---
license: other
license_name: bespoke-lora-trained-license
license_link: https://multimodal.art/civitai-licenses
language:
- en
base_model:
- MiniMaxAI/MiniMax-H3
tags:
- safetensors
- lora
- t2i
- text-to-image
- diffusers
- photorealistic
- woman
- celebrity
- likeness
- MiniMax-H3
widget:
- text: >-
    AlexKingston001
  output:
    url: mmh3_Images/AlexKingston001.png
- text: >-
    AlexKingston002
  output:
    url: mmh3_Images/AlexKingston002.png
- text: >-
    AlexKingston003 
  output:
    url: mmh3_Images/AlexKingston003.png
- text: >-
    AlexKingston003
  output:
    url: mmh3_Images/AlexKingston004.png
---
# MiniMax H3 LoRA Models

=====================

<Gallery />

Testing out my first MiniMax H3 LoRA. I used Fizgig, the default settings, 150 images with captions, no video or audio, so no voice, 60 epochs but using epoch 51. Sorry, no prompts, they have characters the will not work on this page but still can be dragged into ComfyUI. There are some MP4s with the files as well with different samplers and schedulers. None of them got the bits of cotton flying around right.

=====================

Current MiniMax H3 LoRA Models:

Alex Kingston - 413xkin65t0n