suito-venus 5beeedf402 docs(readme): 日本語と英語のREADMEを分離し、テストファイルを削除
README.mdを英語版に全面書き換えし、日本語版をREADME-ja.mdとして新規作成。両言語で同じ内容をカバー(機能説明、使用方法、インストール手順、Gitフック設定、互換性、ライセンス情報)。pre-commitフック検証用のtest_file.txtを削除。
2026-02-11 19:24:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00
2026-02-11 18:59:59 +09:00

UNET Loader with Model Name

A ComfyUI custom node that extracts the model name as a string when loading UNET models (Diffusion Models).

Features

This node extends the standard UNETLoader (Load Diffusion Model) node and provides the following outputs:

  • MODEL: Standard model output
  • STRING: Model name (new feature)

Usage

  1. Open the loaders category in the node menu
  2. Select the Load Diffusion Model (with Name) node
  3. Choose a model from the dropdown
  4. Select the weight data type (default, fp8, etc.)
  5. Use the two outputs:
    • MODEL output → Connect to KSampler or other nodes
    • STRING output → Connect to Save Image filename prefix or metadata

Example Use Case

Include Model Name in Filename

Load Diffusion Model (with Name)
  ├─ MODEL → KSampler → VAE Decode → Save Image
  └─ STRING → (Connect to Save Image's filename_prefix)

This automatically includes the model name used in the generated image filename.

Output String Format

If the model filename is flux1-dev-Q8_0.gguf, the STRING output will be flux1-dev-Q8_0 (filename without extension).

Weight Data Type Options

  • default: Default precision
  • fp8_e4m3fn: FP8 format (E4M3)
  • fp8_e4m3fn_fast: FP8 format (E4M3) optimized version
  • fp8_e5m2: FP8 format (E5M2)

Installation

Simply place this folder in the ComfyUI/custom_nodes/ directory. The node will be automatically loaded when you restart ComfyUI.

For Developers: Git Hooks Setup

If you're contributing to this project, it's recommended to set up Git hooks that check for sensitive information before commits.

Method 1: Use Setup Script (Recommended)

# For Windows
setup-git-hooks.bat

# For Linux/Mac
chmod +x .git-hooks/pre-commit
chmod +x .git-hooks/pre-commit.py
git config core.hooksPath .git-hooks

Method 2: Manual Setup

git config core.hooksPath .git-hooks

Once the hooks are properly configured, the following checks will run automatically on commit:

  • Detection of sensitive information such as API keys, passwords, and tokens
  • Detection of files that should be excluded (__pycache__, .env, etc.)

For more details, see PRE_COMMIT_CHECKLIST.md.

Compatibility

  • Compatible with the latest version of ComfyUI
  • Uses the new node definition format with comfy_api.latest
  • Uses the same model folder (diffusion_models) as the standard UNETLoader

License

This custom node is based on ComfyUI's UNETLoader implementation.

  • License: GPL-3.0
  • Original Project: ComfyUI
  • Original License: GPL-3.0

For details, see the LICENSE file.

S
Description
No description provided
Readme
46 KiB
Languages
Python 65.6%
Shell 28.3%
Batchfile 6.1%