LLM Easy prompt enhancer + modification (both with input image) (krea2/Minimax H3) - Minimax H3 v2.0 Showcase

Minimax H3 V2.0

What's new ?

• Upload up to 4 images to the LLM for prompt generator and modificator

• "Audio Generator" replace the previous "Text Voice Dialogue Generator" (it allows now to generate an audio file from an user input or a prompt, great to control sound and voices when using ref2video)

The "Audio Generator" need your Minimax diffusion model (you don't need to download another model just use the one you usually use). Same for CLIP and Audio VAE.

Audio Generation uses LLM to generate the prompt then Minimax model at min resolution and then extract audio to a file. It's quick but depends on the hardware (avg 15-20 sec for 15 sec audio with RTX 5080).

Hope it helps some of you, let me know in the comments if you have some feedbacks !