Phil2Sat

Only quants of https://huggingface.co/Phr00t/Qwen-Image-Edit-Rapid-AIO No AIO, just NSFW and Lightning loras included, packaging VAE and textencoder into GGUF is not possible ATM, needs the text encoder and the VAE. I personally recommend: Qwen2.5-VL-7B-Instruct-abliterated but the default textencoder does also fine. I have cloned the abliberated files into my repo and fixed the mmproj ready to use. HERE: https://huggingface.co/Phil2Sat/Qwen-Image-Edit-Rapid-AIO-GGUF/tree/main/Qwen2.5-VL-7B-Instruct-abliterated !!! PLACE THE mmproj FILE NEXT TO THE Qwen2.5-VL-7B-Instruct-abliterated.xxx.gguf !!! "mat1 and mat2 shapes cannot be multiplied..." is the error you get if you don't VAE https://huggingface.co/calcuis/pig-vae/blob/main/pigqwenimagevaefp32-f16.gguf Replace comfyextras/nodesqwen.py with https://huggingface.co/Phr00t/Qwen-Image-Edit-Rapid-AIO/blob/main/fixed-textencode-node/nodesqwen.py

This is a fork of the T5-XXL encoder that uses an extended tokenizer to facilitate Flux uncensoring and offer other benefits to tokenization. More info is on the project's GitHub page. As of 17.3.2025, third-party tools like ComfyUI and Kohya's training scripts don't yet support the new architecture. Their respective developers have been notified of this release, and may or may not choose to officially support it in a future update to their code. In the meantime, temporary patches for both tools are available in the GitHub repo linked above.