It's a workflow that I use primarily to work on prompts and test models via near-instant generation using the DMD2 lora and the LCM sampler. By activating ComfyUI's run (on change) function, you type your prompt and click outside the text window and your image will be instantly generated. Then, simply change the seed to check its consistency. The image is intentionally not saved, it's still possible to do so by right-clicking on the image in the "preview image" node and saving it or copying it into another workflow for img2img. <ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/1837476?modelVersionId=2079339">My SDXL merge model</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/tianweiy/DMD2/blob/main/dmd2_sdxl_4step_lora_fp16.safetensors">DMD2 Lora</a></li></ul>

Files with SHA256: 025e1a2b6705c90fca240601a584004e96185bdf162eaead0684f52583292309