Model for generation for all PCs. With updated text_encoder - umt5_xxl_fp8_e4m3fn_scaled.safetensors - <a target="_blank" rel="ugc" href="https://huggingface.co/OreX/Models/blob/main/WAN/umt5_xxl_fp8_e4m3fn_scaled.safetensors">WAN/umt5_xxl_fp8_e4m3fn_scaled.safetensors · OreX/Models at main</a>Model 1 -  <a target="_blank" rel="ugc" href="https://civarchive.com/models/1864281?modelVersionId=2110043">Qwen-Image - Qwen_Image_fp8_e4m3fn | Qwen Checkpoint | Civitai</a>Model 2 - <a target="_blank" rel="ugc" href="https://civarchive.com/models/1820829?modelVersionId=2060943">Wan2.2 I2V A14B GGUF - A14B Low | Wan Video Checkpoint | Civitai</a>Upscale models - <a target="_blank" rel="ugc" href="https://huggingface.co/Kim2091/UltraSharpV2/blob/main/4x-UltraSharpV2.safetensors">4x-UltraSharpV2.safetensors · Kim2091/UltraSharpV2 at main</a>Clip - <a target="_blank" rel="ugc" href="https://civarchive.com/models/1864281?modelVersionId=2110105">Qwen-Image - qwen_2.5_vl_7b_fp8_scaled | Qwen Checkpoint | Civitai</a>Vae - <a target="_blank" rel="ugc" href="https://civarchive.com/models/1864281?modelVersionId=2110009">Qwen-Image - Qwen Image VAE | Qwen Checkpoint | Civitai</a>ComfyUI: v0.3.60Step 1. Update the node in the workflowStep 2. Load the models and update umt5_xxl_fp8_e4m3fn_scaledStep 3. Don't touch KSampler and FaceDetailer until everything is setFor those with weak PCs, loading models and generation may take a long time, more than an hour.

Files with SHA256: 8aabc0f044677d75940387ff083a0ff2bf3f6e41ebb4c1a7fbf694fa7f5bc7b9