{"id":28426277,"url":"https://github.com/ajitashwath/nn-visualization","last_synced_at":"2026-05-03T18:34:01.397Z","repository":{"id":274017552,"uuid":"921645615","full_name":"ajitashwath/nn-visualization","owner":"ajitashwath","description":"A web application for visualizing various aspects of neural networks.","archived":false,"fork":false,"pushed_at":"2025-07-04T14:41:33.000Z","size":24,"stargazers_count":2,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-09-15T15:03:03.057Z","etag":null,"topics":["matplotlib-pyplot","python3","scikit-learn","streamlit","tensorflow"],"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/ajitashwath.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":"2025-01-24T10:44:10.000Z","updated_at":"2025-07-10T05:56:19.000Z","dependencies_parsed_at":"2025-03-03T08:21:18.760Z","dependency_job_id":"3f578179-15d9-427e-9278-702ff0717a7b","html_url":"https://github.com/ajitashwath/nn-visualization","commit_stats":null,"previous_names":["ajitashwathr10/nn-visualization","ajitashwath/nn-visualization"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/ajitashwath/nn-visualization","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ajitashwath%2Fnn-visualization","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ajitashwath%2Fnn-visualization/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ajitashwath%2Fnn-visualization/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ajitashwath%2Fnn-visualization/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ajitashwath","download_url":"https://codeload.github.com/ajitashwath/nn-visualization/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ajitashwath%2Fnn-visualization/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":32443514,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-04-29T20:22:27.477Z","status":"ssl_error","status_checked_at":"2026-04-29T20:22:26.507Z","response_time":110,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"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":["matplotlib-pyplot","python3","scikit-learn","streamlit","tensorflow"],"created_at":"2025-06-05T11:09:28.497Z","updated_at":"2026-04-29T20:32:59.866Z","avatar_url":"https://github.com/ajitashwath.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Neural Network Visualization Toolkit\nA web application for visualizing various aspects of neural networks, including architecture, activations, decision boundaries, gradients, and training progress.\n\n## Features\n- **Architecture Visualization**: Display neural network structure with layer details\n- **Activation Maps**: Visualize feature maps in convolutional layers\n- **Decision Boundaries**: Plot classification boundaries for 2D datasets\n- **Gradient Analysis**: Examine gradient distributions across layers\n- **Training Progress**: Real-time monitoring of loss and accuracy during training\n\n## Installation\n1. Clone the repository:\n```bash\ngit clone https://github.com/ajitashwath/nn-visualization.git\ncd nn-visualization\n```\n\n2. Install required dependencies:\n```bash\npip install streamlit tensorflow numpy matplotlib scikit-learn keras\n```\n\n## Usage\n\nLaunch the application:\n```bash\nstreamlit run app.py\n```\n\n## Application Interface\n\n### Sidebar Controls\n\n- **Visualization Type**: Select from 5 different visualization options\n- **Hyperparameters**: \n  - Learning Rate: 0.001 - 0.1\n  - Batch Size: 16 - 128\n  - Epochs: 1 - 100\n- **Model Type**: Choose between Simple Neural Network or CNN\n- **Custom Data**: Upload CSV datasets for analysis\n\n### Visualization Options\n\n#### 1. Architecture\nDisplays network structure using Keras plot_model functionality.\n\n**Models Available:**\n- Simple Neural Network: 784 → 64 → 64 → 10 (ReLU, Softmax)\n- CNN: Conv2D(32) → MaxPool → Flatten → Dense(64) → Dense(10)\n\n#### 2. Activations\nVisualizes activation maps for convolutional layers.\n\n**Features:**\n- Uses pre-trained VGG16 model\n- Displays up to 16 feature maps in 4x4 grid\n- Only available for CNN models\n\n#### 3. Decision Boundaries\nPlots classification boundaries for 2D datasets.\n\n**Implementation:**\n- Uses make_moons dataset (1000 samples, 0.2 noise)\n- Binary classification with sigmoid activation\n- Visualizes decision contours with data points\n\n#### 4. Gradients\nAnalyzes gradient distributions across network layers.\n\n**Visualization:**\n- Histogram of gradient values per layer\n- Helps identify vanishing/exploding gradient problems\n- Color-coded by layer\n\n#### 5. Training Progress\nReal-time monitoring of training metrics.\n\n**Features:**\n- Live plots of loss and accuracy\n- Updates after each epoch\n- Dual subplot layout\n\n## File Structure (Main)\n\n```\nneural-network-visualization-toolkit/\n├── app.py                      # Main Streamlit application\n├── main/\n│   ├── __init__.py\n│   ├── architecture.py         # Network architecture visualization\n│   ├── activations.py          # Activation map visualization\n│   ├── decision_boundaries.py  # Decision boundary plotting\n│   ├── gradients.py            # Gradient analysis\n│   └── training_process.py     # Training progress callback\n└── README.md\n```\n\n## Module Details\n\n### app.py\nMain application file containing:\n- Streamlit UI configuration\n- Model definitions\n- Data preparation\n- Visualization routing\n\n### main/architecture.py\n- **Function**: `visualize_arch(model)`\n- **Purpose**: Generates and displays network architecture diagrams\n- **Output**: PNG image of model structure\n\n### main/activations.py\n- **Function**: `visualize_act(model, input_data, layer_name)`\n- **Purpose**: Extracts and visualizes activation maps\n- **Parameters**: \n  - `model`: Keras model\n  - `input_data`: Input tensor\n  - `layer_name`: Target layer for visualization\n\n### main/decision_boundaries.py\n- **Function**: `plot_decision_boundary(model, X, y)`\n- **Purpose**: Plots classification decision boundaries\n- **Method**: Meshgrid prediction with contour plotting\n\n### main/gradients.py\n- **Function**: `visualize_grad(model, X, y)`\n- **Purpose**: Analyzes gradient distributions\n- **Method**: GradientTape for automatic differentiation\n\n### main/training_process.py\n- **Class**: `TrainingProcess(Callback)`\n- **Purpose**: Real-time training visualization\n- **Methods**:\n  - `on_epoch_end()`: Updates plots after each epoch\n  - `plot_progress()`: Generates loss/accuracy plots\n\n## Technical Requirements\n- Python 3.7+\n- TensorFlow 2.x\n- Streamlit\n- NumPy\n- Matplotlib\n- Scikit-learn\n- Keras\n\n## Sample Datasets\nThe application includes built-in datasets:\n- **MNIST-like**: 784-dimensional random data for simple networks\n- **Make Moons**: 2D classification dataset for decision boundaries\n- **ImageNet**: Pre-trained VGG16 for activation visualization\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fajitashwath%2Fnn-visualization","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fajitashwath%2Fnn-visualization","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fajitashwath%2Fnn-visualization/lists"}