from datasets import Dataset

# Create a dictionary for leaderboard data
leaderboard_data = {
    "model": ["Model_A", "Model_B", "Model_C", "Model_D"],
    "accuracy": [0.92, 0.87, 0.89, 0.95],  # Example metric
    "inference_time (ms)": [12.3, 15.6, 10.8, 14.2]  # Example inference speed
}

# Convert the dictionary into a Hugging Face Dataset
leaderboard = Dataset.from_dict(leaderboard_data)

# Sort the leaderboard by accuracy (descending)
leaderboard = leaderboard.sort("accuracy", reverse=True)

# Print the sorted leaderboard
print(leaderboard)