def main()

in llm_perf/common/get_top_model_from_hub.py [0:0]

• 22 lines of code
• 4 McCabe index (conditional complexity)

def main():
    # Set up authentication (optional, but recommended)
    huggingface_token = os.environ.get("HUGGINGFACE_TOKEN")
    if huggingface_token:
        os.environ["HUGGINGFACE_HUB_TOKEN"] = huggingface_token
    else:
        print(
            "Warning: HUGGINGFACE_TOKEN not found in environment variables. Running without authentication."
        )

    n = 100
    top_models = get_top_text_generation_models(n)

    print(f"\nTop {n} text generation models on Hugging Face Hub:")
    for i, model in enumerate(top_models, 1):
        print(
            f"{i}. {model['organization']}/{model['model_name']}: {model['downloads']:,} downloads"
        )

    # Upload to Hugging Face dataset
    dataset_name = "optimum-benchmark/top-text-generation-models"
    upload_to_hf_dataset(top_models, dataset_name)

    # Display top 10 organizations by downloads
    print("\nTop 10 organizations by total downloads:")
    org_downloads = compute_org_downloads(top_models)
    sorted_orgs = sorted(org_downloads.items(), key=lambda x: x[1], reverse=True)[:10]
    for i, (org, downloads) in enumerate(sorted_orgs, 1):
        print(f"{i}. {org}: {downloads:,} downloads")