https://github.com/ktiseos-nyx/hfbackup_comfyuinode
ComfyUI Node for backing up checkpoints, and loras etc to huggingface
https://github.com/ktiseos-nyx/hfbackup_comfyuinode
Last synced: 3 months ago
JSON representation
ComfyUI Node for backing up checkpoints, and loras etc to huggingface
- Host: GitHub
- URL: https://github.com/ktiseos-nyx/hfbackup_comfyuinode
- Owner: Ktiseos-Nyx
- License: cc0-1.0
- Created: 2025-01-30T08:48:16.000Z (4 months ago)
- Default Branch: main
- Last Pushed: 2025-01-30T08:51:58.000Z (4 months ago)
- Last Synced: 2025-01-30T09:31:39.127Z (4 months ago)
- Language: Python
- Size: 0 Bytes
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Hugging Face Upload Node for ComfyUI
This is a custom node for ComfyUI that allows you to upload models (checkpoints or Diffusers models) directly to the Hugging Face Hub from your ComfyUI workflows.
## Features
* Uploads models to the Hugging Face Hub.
* Supports both checkpoint (`.ckpt`) and Diffusers model formats.
* Automatically creates a repository on the Hub if it doesn't exist.
* Optionally creates a model card with a preview image.
* Handles cases where no `model_path` is provided by using the ComfyUI output directory.
## Installation
1. **Clone this repository into your `ComfyUI/custom_nodes` directory:**
```bash
cd ComfyUI/custom_nodes
git clone https://github.com/Ktiseos-Nyx/HFBackup_ComfyUINode
```
2. **Install the required Python packages:**
This node requires the `huggingface_hub` and `Pillow` libraries. You need to install them manually in ComfyUI's Python environment.
**Steps:**
* **Find ComfyUI's Python Environment:** The environment is usually located in `ComfyUI/venv` or `ComfyUI/python_embeded` (especially on Windows).
* **Activate the Environment:**
* **Windows:**
```bash
ComfyUI\venv\Scripts\activate
# OR (if the above doesn't work)
ComfyUI\python_embeded\python.exe -m venv .venv
.\venv\Scripts\activate
```
* **macOS/Linux:**
```bash
source ComfyUI/venv/bin/activate
```
* **Install Dependencies:**
```bash
pip install huggingface_hub Pillow
```
3. **Restart ComfyUI:** Restart your ComfyUI instance so that it can load the new custom node and the installed dependencies.
## Usage
1. **Add the Node:** In the ComfyUI interface, add the "Hugging Face Upload" node to your workflow. You'll find it in the "utils" category.
2. **Connect Inputs:**
* **`hf_token`:** (Required) Your Hugging Face Hub write access token. Get it from your [Hugging Face profile settings](https://huggingface.co/settings/tokens).
* **`repo_id`:** (Required) The ID of your Hugging Face repository (e.g., `your_username/your_repo_name`).
* **`model_path`:** (Optional) The path to the model file or directory. If left empty, it defaults to the ComfyUI output directory.
* **`path_in_repo`:** (Optional) The path within the repository to store the model.
* **`model_type`:** Select either `ckpt` or `diffusers`.
* **`commit_message`:** A message for the commit.
* **`create_model_card`:** Check to create a model card.
* **`images`:** (Optional) Connect an `IMAGE` tensor (e.g., from a `Preview Image` node) to include a preview in the model card.
3. **Run the Workflow:** Execute your workflow. The node will upload the model to your specified Hugging Face Hub repository.
## Troubleshooting
* If you encounter errors, check the ComfyUI console for messages.
* Make sure you have activated the correct Python environment before installing dependencies.
* Ensure that your Hugging Face token has write access to the specified repository.
## Notes
* Uploading large models may take time. There is no progress bar in the ComfyUI interface.
* This README assumes you are familiar with basic ComfyUI concepts and have a Hugging Face Hub account.
## License
see the LICENSE file for details.