<h2 id="clone-into-custom-nodes:"><strong>Clone into custom nodes:</strong></h2><h2 id="https:github.commidfr-caimidfr-comfyui-nodes"><a target="_blank" rel="ugc" href="https://github.com/MidFR-CAI/midfr-comfyui-nodes">https://github.com/MidFR-CAI/midfr-comfyui-nodes</a></h2><p>I'm gonna keep updating these as I am also using them. Comes with a 'LoRA Vault' for use with Firefox to manage LoRAs without having to keep them all downloaded.</p><h3 id="moss-single-image-edit-notes">MOSS Single Image Edit Notes</h3><p>In the LoRA loader for styles or concepts, styles should load starting at layer 14, and probably at a lower weight.</p><p>Edit LoRA from:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/akhaliq/MOSS-Image-Edit-HQ">https://huggingface.co/akhaliq/MOSS-Image-Edit-HQ</a></p><p>In models/LLM, needs LLM from :</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/svjack/Qwen3-VL-4B-Instruct-heretic-7refusal">https://huggingface.co/svjack/Qwen3-VL-4B-Instruct-heretic-7refusal</a><br /></p><h3 id="minimax-h3-ref2v-notes">Minimax H3 Ref2V Notes</h3><p>In comfyui/models/gguf:</p><p>Need Q8 gguf from:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/DreamFast/Qwen3-VL-4b-Heretic-GGUF/tree/main">https://huggingface.co/DreamFast/Qwen3-VL-4b-Heretic-GGUF/tree/main</a></p><p>Need mmproj Q8 gguf from:</p><p><a target="_blank" rel="ugc" href="https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/tree/main">https://huggingface.co/Qwen/Qwen3-VL-4B-Instruct-GGUF/tree/main</a></p><p>Clone into comfyui/custom_nodes:</p><p><a target="_blank" rel="ugc" href="https://github.com/T8mars/comfyui-minimax-h3-audio-T8">https://github.com/T8mars/comfyui-minimax-h3-audio-T8</a></p>

ComfyUI Workflows

<p>Mostly just changes to the Autoprompt nodes</p><p>Also removed spectrum for turbo as it's a decent quality hit for a small performance boost at low steps</p>

Files with SHA256: 5c01785e8af2b932bccace818cdee2e3c804911caf3ea220e77bcb1caa2ce103