{"id":26544729,"url":"https://github.com/ryanontheinside/ComfyUI_SuperResolution","last_synced_at":"2025-03-22T04:01:49.653Z","repository":{"id":282740332,"uuid":"949514479","full_name":"ryanontheinside/ComfyUI_SuperResolution","owner":"ryanontheinside","description":"Super Resolution Implementation for ComfyUI","archived":false,"fork":false,"pushed_at":"2025-03-16T16:26:21.000Z","size":6,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-03-16T17:29:07.296Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"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/ryanontheinside.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}},"created_at":"2025-03-16T16:23:55.000Z","updated_at":"2025-03-16T16:26:24.000Z","dependencies_parsed_at":"2025-03-16T17:39:28.451Z","dependency_job_id":null,"html_url":"https://github.com/ryanontheinside/ComfyUI_SuperResolution","commit_stats":null,"previous_names":["ryanontheinside/comfyui_superresolution"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryanontheinside%2FComfyUI_SuperResolution","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryanontheinside%2FComfyUI_SuperResolution/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryanontheinside%2FComfyUI_SuperResolution/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ryanontheinside%2FComfyUI_SuperResolution/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ryanontheinside","download_url":"https://codeload.github.com/ryanontheinside/ComfyUI_SuperResolution/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244902929,"owners_count":20529115,"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":[],"created_at":"2025-03-22T04:01:48.775Z","updated_at":"2025-03-22T04:01:49.647Z","avatar_url":"https://github.com/ryanontheinside.png","language":"Python","funding_links":[],"categories":["Workflows (3395) sorted by GitHub Stars","All Workflows Sorted by GitHub Stars"],"sub_categories":[],"readme":"# ComfyUI Super Resolution\n\nA collection of high-performance neural network-based Super Resolution models for ComfyUI.\n\n## Features\n\n- **Multiple SR Models**: Choose from FSRCNN, ESPCN, LapSRN, and EDSR with different quality/speed tradeoffs\n- **Modular Design**: Load models once and reuse them across multiple upscaling operations\n- **CUDA Acceleration**: Automatic GPU acceleration when available\n- **Multiple Scale Factors**: Support for 2x, 3x, and 4x upscaling\n\n## Installation\n\n1. Clone this repository to your ComfyUI custom_nodes directory:\n   ```\n   cd ComfyUI/custom_nodes\n   git clone https://github.com/yourusername/ComfyUI_SuperResolution\n   ```\n\n2. Restart ComfyUI - the necessary models will be automatically downloaded on first use\n\n## Usage\n\n1. Add a **SR Model Loader** node to your workflow\n2. Select your desired model type and scale factor\n3. Connect the model output to a **SR Upscale** node\n4. Connect an image to the upscale node\n5. Run your workflow to get high-quality upscaled images!\n\n## Model Comparison\n\n| Model | Architecture | Features | Best For | Speed | Quality |\n|-------|-------------|----------|----------|-------|---------|\n| **FSRCNN-small** | Lightweight CNN | Fast, minimal memory use | Real-time processing, mobile | ★★★★★ | ★★ |\n| **FSRCNN** | CNN with larger features | Good balance of speed/quality | General purpose | ★★★★ | ★★★ |\n| **ESPCN** | Sub-pixel convolutions | Efficient upscaling at end | Text/line drawings | ★★★★ | ★★★ |\n| **VDSR** | Very deep CNN | Better edge reconstruction | Detailed images with edges | ★★★ | ★★★★ |\n| **LapSRN** | Laplacian pyramid | Progressive upscaling | Sharp edges, details | ★★★ | ★★★★ |\n| **EDSR** | Deep residual network | Most parameters, best quality | Maximum detail | ★★ | ★★★★★ |\n\n## Technical Details\n\nEach model has unique architectural characteristics:\n\n1. **FSRCNN / FSRCNN-small**: Direct mapping from low to high resolution, lightweight with fewer parameters.\n\n2. **ESPCN**: Uses the efficient sub-pixel convolution technique that processes at low resolution and only expands dimensions at the final layer.\n\n3. **VDSR**: Very Deep Super Resolution network with 20 convolutional layers, uses global residual learning.\n\n4. **LapSRN**: Uses a Laplacian pyramid structure to progressively upscale images, which preserves edges better than the others.\n\n5. **EDSR**: Enhanced Deep Super-Resolution, a residual network with significantly more parameters, offering the highest quality but slower processing.\n\n## Performance Expectations (RTX 4090)\n\n| Model | 2x Upscaling | 4x Upscaling | Memory Usage |\n|-------|--------------|--------------|--------------|\n| FSRCNN-small | 200+ FPS | 150+ FPS | \u003c 100MB |\n| FSRCNN | 100+ FPS | 80+ FPS | \u003c 150MB |\n| ESPCN | 90+ FPS | 70+ FPS | \u003c 150MB |\n| LapSRN | 60+ FPS | 40+ FPS | \u003c 200MB |\n| EDSR | 40+ FPS | 25+ FPS | \u003c 500MB |\n\n## When to Use Each Model\n\n- **FSRCNN-small**: When maximum speed is required\n- **FSRCNN**: For a good balance of quality and performance\n- **ESPCN**: For text and line art upscaling\n- **LapSRN**: When you need better edge preservation than FSRCNN\n- **EDSR**: When maximum quality is desired and performance is secondary\n\n## Comparison to Traditional Methods\n\nThese neural network-based upscaling methods offer significantly better quality than traditional algorithms like Lanczos, Bicubic, or Bilinear interpolation, while often being just as fast thanks to CUDA acceleration.\n\n## Credits\n\nThis nodepack implements models originally created by:\n\n- FSRCNN: Dong et al. (https://github.com/ryanontheinside/FSRCNN_Tensorflow)\n- EDSR: Lim et al. (https://github.com/ryanontheinside/EDSR_Tensorflow)\n- ESPCN: Shi et al. (https://github.com/ryanontheinside/TF-ESPCN)\n- LapSRN: Lai et al. (https://github.com/ryanontheinside/TF-LapSRN)\n\n## License\n\nMIT ","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fryanontheinside%2FComfyUI_SuperResolution","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fryanontheinside%2FComfyUI_SuperResolution","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fryanontheinside%2FComfyUI_SuperResolution/lists"}