---
library_name: transformers
license: mit
datasets:
- detection-datasets/coco
pipeline_tag: object-detection
---

# Model Card for DiffusionDet

DiffusionDet is a diffusion-based object detection model that formulates object detection as a denoising diffusion process. It iteratively refines noisy box predictions to generate high-quality detection outputs. This approach provides a flexible and unified framework for object detection, offering advantages over traditional proposal-based methods.

## 🔧 Uses

You can load and use the model with Hugging Face's 🤗 `transformers` or via the original repository.

- 📦 [Original GitHub repo](github.com/pierlj/fsdiffusiondet)
- 🚀 [Few-shot cross-domain adaptation repo](https://github.com/ShoufaChen/DiffusionDet)

This model has been adapted for cross-domain few-shot object detection using LoRA (Low-Rank Adaptation).
📄 Check out the paper: [LoRA for Cross-Domain Few-Shot Object Detection](https://huggingface.co/papers/2504.06330)