download_model.py

import os
import sys

MODEL_ID = "Qwen/Qwen2.5-VL-7B-Instruct"
# MODEL_ID = "Qwen2.5-VL-3B-Instruct"

def main():
    try:
        from huggingface_hub import snapshot_download
    except ImportError:
        print("[error] huggingface_hub not installed. Run: pip install huggingface_hub")
        sys.exit(1)

    token = os.environ.get("HF_TOKEN")
    if not token:
        print("[info] No HF_TOKEN set. Download will work but rate-limited.")
        print("[info] Optional: set HF_TOKEN env var for faster downloads.")

    print(f"[info] Downloading {MODEL_ID} to HF cache...")
    print("[info] This is ~16 GB. First time: ~5-30 min depending on connection.")
    print("[info] Safe to Ctrl+C and resume — will pick up where it left off.\n")

    try:
        path = snapshot_download(
            repo_id=MODEL_ID,
            token=token,
            resume_download=True,
        )
        print(f"\n[done] Model cached at: {path}")
        print("[done] You can now run: python app.py")
    except KeyboardInterrupt:
        print("\n[interrupted] Partial download saved. Run again to resume.")
        sys.exit(0)
    except Exception as e:
        print(f"\n[error] Download failed: {e}")
        sys.exit(1)


if __name__ == "__main__":
    main()