ComfyUI Workflow<ul><li>Needed Custom Node:<ul><li>Git-Hub: <a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI-RvTools_v2">RvTools_v2 (Updated: 2.4.37)</a></li><li>Comfyui-Registry: <a target="_blank" rel="ugc" href="https://registry.comfy.org/nodes/comfyui-rvtools_v2">RvTools_v2</a></li><li>if you install the node with the manager choose nightly because i don't upload every change to comfyui registry.</li></ul></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/15787/listing-links-resources">Resource Listing</a></li><li>How to install <a target="_blank" rel="ugc" href="https://www.reddit.com/r/StableDiffusion/comments/1h7hunp/how_to_run_hunyuanvideo_on_a_single_24gb_vram_card/">Triton &amp; Sage 2 </a>and <a target="_blank" rel="ugc" href="https://huggingface.co/lldacing/flash-attention-windows-wheel/tree/main">Flash Attention 2 Wheels (for e.g. Florence2)</a></li><li>when using torch 2.7.x+cu128 use triton 3.3.1 (3.4 will cause issues) torch compile don't work<ul><li>-m pip install -U "triton-windows==3.3.1.post19"</li></ul></li></ul>

iGEN II for FLUX/SD/HiDream/Qwen (STD/GGUF)

<ul><li>updated all model loader to multi gpu for a better memory managment</li><li>added flux torch compile to the detailer groups for more speed</li><li>removed the yolo detailers in my tests they returning errors and florence2 with the flux large model detects almost everything</li><li>the other groups are untouched</li></ul>

Files with SHA256: fae67aa3713fcc2b6d654f820b0c1aca5084dfc6ba2c681269806e2c9c9168cb