abenzerps

/

Qwen-Image-2.1-Uncensored-GGUF

Like 1.51k

Hi so do I install the model through like LM Studio? then install Comfy and Comfy-GGUF? or is it not even compatible for LM Studio?

Hey, it's not compatible with LM Studio. You can simply run it with Unsloth Desktop, or directly in ComfyUI using the ComfyUI-GGUF node.

I've tried LM Studio and ComfyUI and always get error "Failed to load model: The installed llama.cpp does not recognise this GGUF's model architecture ('qwen_image21')." How to solve that issue?

as I know, it can't be used in LM Studio, so you need to use ComfyUI; however, to use GGUF files there, you have to use the https://github.com/leejet/ComfyUI-GGUF repository—there is a guide on the model card explaining how to do it.

as I know, it can't be used in LM Studio, so you need to use ComfyUI; however, to use GGUF files there, you have to use the https://github.com/leejet/ComfyUI-GGUF repository—there is a guide on the model card explaining how to do it.

ComfyUI got the same error: # ComfyUI Error Report

Error Details

• Node ID: 72
• Node Type: UnetLoaderGGUF
• Exception Type: ValueError
• Exception Message: ValueError: Unexpected architecture type in GGUF file: 'qwen_image21'

Macbook pro M5Pro.

as I know, it can't be used in LM Studio, so you need to use ComfyUI; however, to use GGUF files there, you have to use the https://github.com/leejet/ComfyUI-GGUF repository—there is a guide on the model card explaining how to do it.

ComfyUI got the same error: # ComfyUI Error Report

Error Details

• Node ID: 72
• Node Type: UnetLoaderGGUF
• Exception Type: ValueError
• Exception Message: ValueError: Unexpected architecture type in GGUF file: 'qwen_image21'

Macbook pro M5Pro.

this happens because ComfyUI-GGUF is missing qwen_image21 in its supported architectures list.

Here is the quick fix:

1. Open ComfyUI/custom_nodes/ComfyUI-GGUF/loader.py
2. Go to line 12 and find IMG_ARCH_LIST
3. Add "qwen_image21" to the list so it looks like this:

IMG_ARCH_LIST = {"flux", "sd1", "sdxl", "sd3", "aura", "hidream", "cosmos", "ltxv", "hyvid", "wan", "lumina2", "qwen_image21"}

4. Save the file and restart ComfyUI.

(Alternatively, you can use qwen-image-2.1-UC-fp8.safetensors from this repo with ComfyUI's standard Load Diffusion Model node without needing the GGUF custom node at all).

Thank you so much! Now it works!! =))))

Hi so do I install the model through like LM Studio? then install Comfy and Comfy-GGUF? or is it not even compatible for LM Studio?

Hey, it's not compatible with LM Studio. You can simply run it with Unsloth Desktop, or directly in ComfyUI using the ComfyUI-GGUF node.

Ok thanks do I have to other Dependencies on Unsloth to run this exact model?

I haven't checked exactly, but I assume unsloth has included them in its own list to make downloading them easy.