text encoder gguf version quant 4 qwen 3 vl ?
hi is possible to make ggu version of text encoder and use it at comfui ? thanks
I was looking for that too, but ComfyUI does only safetensor files for models, not gguf. So?
Not all of them can be used in ComfyUI and that is sadly no part of the search results.
The full-text search doesn't sort tags after the importance in the search line.
Nevertheless... The file I chose works in ComfyUI:
https://huggingface.co/mradermacher/Qwen3-VL-4B-Instruct-c_abliterated-v2-GGUF
https://huggingface.co/mradermacher/Qwen3-VL-4B-Instruct-c_abliterated-v2-GGUF/blob/main/Qwen3-VL-4B-Instruct-c_abliterated-v2.Q8_0.gguf
https://huggingface.co/mradermacher/Qwen3-VL-4B-Instruct-c_abliterated-v2-GGUF/blob/main/Qwen3-VL-4B-Instruct-c_abliterated-v2.mmproj-Q8_0.gguf
You select one .gguf model of your preference (Q4/Q8 etc.) and you also need the .mmproj-f16.gguf file from there. Both files have to be in the same folder within your Comfy's text_encoders folder. You may need to use another custom-node for gguf (and delete the one from city96 if you use that one) to make this work, read the discussion about it here: https://github.com/city96/ComfyUI-GGUF/issues/464 - it includes the command to clone a fork from @city96 's repository by @m8rr.