SwiftVR is a streaming one-step generative video restoration (VR) framework presented in SwiftVR: Real-Time One-Step Generative Video Restoration. - [2026/06] Release the inference code and pretrained weights LightX2V brings faster inference and lower GPU memory usage to SwiftVR (1.91× speedup and 65.71% lower peak GPU memory per request on a single H100). It also supports multi-GPU acceleration. Ready-to-use scripts cover image and video super-resolution, offline inference, and API serving. Get started with LightX2V → - Mask-free shifted-window self-attention (MFSWA). Each spatial window is pre-gathered into a dense tensor, so every attention call reduces to a single standard…
Viggle-Animate is an open-weight model for video to video from Viggle AI, released under other. It has 33.1B parameters. At 16-bit it needs about 79.5 GB of GPU memory, which fits on 1x MI300X from $1.85 an hour, at the lowest prices in the SAVRN Index. It draws 32k downloads a month.
Built on MiniMaxAI/MiniMax-H3 this repository. The cast of a two-minute music video replaced through its fast cuts, from at most one repainted frame per shot. AI-generated parody: the people shown did not take part in, approve or endorse it.
Parameters33.1B
Context—
Weights68.9 GB
Licenseother
AccessOpen weights
Monthly Downloads32k
Runs On
What it takes to serve Viggle-Animate (33.1B parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.
| Precision | Weights | Memory needed | Cheapest setup | Per hour | Also fits |
|---|---|---|---|---|---|
| 16-bit | 66.2 GB | 79.5 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 8-bit | 33.1 GB | 39.7 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
| 4-bit | 16.6 GB | 19.9 GB | 1x MI300X (192 GB) Vultr |
$1.85 | 1x H100 $1.99 · 1x MI325X $2.00 |
Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Oct 18, 2026.
Viggle-Animate on every accelerator the SAVRN Index prices, at every precision
Model Card
Built on MiniMaxAI/MiniMax-H3 this repository. The cast of a two-minute music video replaced through its fast cuts, from at most one repainted frame per shot. AI-generated parody: the people shown did not take part in, approve or endorse it. Viggle-Animate replaces the character in a video with whatever you paint into one of its own frames — motion, camera and timing untouched. You prepare that one frame in an image editor; from there the video stage runs no pose estimator, segmenter, face tracker or text encoder. Two inputs, three forward passes, 26 seconds a shot on one GPU. It is strongest where replacement is hardest: fast motion, and pose transfer accurate enough to follow it. Whipping…
Excerpt from the card by Viggle AI, licensed other.
Identity and Version
- Repository
- Viggle/Viggle-Animate
- Publisher
- Viggle AI
- Task
- Video to video
- Modality
- Other
- Library
- minimax-h3
- Parameters
- 33.1B parameters
- Languages
- dmd
- Revision
- c09398730222a1eb9b7bf9e4141d439ebef91a68
- First published
- 2026-08-31
- Last updated
- 2026-10-17
Files and Weights
60 files, 69.0 GB in total. The weights are 16 files totalling 68.9 GB in pt, safetensors.
Weights16 files · 68.9 GB
Configuration3 files · 77.5 KB
Documentation5 files · 55.0 KB
Other35 files · 106.2 MB
Repository1 file · 3.9 KB
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| assets/fixed_embed_fwd_anyframe.pt | Weights | 3.7 MB | e0ae90929caf |
| lora/pytorch_lora_weights.safetensors | Weights | 2.7 GB | bcdd60965d95 |
| transformer/diffusion_pytorch_model-00001-of-00014.safetensors | Weights | 4.9 GB | fb30791125c5 |
| transformer/diffusion_pytorch_model-00002-of-00014.safetensors | Weights | 4.5 GB | 58d254c72eb9 |
| transformer/diffusion_pytorch_model-00003-of-00014.safetensors | Weights | 4.7 GB | bd8410138703 |
| transformer/diffusion_pytorch_model-00004-of-00014.safetensors | Weights | 4.9 GB | c42d981e9fa7 |
| transformer/diffusion_pytorch_model-00005-of-00014.safetensors | Weights | 4.6 GB | 2b6fb580aef3 |
| transformer/diffusion_pytorch_model-00006-of-00014.safetensors | Weights | 4.7 GB | a3183d396d2f |
| transformer/diffusion_pytorch_model-00007-of-00014.safetensors | Weights | 4.9 GB | 89fb79d36334 |
| transformer/diffusion_pytorch_model-00008-of-00014.safetensors | Weights | 4.6 GB | 9b2613f24f59 |
| transformer/diffusion_pytorch_model-00009-of-00014.safetensors | Weights | 4.7 GB | 82f6272df4fd |
| transformer/diffusion_pytorch_model-00010-of-00014.safetensors | Weights | 4.9 GB | a21d6103074b |
| transformer/diffusion_pytorch_model-00011-of-00014.safetensors | Weights | 4.6 GB | 2a2b57d4cccf |
| transformer/diffusion_pytorch_model-00012-of-00014.safetensors | Weights | 4.7 GB | 2cf2825b5658 |
| transformer/diffusion_pytorch_model-00013-of-00014.safetensors | Weights | 4.9 GB | 291fded8f271 |
| transformer/diffusion_pytorch_model-00014-of-00014.safetensors | Weights | 4.6 GB | 145d6c9dfc17 |
| inference/sample.py | Configuration | 12.4 KB | — |
| transformer/config.json | Configuration | 546 B | — |
| transformer/diffusion_pytorch_model.safetensors.index.json | Configuration | 64.5 KB | — |
| LICENSE | Documentation | 17.6 KB | — |
| LICENSE-CODE | Documentation | 11.4 KB | — |
| MODIFICATIONS.md | Documentation | 4.7 KB | — |
| NOTICE | Documentation | 515 B | — |
| README.md | Documentation | 20.9 KB | — |
| MiniMaxH3TokenRefinerBlock/package.pt2 | Other | 604.6 KB | 1fc10961935e |
| MiniMaxH3TransformerBlock/package.pt2 | Other | 853.1 KB | 58c651aa49a7 |
| assets/fixed_prompt.txt | Other | 1.5 KB | — |
| examples/demo.sh | Other | 1.8 KB | — |
| examples/media/audio-pin-driving.mp4 | Other | 4.2 MB | fc3dd85a5be2 |
| examples/media/audio-pin-out.mp4 | Other | 389.2 KB | 25fc5938c25a |
| examples/media/audio-pin-ref.png | Other | 442.1 KB | 6bf55661c6e0 |
| examples/media/before-after.png | Other | 1.3 MB | 958c7d84b72c |
| examples/media/community-tweet.png | Other | 360.3 KB | 84ce1d8e96c0 |
| examples/media/compare-cosplay.mp4 | Other | 1.1 MB | 6bd63ca750de |
| examples/media/compare-costume.mp4 | Other | 447.3 KB | 2a45ac3e161d |
| examples/media/compare-fastmotion.mp4 | Other | 1.2 MB | b79b19f959b3 |
| examples/media/compare-gamechar.mp4 | Other | 858.2 KB | 69e34a0ec75f |
| examples/media/compare-highkick.mp4 | Other | 847.8 KB | 6a137d9d1970 |
| examples/media/compare-prop.mp4 | Other | 678.1 KB | 4128c0e35bf1 |
| examples/media/compare-style.mp4 | Other | 536.8 KB | 692ff38363a0 |
| examples/media/hero-corgi.mp4 | Other | 426.0 KB | 4cb6c4bb9252 |
| examples/media/hero-duo.mp4 | Other | 187.0 KB | cbf81086b77e |
| examples/media/output.mp4 | Other | 880.4 KB | 7b89a33a902c |
| examples/media/pipeline.png | Other | 105.1 KB | 451cbcf83228 |
| examples/media/pipeline.svg | Other | 4.9 KB | — |
| examples/media/reference.png | Other | 541.5 KB | 5a01824f3023 |
| examples/media/swap-airliner.mp4 | Other | 397.3 KB | 5a7a31fd225e |
| examples/media/swap-anime-duo.mp4 | Other | 109.0 KB | 205700e55841 |
| examples/media/swap-claymation.mp4 | Other | 286.3 KB | dd08b83d8a57 |
| examples/media/swap-corgi.mp4 | Other | 272.4 KB | 5954aeab0360 |
| examples/media/swap-panda.mp4 | Other | 390.5 KB | bec174f093bf |
| examples/media/swap-penguin.mp4 | Other | 346.1 KB | ea2ccf8ce410 |
| examples/media/swap-photoreal-duo.mp4 | Other | 108.3 KB | a2d7c7fb36c5 |
| examples/media/swap-robot.mp4 | Other | 194.5 KB | e8147fb8f85e |
| examples/media/swap-wushu-animals.mp4 | Other | 204.0 KB | 1ed178d0c880 |
| examples/media/swap-wushu-robots.mp4 | Other | 201.5 KB | 63cc5dfe297c |
| examples/media/teaser.mp4 | Other | 14.4 MB | cbe5dd05057b |
| examples/media/white-house-funk.mp4 | Other | 73.2 MB | 810450b4afdb |
| requirements.txt | Other | 297 B | — |
| .gitattributes | Repository | 3.9 KB | — |
License and Download
- License
- other
- Access
- Open weights, no gate
- Download size
- 68.9 GB
Download from Viggle AI
Released by Viggle AI through its official repository on Hugging Face.
Built From
- Derived from MiniMaxAI/MiniMax-H3
Memory Requirements
| Precision | Weights in memory |
|---|---|
| As published | 68.9 GB |
| 16-bit | 66.2 GB |
| 8-bit | 33.1 GB |
| 4-bit | 16.6 GB |
Weights only, from the published parameter count; the key-value cache and runtime add to this.
Questions About Viggle-Animate
How much GPU memory does Viggle-Animate need?
About 79.5 GB at 16-bit and 19.9 GB at 4-bit: the weights (33.1B parameters) plus a working margin. A long context needs more.
What is the cheapest GPU to run Viggle-Animate on?
At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.
What license is Viggle-Animate released under?
other, as its publisher declares it. Read the license text before commercial use.
Similar Models
The six-H200 full-corpus run is in progress. Its rolling latest checkpoint is in a mutable Storage Bucket, with the verified step recorded in latest.json. The final weights will be published in this model repository. Live training and held-out validation metrics are updated throughout the run. The model initializes from OpenWAM-α at revision 52df4e66c82c5c8b480adcc8d01f4db7415dfb56. Its Slither adapter replaces the 80-dimensional robot action head with three outputs: cosine and sine of estimated mouse angle, and binary boost. It removes robot proprioception and uses fixed empty-prompt conditioning. Each example uses 33 frames at 30 Hz and 32 transition actions; the nine video frames are…
Scene-specialized causal video renderers for Medieval Castle, Ancient Egypt, and City Racing. An executable game determines gameplay, physics and input, while the model re-renders clean low-poly RGB. HUD and sound remain separate, frame-aligned presentation layers. Each scene checkpoint has 5,001,016,512 BF16 parameters. The weights require the Matrix-Game 4.0 inference source; they are not drop-in Diffusers pipelines. Training code and training data are not included. Shared encoders/codecs are installed once for all scenes. See game asset details. Use the source release's manifest downloader with a pinned Hub commit revision and --repo-id Skywork/Matrix-Game-4.0 --repo-type model for both…