AlperKTS/Qwen-Image-2.1-GGUF

Qwen-Image 2.1 GGUF (ComfyUI)

Quantized GGUF weights of the state-of-the-art Qwen-Image 2.1 diffusion model for ComfyUI using ComfyUI-GGUF.

These models allow running Qwen-Image 2.1 on consumer GPUs with significantly reduced VRAM footprint while preserving maximum visual quality, text-rendering capabilities, and composition fidelity.

---

📦 Quantization Versions

File	Size	VRAM Recommended	Description
qwen_image_2.1_Q8_0.gguf	~7.2 GB	12GB+	Near-lossless precision. Best quality & fidelity.
qwen_image_2.1_Q6_K.gguf	~5.5 GB	8GB - 12GB	Optimal balance of high quality and memory efficiency.
qwen_image_2.1_Q5_K_M.gguf	~4.6 GB	8GB	Fast inference, lowest VRAM consumption, excellent quality.

Note: The original BF16 diffusion model is ~14.2 GB.

---

🚀 How to Use in ComfyUI

1. Install Prerequisites

Make sure you have ComfyUI-GGUF installed in your custom_nodes folder:

cd ComfyUI/custom_nodes
git clone https://github.com/city96/ComfyUI-GGUF

2. Download Required Files

1. Diffusion Model (GGUF):

   • Download any of the .gguf files from this repo (e.g. qwen_image_2.1_Q8_0.gguf).
   • Place it in: ComfyUI/models/diffusion_models/ (or ComfyUI/models/unet/)

2. VAE:

   • Download qwen_image_2.1_vae_bf16.safetensors from Comfy-Org/Qwen-Image-2.1.
   • Place it in: ComfyUI/models/vae/

3. Text Encoder / CLIP:

   • Download qwen3vl_8b_fp8_scaled.safetensors or qwen3vl_8b_int8_convrot.safetensors (or qwen3vl_8b_bf16.safetensors) from Comfy-Org/Qwen-Image-2.1.
   • Place it in: ComfyUI/models/text_encoders/ (or ComfyUI/models/clip/)
   • Note: In ComfyUI's CLIPLoader, select loader type qwen_image.

---

🎨 Workflows

Pre-configured workflows are included in the workflows/ folder:

• Qwen_Image_2.1_GGUF_Text_to_Image.json: Ready-to-run ComfyUI text-to-image workflow using UnetLoaderGGUF.
• Qwen_Image_2.1_GGUF_Image_Edit.json: Ready-to-run ComfyUI image-to-image / editing workflow using UnetLoaderGGUF.
• Drag and drop either workflow JSON directly into ComfyUI to start generating immediately.

Key Node Configuration

• Node: UnetLoaderGGUF (from ComfyUI-GGUF)
• unet_name: Select qwen_image_2.1_Q8_0.gguf (or Q6_K / Q5_K_M)
• Connect to standard KSampler / ModelSamplingDiscrete nodes.

---

📜 Credits & License

• Original model created by the Qwen Team (Alibaba).
• GGUF conversion & quantization by AlperKTS.
• Compatible with ComfyUI-GGUF by city96.
• Licensed under the original Qwen Research License.