| # ComfyUI-RealRestorer |
|
|
| Standalone ComfyUI node pack for **RealRestorer** image restoration. |
|
|
| **No vendored diffusers fork.** All model code is self-contained. Safe to install |
| into any existing ComfyUI environment without breaking dependencies. |
|
|
| - Paper: https://arxiv.org/abs/2603.25502 |
| - Model: https://huggingface.co/RealRestorer/RealRestorer |
| - Official repo: https://github.com/yfyang007/RealRestorer |
|
|
|
|
| ## Why This Exists |
|
|
| The official RealRestorer repo requires a patched fork of `diffusers` that will |
| break existing ComfyUI environments. This node pack reimplements the full pipeline |
| from scratch using only standard PyTorch, transformers, einops, and safetensors. |
|
|
|
|
| ## Requirements |
|
|
| Everything is already in a standard ComfyUI venv. No `pip install` needed. |
|
|
|
|
| ## VRAM Usage |
|
|
| At 1024x1024 with bfloat16: |
|
|
| | Strategy | Peak VRAM | Speed | |
| |---|---|---| |
| | full_gpu | ~34 GB | Fastest | |
| | offload_to_cpu | ~18 GB peak | Slower | |
| |
| With 96GB VRAM, `full_gpu` is recommended. |
|
|
|
|
| ## Credits |
|
|
| - RealRestorer: Yufeng Yang et al. (https://arxiv.org/abs/2603.25502) |
| - Base model: Step1X-Edit by StepFun AI |
| - ComfyUI implementation: Built with Claude (Anthropic) |
|
|
