Spaces:
Running on Zero
Running on Zero
| title: AD-Copilot | |
| emoji: ๐ | |
| colorFrom: blue | |
| colorTo: indigo | |
| sdk: gradio | |
| sdk_version: 5.23.0 | |
| app_file: app.py | |
| hardware: zero-a10g | |
| pinned: true | |
| license: apache-2.0 | |
| models: | |
| - jiang-cc/AD-Copilot | |
| tags: | |
| - anomaly-detection | |
| - vision-language-model | |
| - industrial-inspection | |
| - comparison-aware | |
| short_description: "Comparison-Aware Anomaly Detection VLM" | |
| # AD-Copilot: Comparison-Aware Anomaly Detection | |
| **AD-Copilot** extends Qwen2.5-VL with a novel comparison-aware visual encoder that generates | |
| special comparison tokens capturing differences between a reference image and a test image, | |
| achieving state-of-the-art results on industrial anomaly detection benchmarks (82.3% on MMAD). | |
| ## Usage | |
| 1. Upload a **reference (good) image** of the product | |
| 2. Upload a **test image** to inspect | |
| 3. Choose a prompt (classification or detailed description) | |
| 4. Click **Detect Anomaly** | |
| ## Links | |
| - **Paper:** [arXiv:2603.13779](https://arxiv.org/abs/2603.13779v1) | |
| - **Code:** [GitHub](https://github.com/jam-cc/AD-Copilot) | |
| - **Model:** [HuggingFace](https://huggingface.co/jiang-cc/AD-Copilot) | |