{"id":29092235,"url":"https://github.com/nevrohelios/deep-viz","last_synced_at":"2026-04-05T11:32:30.801Z","repository":{"id":301422482,"uuid":"1008217740","full_name":"NevroHelios/deep-viz","owner":"NevroHelios","description":"🔬 Deep-Viz: Unveiling the Black Box of Deep Learning","archived":false,"fork":false,"pushed_at":"2025-06-26T19:09:28.000Z","size":9,"stargazers_count":1,"open_issues_count":0,"forks_count":1,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-06-26T20:20:37.444Z","etag":null,"topics":["captum","opencv","pytorch","streamlit","torchvision"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/NevroHelios.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-06-25T07:54:49.000Z","updated_at":"2025-06-26T19:10:50.000Z","dependencies_parsed_at":"2025-06-26T20:20:39.732Z","dependency_job_id":"d0741a2b-69f4-4f75-a9ff-efe39d0bb705","html_url":"https://github.com/NevroHelios/deep-viz","commit_stats":null,"previous_names":["nevrohelios/deep-viz"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/NevroHelios/deep-viz","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/NevroHelios%2Fdeep-viz","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/NevroHelios%2Fdeep-viz/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/NevroHelios%2Fdeep-viz/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/NevroHelios%2Fdeep-viz/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/NevroHelios","download_url":"https://codeload.github.com/NevroHelios/deep-viz/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/NevroHelios%2Fdeep-viz/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":262389457,"owners_count":23303341,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["captum","opencv","pytorch","streamlit","torchvision"],"created_at":"2025-06-28T07:04:04.560Z","updated_at":"2026-04-05T11:32:30.761Z","avatar_url":"https://github.com/NevroHelios.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🔬 Deep-Viz: Unveiling the Black Box of Deep Learning\n\n\u003cdiv align=\"center\"\u003e\n\n![Deep Learning Visualization](https://img.shields.io/badge/Deep%20Learning-Visualization-blue?style=for-the-badge\u0026logo=pytorch)\n![Streamlit](https://img.shields.io/badge/Streamlit-FF4B4B?style=for-the-badge\u0026logo=streamlit\u0026logoColor=white)\n![PyTorch](https://img.shields.io/badge/PyTorch-EE4C2C?style=for-the-badge\u0026logo=pytorch\u0026logoColor=white)\n\n*\"Seeing is believing, but understanding is enlightening\"* 🧠✨\n\n\u003c/div\u003e\n\n## 🌟 What is Deep-Viz?\n\nDeep-Viz is an interactive web application that makes AI interpretability accessible to everyone! Ever wondered what a neural network \"sees\" when it looks at an image? This tool pulls back the curtain on deep learning models, revealing the hidden patterns and decision-making processes that drive AI predictions.\n\n## 🎯 Features\n\n### 🔍 **Model Interpretability Made Simple**\n- **Class Activation Maps (CAM)** using SmoothGradCAM++ 🎨\n- **Integrated Gradients** for pixel-level feature attribution 📊\n- **Real-time visualization** of model predictions 🚀\n- **Multi-model support** (ResNet18 \u0026 ResNet50) ⚙️\n\n### 🎨 **Interactive Experience**\n- **Drag-and-drop** image upload 📤\n- **Layer-wise exploration** of neural network activations 🔬\n- **Top-5 predictions** with confidence scores 📈\n- **Side-by-side comparison** of different visualization techniques 🔄\n\n## 🧠 How It Works\n\n```mermaid\ngraph TD\n    A[📷 Upload Image] --\u003e B[🔄 Preprocessing]\n    B --\u003e C[🤖 ResNet Model]\n    C --\u003e D[📊 Predictions]\n    C --\u003e E[🎨 SmoothGradCAM++]\n    C --\u003e F[🔍 Integrated Gradients]\n    \n    D --\u003e G[📈 Top-5 Results]\n    E --\u003e H[🗺️ Class Activation Map]\n    F --\u003e I[⚡ Attribution Heatmap]\n    \n    G --\u003e J[📱 Interactive Dashboard]\n    H --\u003e J\n    I --\u003e J\n    \n    style A fill:#ff6b6b\n    style C fill:#4ecdc4\n    style J fill:#45b7d1\n```\n\n## 🚀 Quick Start\n\n### Prerequisites\nMake sure you have Python 3.7+ installed on your system.\n\n### Installation\n\n1. **Clone the repository**\n```bash\ngit clone https://github.com/nevrohelios/deep-viz.git\ncd deep-viz\n```\n\n2. **Install dependencies**\n```bash\npip install -r requirements.txt\n```\n\n3. **Launch the application**\n```bash\nstreamlit run main.py\n```\n\n4. **Open your browser** and navigate to `http://localhost:8501` 🌐\n\n## 📚 Understanding the Visualizations\n\n### 🎨 Class Activation Maps (CAM)\nCAMs highlight the regions in an image that are most important for the model's prediction. Think of it as the model's \"attention map\" - where is it looking when making decisions?\n\n```\n🖼️ Original Image → 🔥 Heat Map → 🎯 Important Regions\n```\n\n### ⚡ Integrated Gradients\nThis technique provides pixel-level attributions, showing how much each pixel contributes to the final prediction. It's like having X-ray vision into the model's decision process!\n\n```\n🔍 Pixel Analysis → 📊 Attribution Scores → 🎨 Visualization\n```\n\n## 🔧 Technical Architecture\n\n### 📦 Core Components\n\n| Component | Purpose | Technology |\n|-----------|---------|------------|\n| 🎨 **Frontend** | Interactive UI | Streamlit |\n| 🤖 **Model Backend** | Deep Learning | PyTorch + TorchVision |\n| 🔍 **Interpretability** | CAM Generation | TorchCAM |\n| ⚡ **Attribution** | Gradient Analysis | Captum |\n| 🖼️ **Image Processing** | Preprocessing | PIL + Transforms |\n\n### 🏗️ Model Architecture\n\n```\nInput Image (224×224×3)\n     ↓\nResNet18/50 Backbone\n     ↓\nFeature Extraction\n     ↓\n┌─────────────────┬─────────────────┐\n│  CAM Extraction │ Gradient Flow   │\n│  (TorchCAM)     │ (Captum)        │\n└─────────────────┴─────────────────┘\n     ↓\nVisualization Pipeline\n     ↓\nInteractive Dashboard\n```\n\n## 🎮 Usage Examples\n\n### 🐕 Example 1: Animal Classification\nUpload a photo of your pet and see which features the model uses to identify the animal!\n\n### 🏞️ Example 2: Scene Understanding\nTry landscape photos to understand how the model recognizes different environments.\n\n### 🚗 Example 3: Object Detection\nUpload images with vehicles to see how the model focuses on distinctive features.\n\n## 🔬 Advanced Features\n\n### 🎛️ **Layer Selection**\nExplore different layers of the neural network to understand how features evolve from simple edges to complex patterns:\n\n- **Early Layers**: Edge detection, basic shapes 📐\n- **Middle Layers**: Textures, patterns 🌀\n- **Deep Layers**: Complex objects, semantic features 🏗️\n\n### 📊 **Model Comparison**\nSwitch between ResNet18 and ResNet50 to see how model depth affects interpretation:\n\n| Model | Parameters | Speed | Accuracy |\n|-------|------------|-------|----------|\n| ResNet18 | 11.7M | ⚡ Fast | 🎯 Good |\n| ResNet50 | 25.6M | 🐌 Slower | 🎯 Better |\n\n## 🔮 What Makes This Special?\n\n### 🎨 **Visual Excellence**\n- Clean, intuitive interface designed for both beginners and experts\n- Real-time processing with smooth animations\n- High-quality visualizations that reveal model insights\n\n### 🧪 **Educational Value**\n- Perfect for students learning about AI interpretability\n- Demonstrates cutting-edge explainable AI techniques\n- Bridges the gap between theory and practice\n\n### 🚀 **Production Ready**\n- Robust error handling and user feedback\n- Optimized for performance\n- Extensible architecture for adding new models\n\n## 🛠️ Dependencies\n\n```python\n# Core ML Framework\ntorch              # PyTorch deep learning framework\ntorchvision        # Computer vision utilities\ntorchcam           # Class Activation Mapping\n\n# Interpretability\ncaptum             # Model interpretability library\n\n# Web Framework\nstreamlit          # Interactive web applications\n\n# Image Processing\npillow             # Python Imaging Library\nnumpy              # Numerical computing\nmatplotlib         # Plotting library\n```\n\n## 🔄 Future Enhancements\n\n- [ ] 🎯 **More Models**: Add support for Vision Transformers, EfficientNet\n- [ ] 🎨 **Advanced Visualizations**: GradCAM, LIME, SHAP\n- [ ] 📱 **Mobile Support**: Responsive design for mobile devices\n- [ ] 🔗 **API Integration**: REST API for programmatic access\n- [ ] 🎥 **Video Analysis**: Extend to video classification\n- [ ] 🌐 **Multi-language**: Support for multiple languages\n\n## 🤝 Contributing\n\nWe welcome contributions! Whether it's:\n- 🐛 Bug fixes\n- ✨ New features\n- 📚 Documentation improvements\n- 🎨 UI/UX enhancements\n\nPlease feel free to open issues and pull requests!\n\n## 📄 License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\n## 🙏 Acknowledgments\n\n- **PyTorch Team** for the amazing deep learning framework 🔥\n- **Streamlit** for making web apps incredibly simple 🚀\n- **Captum** for state-of-the-art interpretability tools 🔍\n- **TorchCAM** for excellent CAM implementations 🎨\n\n---\n\n\u003cdiv align=\"center\"\u003e\n\n**Made with ❤️ and lots of ☕**\n\n*Star ⭐ this repo if you found it helpful!*\n\n\u003c/div\u003e\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnevrohelios%2Fdeep-viz","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnevrohelios%2Fdeep-viz","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnevrohelios%2Fdeep-viz/lists"}