Ultralytics

🚀 ComfyUI Ultralytics Integration – Midnight1111 Model Collection

This project offers an integration of Ultralytics models into ComfyUI, providing a seamless experience for object - detection tasks. It supports a variety of YOLO models and datasets, ensuring high - performance and accurate results.

🚀 Quick Start

Loading Models in ComfyUI

1. Find your ComfyUI directory: Locate the root folder of your ComfyUI installation (e.g., ~/ComfyUI/ or C:\ComfyUI\).
2. Create necessary directories:

   cd /path/to/ComfyUI/models
   mkdir -p ultralytics/segm ultralytics/bbox
   

3. Place your model files:
   • Segmentation models: Put your .pt segmentation files into models/ultralytics/segm/.
   • Detection models: Place your .pt detection (bbox) files into models/ultralytics/bbox/.
4. Restart ComfyUI:

   cd /path/to/ComfyUI
   

5. Use in the UI:
   • In the UI, go to Add Node → Model → Ultralytics, then choose the appropriate model from segm/… or bbox/….
   • Connect an Image Loader to the Ultralytics node, and then to a Previewer.

📄 License

This project is licensed under the AGPL - 3.0 license.

⚠️ Important Note

Unsafe files

Since getattr is considered a dangerous pickle function, any segmentation model that uses it is classified as unsafe. All models in this collection were created and saved using the official Ultralytics library. So, it's safe to use files downloaded from a trusted source. For more information, see [Hugging Face Security Pickle Docs](https://huggingface.co/docs/hub/security - pickle).

Note on Loading Weights

If you encounter an error like:

Weights - only load failed. Unsupported global: builtins.set

This occurs because PyTorch (for safety) blocks loading some objects by default. To fix it safely, add the following code before loading:

import torch
torch.serialization.add_safe_globals([set])

This whitelists set and allows safe loading with weights_only = True. Otherwise, never disable weights_only unless you fully trust the file. Check PyTorch's torch.load docs for more details.

📋 Model Information