RT-DETRv4: Real-Time Detection Transformer
RT-DETRv4 is a real-time object detection transformer model repackaged for ComfyUI, available in fp16 and fp32 variants.
RT-DETR (Real-Time Detection Transformer) is a real-time object detection model repackaged for use in ComfyUI. Despite being stored in diffusion_models, this is an object detection model — not a diffusion/ generation model. It is designed for ComfyUI custom nodes that load detection models from the diffusion_models directory.
Usage in ComfyUI
Place the model files in ComfyUI/models/diffusion_models/. The model is loaded by ComfyUI custom nodes that support RT-DETR-based object detection.
| Developer | RT-DETR Community (Baidu) |
| Release Date | 2025 |
| Architecture | RT-DETRv4 X-HGNet |
| License | Apache-2.0 |
| Type | Object Detection |
| Upstream | https://github.com/RT-DETRs/RT-DETRv4 |
Upstream Repository
- GitHub: https://github.com/RT-DETRs/RT-DETRv4
- Hugging Face: https://huggingface.co/Comfy-Org/RT-DETR
Guides and workflows related to this model series.
No articles found.