Model for generation for all PCs. With updated text_encoder - umt5_xxl_fp8_e4m3fn_scaled.safetensors - WAN/umt5_xxl_fp8_e4m3fn_scaled.safetensors · OreX/Models at main

Model 1 -  Qwen-Image - Qwen_Image_fp8_e4m3fn | Qwen Checkpoint | Civitai

Model 2 - Wan2.2 I2V A14B GGUF - A14B Low | Wan Video Checkpoint | Civitai

Upscale models - 4x-UltraSharpV2.safetensors · Kim2091/UltraSharpV2 at main

Clip - Qwen-Image - qwen_2.5_vl_7b_fp8_scaled | Qwen Checkpoint | Civitai

Vae - Qwen-Image - Qwen Image VAE | Qwen Checkpoint | Civitai

ComfyUI: v0.3.60

Step 1. Update the node in the workflow

Step 2. Load the models and update umt5_xxl_fp8_e4m3fn_scaled

Step 3. Don't touch KSampler and FaceDetailer until everything is set

For those with weak PCs, loading models and generation may take a long time, more than an hour.