Extensions/ComfyUI LLM SDXL Adapter
ComfyUI Extension

ComfyUI LLM SDXL Adapter

A comprehensive set of ComfyUI nodes for using Large Language Models (LLM) as text encoders for SDXL image generation through a trained adapter.

By NeuroSenko·Created about a year ago·Updated 10 months ago· 68
NeuroSenko/ComfyUI_LLM_SDXL_Adapter
Nodes9
On cloudLocal install
Categoryllm_sdxl
Stars68
Updated10 months ago

Nodes (9)

Apply LLM To SDXL Adapter

The bridge that turns LLM embeddings into SDXL conditioning

llm_sdxl
LLM Adapter Loader

Where the trained weights actually come in

llm_sdxl
LLM Adapter Loader Custom

Same adapter loader, but every knob is yours now

llm_sdxl
LLM GGUF Model Loader

The SDXL text encoder that fits on a low-VRAM card

llm_sdxl
LLM Model Loader

The plain LLM loader this whole pack is built around

llm_sdxl
LLM Text Encoder

SDXL finally reads full sentences instead of tag soup

llm_sdxl
Apply T5Gemma LLM to Adapter

T5Gemma's handoff to the sampler, size conditioning included

llm_sdxl
T5Gemma Loader

Load the encoder half of Google's T5Gemma as your SDXL text encoder

llm_sdxl
T5Gemma Text Encoder

The T5Gemma branch's text front end

llm_sdxl
Readme

ComfyUI LLM SDXL Adapter

Version License Python ComfyUI

A comprehensive set of ComfyUI nodes for using Large Language Models (LLM) as text encoders for SDXL image generation through a trained adapter.

<img width="1803" height="904" alt="image" src="https://github.com/user-attachments/assets/e8e5f047-37e7-4f8b-9bbd-78d70e2a7d80" />

Image with workflow

🎯 Available Adapters

RouWei-Gemma Adapter

Trained adapter for using Gemma-3-1b as text encoder for Rouwei v0.8 (vpred or epsilon or base).

Download Links:

📦 Installation

Install Nodes

  1. Clone the repository to ComfyUI/custom_nodes/:
cd ComfyUI/custom_nodes/
git clone https://github.com/NeuroSenko/ComfyUI_LLM_SDXL_Adapter.git
  1. Restart ComfyUI

Setup RouWei-Gemma Adapter

  1. Download the adapter:

    • Download from CivitAI or HuggingFace
    • Place the adapter file in ComfyUI/models/llm_adapters/

  2. Download Gemma-3-1b-it model:

    • Download gemma-3-1b-it (non-gated mirror)
    • Place in ComfyUI/models/llm/gemma-3-1b-it/
    • Note: You need ALL files from the original model for proper functionality (not just .safetensors)

  3. Download Rouwei checkpoint:

    • Get Rouwei v0.8 (vpred, epsilon, or base) if you don't have it
    • Place in your regular ComfyUI checkpoints folder

📁 File Structure Example

ComfyUI/models/
├── llm/gemma-3-1b-it/
│   ├── added_tokens.json
│   ├── config.json
│   ├── generation_config.json
│   ├── model.safetensors
│   ├── special_tokens_map.json
│   ├── tokenizer.json
│   ├── tokenizer.model
│   └── tokenizer_config.json
├── llm_adapters/
│   └── rouweiGemma_g31b27k.safetensors
└── checkpoints/
    └── rouwei_v0.8_vpred.safetensors

🔍 Debugging

To enable detailed logging, edit __init__.py:

# Change from:
logger.setLevel(logging.WARN)
# To:
logger.setLevel(logging.INFO)