I updated the Z-Image simple workflow. I added a Lora Loader, the ability to choose between using the gguf or safetensors version of the model (just bypass the one you don't want to use) upscaling using SeedVR2, ability to add a watermark (bypass if you don't want to use) and eye detailer.Feeling generous? <a rel="ugc" href="https://buymeacoffee.com/crod2730j">Buy Me Coffee</a>

Z-Image w/ SeedVR2 Upscaler, GGUF and Safetensors model supported

Added the following to v3<ul><li>WAN I2I Passthrough - zImage created in flow will get passed into a connected WAN I2I flow. Results will be a zImage generated image and a WAN generated image. WAN sampler may need to get tweaked based on the output you are looking for</li></ul>

Files with SHA256: 859aa81a1d0f3acd3c2da12e6c2b4be7bfc2975b6340d7336e68f5364fedfe3f