CoolFace
Modelpublic

EllipsesMark/Minimax-h3_Singularity

sourceHugging Faceapache-2.0updated 11d agoView on Hugging Face
0likes952downloads
Model Card

Minimax-h3_Singularity

<p align="center"> <a href="https://www.runninghub.ai/post/2096339589492432897/?inviteCode=rh-v1559"> <img src="https://img.shields.io/badge/๐Ÿš€OnlineDemo-RunningHub-blue.svg" alt="Online Demo"> </a> <a href="https://youtube.com/@AIGC-Singularity"> <img src="https://img.shields.io/badge/YouTube-AIGC--Singularity-red.svg" alt="YouTube"> </a> <a href="https://space.bilibili.com"> <img src="https://img.shields.io/badge/Bilibili-AIGC--Singularity_Space-00a1d6.svg" alt="Bilibili"> </a> </p>

๐Ÿ“– Model Overview

Minimax-h3_Singularity is a comprehensive fine-tuned fusion model specialized in enhancing the capabilities of MiniMax-H3. Designed as a versatile multimodal video generation model, it natively supports Text-to-Video (T2V), Image-to-Video (I2V), Reference-to-Video (Ref2V), and Video-to-Video (V2V) workflows within ComfyUI.

Built upon a strategic fusion of key checkpoints (including ref, fl, b25-49, etc.), this model underwent deep high-step fine-tuning. To preserve the original model's foundational strengths and broad generalization while solving artifacts introduced by high-step training, we spent 3 full days on precise model pruning and weight optimization. The result is a clean, sharp, and highly dynamic video generation model.


โœจ Key Improvements & Features

  • โ€”๐ŸŽฌ HDR Image Quality & Blur Reduction: Fine-tuned on high-dynamic-range (HDR) video datasets to significantly enhance visual clarity and eliminate motion blur during high-speed action.
  • โ€”๐Ÿ‘ค Distant Face Restoration: Drastically reduces facial distortion, blurriness, and collapsing in medium-to-long shots.
  • โ€”๐ŸŽจ Clean & De-Oiled Aesthetic: Removes heavy, unnatural skin shine and glossy textures, rendering natural lighting and photorealistic materials.
  • โ€”โš”๏ธ Enhanced Dynamic Motion: Boosts motion fluidity and physical impact, excels in complex action sequences such as sword fighting and martial arts/melee combat.
  • โ€”๐ŸŒŒ VFX & Fantasy Effects: Specifically optimized for fantasy spellcasting, particle aura, and magical combat visual effects.
  • โ€”๐ŸŽญ Expressive Facial Dynamics: Captures subtle facial expressions and emotional nuances more vividly.
  • โ€”๐Ÿ“น Cinematography & Camera Control: Strengthens responsiveness to camera movements (pan, tilt, zoom, tracking shots) for cinematic storytelling.
  • โ€”๐Ÿ›ก๏ธ Full Base Capability Retention: 100% preserves MiniMax-H3's original prompt adherence, style adaptability, and base multimodal generation strength.

๐ŸŽฌ Showcase

<video src="https://huggingface.co/WarmBloodAban/Minimax-h3Singularity/resolve/main/video/1.mp4" controls autoplay loop muted width="20%"></video> <video src="https://huggingface.co/WarmBloodAban/Minimax-h3Singularity/resolve/main/video/3.mp4" controls autoplay loop muted width="20%"></video> <video src="https://huggingface.co/WarmBloodAban/Minimax-h3Singularity/resolve/main/video/AIGCTYD200001p84-audioganvc1788634594%20(1).mp4" controls autoplay loop muted width="20%"></video> <video src="https://huggingface.co/WarmBloodAban/Minimax-h3Singularity/resolve/main/video/AIGCTYD200010p87-audioalipa1788652671.mp4" controls autoplay loop muted width="20%"></video>


๐Ÿ’ก Usage Guide

Multimodal Pipeline Support

This model is fully compatible with ComfyUI and supports:

  • โ€”Text-to-Video (T2V)
  • โ€”Image-to-Video (I2V)
  • โ€”Reference-to-Video (Ref2V)
  • โ€”Video-to-Video (V2V)

๐Ÿš€ Recommended Acceleration LoRA

For high-speed generation with minimal quality loss, we strongly recommend pairing with:

  • โ€”`minimax_h3_ref2v_turbo_4step_v0.1` (Enables 4-step fast inference)

๐ŸŒ Online Interactive Demo

Test the model directly in your browser without local GPU setup: ๐Ÿ‘‰ **Try it on RunningHub Workflows**


๐Ÿ™ Acknowledgements

Special thanks to the MiniMax open-source team for creating and releasing the powerful MiniMax-H3 multimodal video model, providing a solid foundation for the open-source community! ๐Ÿค


๐Ÿค Community & Commercial Inquiries

Feel free to connect for tutorials, community discussions, workflow sharing, or commercial collaborations:

  • โ€”YouTube Channel: AIGC-Singularity
  • โ€”Bilibili Channel: AIGC-Singularity Space
  • โ€”QQ Group 1: 1058747239 (Request to join)
  • โ€”QQ Group 2: 1072010342 (Request to join)
  • โ€”Business Inquiries (WeChat): aigctyd
  • โ€”Email: a592991299@gmail.com