<h2 id="a-custom-fork-of-comfyui-gguf-is-required">A custom fork of ComfyUI-GGUF is required</h2>I would suggest following these steps:<ul><li>Stop ComfyUI if it is running.</li><li>Open your terminal and cd to ComfyUI/custom_nodes.</li><li>Type the following command <code>git clone https://github.com/RealRebelAI/ComfyUI-GGUF_KREA-2.git</code></li><li>Restart ComfyUI</li></ul>Once clone the custom node folder structure should look like this:<pre><code>📂 ComfyUI/
├── 📂 custom_nodes/
│ ├── 📂 ComfyUI-GGUF_KREA-2/</code></pre><h2 id="installing-the-lmstudio-prompt-enhancer-custom-node.">Installing the LMStudio Prompt Enhancer custom node.</h2><ul><li>Click the "Extensions" button at the top left hand corner of ComfyUI.</li><li>Search for a Node Pack called "comfyui-lmstudio-prompt-enhancer "</li><li>Install "comfyui-lmstudio-prompt-enhancer "</li><li>Restart ComfyUI and refresh this page.</li></ul><a target="_blank" rel="ugc" href="https://github.com/AllenCraigBarnard/comfyui-lmstudio-prompt-enhancer">Learn more about this custom node</a>I found 2 models in LMStudio model catalogue that fit the 32GB or less requirement for this workflow they are:<ul><li>huihui-ornith-1.0-9b-abliterated-mtp you will want to set llm_max_tokens to 5000.</li><li>meta-llama-3.1-8b-instruct-abliterated you will want to set llm_max_tokens to 2500.</li></ul>🛈 IMPORTANT!!!You will need to search for, download, and load the model huihui-ornith-1.0-9b-abliterated-mtp or meta-llama-3.1-8b-instruct-abliterated in LM studio. You will also need to start the API server and generate an API Key. <a target="_blank" rel="ugc" href="https://github.com/AllenCraigBarnard/comfyui-lmstudio-prompt-enhancer#lm-studio-setup">Learn how to configure LM Studio here.</a><h3 id="notes-for-using-this-32gb-workflow">Notes for using this 32GB workflow</h3><ul><li>If the image looks flat, check the logs in LM Studio, the last log entry will contain a JSON path x.choices[0].message.content if the content field is empty that means the LLM ran out of reasoning tokens, this can be resolved by increasing max_tokens.</li></ul><h2></h2>

Files with SHA256: abf035412f09df3fbf03c7caa63ec46419d07425137c59ca6a175e96402e207b