<h3 id="comfyui-workflow-for-image-generation">Comfyui Workflow for Image Generation</h3><ul><li>needed custom node:<ul><li>Github: <a target="_blank" rel="ugc" href="https://github.com/r-vage/ComfyUI-RvTools-X">RvTools-X (Version 2.5.10)</a></li><li>ComfyUI Registry: <a target="_blank" rel="ugc" href="https://registry.comfy.org/nodes/ComfyUI-RvTools-X">RvTools-X</a></li><li>If you use the Manager to intall the node choose nightly to install the latest Version from github</li></ul></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/articles/15787/listing-links-resources">List of Resources</a> accross the different model types</li></ul>

iGEN II-X for FLUX | SD | HiDream | Qwen (STD | GGUF)

<ul><li>this update needs RvTools-X v2.5.10<ul><li>i've added some multi-switches that allows me to use mute now in the sampler group</li><li>updated replace string v2</li></ul></li><li>i've reworked the sampler group to simplify the connections with the new multi-switches, this group can use mute now instead of bypass. the connections are also cleaner now i.e. the conditionings are not passed through different nodes anymore.</li><li>i've also removed the neg. conditioning zero out from the initial render group. it is now in the prompt group. the reason for that is that you can disable the negative prompt now for flux models. in the versions before the negative conditioning was always encoded just to zero it out -.^ if negative prompt is disabled, the positive conditioning is zeroed out and used as negative conditioning. When enabled, the Initial Render option also mutes the text encodings in the prompt area, allowing you to load an image and bypass the sampling process, but still use the prompt nodes and continue sending the image to the Detailers.</li><li>i've replace pulid for sdxl because it sux anyways, instead i've added face swap for input images. i've also added a video input to use specific frames, send them to swap and to the sampler group. can be used i.e. with denoise to rerender or improve the image, with flux depth lora, redux, kontext or whatever.</li></ul>

Files with SHA256: 1fd0fd3d6e6070ee1b38059d8243c36aa131b9b1fd1da9e47cdb48227969dfb4