A ComfyUI implementation of Latent Bridge Matching (LBM) for efficient image relighting. This node utilizes the LBM algorithm to perform single-step image-to-image translation specifically for relighting tasks.
- 2025/05/22: Update ComfyUI-LBM to v1.1.0 ( update.md )
Added Depth/Normal map generation support
- Fast image relighting with a single inference step
- Simplified workflow with just one node
- Optimized memory usage
- Automatic model download - the model will be downloaded automatically and properly renamed on first use
- Support for depth and normal map generation
- Mask support for selective processing
- Multiple precision options (fp32, bf16, fp16)
- Clone this repository to your
ComfyUI/custom_nodesdirectory:
cd ComfyUI/custom_nodes
git clone https://github.com/1038lab/ComfyUI-LBM.git- Install the required dependencies:
cd ComfyUI/custom_nodes/ComfyUI-LBM
pip install -r requirements.txtThe models will be automatically downloaded and renamed on first use, or you can manually download them:
| Model | Description | Link |
|---|---|---|
| LBM Relighting | Main model for image relighting | Download |
| LBM Depth | Model for depth map generation | Download |
| LBM Normals | Model for normal map generation | Download |
After downloading, place the model files in your ComfyUI/models/diffusion_models/LBM directory.
- Add the "Relighting (LBM)" node from the
🧪AILab/🔆LBMcategory - Connect an image source to the node
- Select the model file (defaults to
LBM_relighting.safetensors) - Adjust the parameters as needed
- Run the workflow
- Add the "Depth / Normal (LBM)" node from the
🧪AILab/🔆LBMcategory - Connect an image source to the node
- Select the task type (depth or normal)
- Adjust the parameters as needed
- Run the workflow
| Parameter | Description | Default | Range |
|---|---|---|---|
| Model | The LBM model file to use | LBM_relighting.safetensors |
- |
| Steps | Number of inference steps | 28 | 1-100 |
| Precision | Inference precision | bf16 | fp32, bf16, fp16 |
| Bridge Noise Sigma | Controls diversity of results | 0.005 | 0.0-0.1 |
| Parameter | Description | Default | Range |
|---|---|---|---|
| Task | Select task type | depth | depth, normal |
| Steps | Number of inference steps | 28 | 1-100 |
| Precision | Inference precision | bf16 | fp32, bf16, fp16 |
| Bridge Noise Sigma | Controls diversity of results | 0.1 | 0.0-0.1 |
| Mask | Optional mask for selective processing | None | - |
| Setting | Recommendation |
|---|---|
| Steps | For most images, 20-30 steps provides a good balance between quality and speed |
| Input Resolution | The model works best with images of 512x512 or higher resolution |
| Memory Usage | If you encounter memory issues, try using fp16 precision or processing images at a lower resolution |
| Performance | For batch processing, consider reducing steps to 15-20 for faster throughput |
| Bridge Noise Sigma | Lower values (0.005) for relighting, higher values (0.1) for depth/normal maps |
This implementation uses the Latent Bridge Matching (LBM) method from the paper "LBM: Latent Bridge Matching for Fast Image-to-Image Translation". The model is designed for fast image relighting, transforming the lighting of objects in an image.
LBM offers:
- Fast processing with a single inference step
- High-quality relighting effects
- Memory-efficient operation
- Consistent results across various image types
- Support for depth and normal map generation
- Mask-based selective processing
The model is trained on a diverse dataset of images with different lighting conditions, ensuring:
- Balanced representation across different image types
- High accuracy in various scenarios
- Robust performance with complex lighting
Future plans for this repository include:
- LBM Depth - for depth map estimation
- LBM Normal - for normal map generation
- Additional optimization options
- ComfyUI
- Python 3.10+
- Required packages (automatically installed via requirements.txt):
- torch>=2.0.0
- torchvision>=0.15.0
- Pillow>=9.0.0
- numpy>=1.22.0
- huggingface-hub>=0.19.0
- tqdm>=4.65.0
- diffusers>=0.19.0
- accelerate>=0.20.0
- transformers>=4.30.0
- safetensors>=0.3.1
- requests>=2.25.0
- LBM Model: Hugging Face Model
- Original Implementation: GitHub Repository
- Paper: "LBM: Latent Bridge Matching for Fast Image-to-Image Translation" by Clément Chadebec, Onur Tasar, Sanjeev Sreetharan, and Benjamin Aubin
- Created by: 1038lab
This repository's code is released under the GNU General Public License v3.0 (GPL-3.0).
The LBM model itself is released under the Creative Commons BY-NC 4.0 license, following the original LBM implementation. Please refer to the original repository for more details on model usage restrictions.