting logic here to match Qwen-2.5 chat template
    print(f"Dataset {dataset_id} loaded and formatted.")
    return dataset

if __name__ == "__main__":
    format_dataset("Nanthasit/hermes-dataset")