VideoRebirthAI/hyperflow-h3-flash
AI model · https://huggingface.co/VideoRebirthAI/hyperflow-h3-flash
First seen by Oriel on .
AI summary
HyperFlow H3 Flash is a compressed, few-step version of MiniMax-H3 that generates video and synchronized stereo audio, targeting consumer and data-center GPUs.
- It supports text-to-video, keyframe-based generation, and reference-conditioned generation, with clips of 5.2 to 15 seconds at 24 fps and 32 kHz stereo audio.
- Sampling uses 4, 6, or 8 steps without classifier-free guidance; the 6-step default generates a 5.2-second clip in 8.7 seconds on 4× B200 GPUs.
- The model has 13.7B or 13.9B parameters; GGUF weights range from 7.2 to 15.2 GB, while INT8 weights are about 16 GB.
- GGUF supports ComfyUI, Unsloth Studio, and stable-diffusion.cpp; the Python package and diffusers support INT8, including video and audio references for ref2va.
- The MiniMax H3 Community License Agreement restricts use in several territories without prior authorization and requires disclosure of publicly distributed AI-generated content.
#video generation #audio generation #diffusers #gguf
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Likes
Total likes reported by the source.
- Hugging Face Models: 333, latest observation ; +232 over observed period since
Downloads (30d)
Rolling 30-day downloads. A change in this value is not new downloads during the observed period.
- Hugging Face Models: 572,345, latest observation ; +531,557 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Hugging Face Models | VideoRebirthAI/hyperflow-h3-flash | https://huggingface.co/VideoRebirthAI/hyperflow-h3-flash Read discussion |