Instructions to use t8star/qwen-image-2.1-comfy with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use t8star/qwen-image-2.1-comfy with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf t8star/qwen-image-2.1-comfy:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf t8star/qwen-image-2.1-comfy:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf t8star/qwen-image-2.1-comfy:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf t8star/qwen-image-2.1-comfy:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Use Docker

docker model run hf.co/t8star/qwen-image-2.1-comfy:Q4_K_M

LM Studio
Jan
Ollama
How to use t8star/qwen-image-2.1-comfy with Ollama:
```
ollama run hf.co/t8star/qwen-image-2.1-comfy:Q4_K_M
```
Unsloth Desktop

How to use t8star/qwen-image-2.1-comfy with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "t8star/qwen-image-2.1-comfy:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use t8star/qwen-image-2.1-comfy with Docker Model Runner:
```
docker model run hf.co/t8star/qwen-image-2.1-comfy:Q4_K_M
```

Lemonade

How to use t8star/qwen-image-2.1-comfy with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull t8star/qwen-image-2.1-comfy:Q4_K_M

Run and chat with the model

lemonade run user.qwen-image-2.1-comfy-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use t8star/qwen-image-2.1-comfy with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default t8star/qwen-image-2.1-comfy:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use t8star/qwen-image-2.1-comfy with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf t8star/qwen-image-2.1-comfy:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "t8star/qwen-image-2.1-comfy:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Qwen Image 2.1 Prompt Enhancer GGUF for ComfyUI / Qwen Image 2.1 提示词改写模型

This repository is a convenience mirror of three existing Q4_K_M prompt-enhancer GGUF models and one BF16 vision projector for ComfyUI-Qwen-Image-Prompt-Rewrite-T8, also available in the ComfyUI Registry. The files were not trained or quantized by T8star. They retain their original filenames and SHA256 checksums. This is a prompt-rewriting collection, not the Qwen Image 2.1 image-generation diffusion model, text encoder, or VAE.

本仓库集中存放用于 ComfyUI 提示词改写节点的三个 Q4_K_M GGUF 主模型及一个 BF16 视觉组件。T8star 并非原模型训练者或量化作者。这里的权重只负责改写提示词，不能单独出图。

File / 文件	Purpose / 用途	Original repository / 原仓库	SHA256
`Qwen-Image-2.1-PE-T2I.Q4_K_M.gguf`	Standard text-to-image rewriting / 标准文生图改写	prithivMLmods/Qwen-Image-2.1-PE-T2I-GGUF	`340feb42c784e35b704a0f0d8d1c679a3fe60437bd9ea44a7a6fd2d730470506`
`Qwen-Image-2.1-PE-I2I.Q4_K_M.gguf`	Image-aware edit rewriting / 看图编辑改写	prithivMLmods/Qwen-Image-2.1-PE-I2I-GGUF	`a5c6cb28cbaf838834d1335c8392619af5809d9fe95d0ebd321e15751866dfe9`
`Qwen-Image-2.1-PE-I2I.mmproj-bf16.gguf`	I2I vision projector / I2I 视觉组件	prithivMLmods/Qwen-Image-2.1-PE-I2I-GGUF	`8dedb71dbc3092dc47de9108ad373d68a12854e2527d59bd9399601738f3bce1`
`pe_t2i_heretic-Q4_K_M.gguf`	Optional Heretic T2I rewriting / 可选 Heretic 文生图改写	pottokao/Qwen-Image-2.1-PE-T2I-Heretic-GGUF	`fe176ded062942ac8858290a33c9303a6e9c09c31405020a22ad0d47fa7b9b69`

The original Qwen prompt-enhancer checkpoints are Qwen/Qwen-Image-2.1-PE-T2I and Qwen/Qwen-Image-2.1-PE-I2I. The Heretic GGUF is based on pottokao/Qwen-Image-2.1-PE-T2I-Heretic. Built with Qwen. See each source repository for its own authorship and technical details.

Install / 安装

Download the four files into ComfyUI/custom_nodes/Comfyui-Qwen-Image-Prompt-Rewrite-T8/models/llm/qwenimage-pe/, or into ComfyUI/models/llm/qwenimage-pe/. The node can also read a directory selected by QWEN_PE_MODEL_DIR. Use the node repository for installation steps, runtime requirements, and drag-and-drop workflows. The I2I model requires the matching mmproj-bf16.gguf; the source repositories do not provide a Q4 vision projector. Do not pair this projector with either T2I weight.

将四个文件放进节点目录的 models/llm/qwenimage-pe/，也可放进 ComfyUI/models/llm/qwenimage-pe/，或用 QWEN_PE_MODEL_DIR 指定目录。带图编辑必须同时使用对应的 BF16 mmproj。节点 README 提供安装和工作流说明。

The default node profile uses the standard T2I model for text-only requests and I2I plus its vision projector for requests with 1–10 reference images. Heretic is an optional T2I alternative; it has not demonstrated a reliable quality advantage in this project's small comparison. Image generation still needs a separate Qwen Image 2.1 diffusion pipeline.

License and redistribution / 许可与转载

These Qwen-derived weights are distributed under the Qwen Research License Agreement, included as LICENSE. It grants use and redistribution for non-commercial research or evaluation only; commercial use requires a separate license from the Qwen rightsholder. Preserve the agreement and NOTICE when redistributing. No commercial-use license is supplied by this mirror. Source and derivative authors retain their respective rights.

本镜像附带 Qwen Research 许可证和 NOTICE。其授权范围为非商业研究或评估；商业使用需另向权利人取得许可。本页的其他链接不改变模型许可。

T8star links / 相关链接

Downloads last month: 1,774

GGUF

Model size

9B params

Architecture

qwen35

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support