Minimax H3 Photorealistic Image Generator (LORA + Workflow)

Minimax H3 Photorealistic Image Generator (LORA + Workflow)

Consider to download the workflow (from 'optional downloads') for easier image extraction.

A conceptual lora and workflow to create high resolution photorealistic images with Minimax H3.

Remember image generation on WAN? Now you can generate images with H3 too.

Two LoRas have been trained on photorealistic datasets, care has been taken to generate as little motion as possible.

The workflow uses the photorealism image lora to capture the first and last frame out of a 22 frame video.

Pick either the 3780 or 3990 step variant. Lora 3990 seems to be more vibrant / contrasty and sharper detailed.

Example from Text to Image

Fun fact, you can also use the "Image2Video" mode to inject an input image into a new output. Results may vary, lora strength may play a factor. Consider it a nice bonus feature.

Usage:

• Preferably, use the following sentence to activate the image generation

  • ph0t0r34l

• Seems to work very well on resolutions up to 2560x1440 resolution

• Feel free to post your results, though keep it decent.

• As an experiment, you could try increasing/decreasing lora strength and maybe combining both loras with different strengths. 1.0 lora strenght is standard.

• Let me know your feedback, as this is the only photorealistic image generator that exists for H3 so far :)

Known issues:

• The workflow may complain about an image needed to be loaded, just feed it an image (ignore that it will have any effect though)

• You may need install some frequently used nodes, mostly Kijai's nodes and VideoHelperSuite

  • https://github.com/kijai/ComfyUI-KJNodes

  • 

Why make this?

I was curious to see what H3 could do when tricked into creating static images. I think these images look really good.

Comparisons:

Left is with lora, right is without lora.

Right click, open image in new tab, notice the difference in sharpness and contrast.