whatmodel.sh / video / Stable Video Diffusion (SVD, 14 frames)

Stable Video Diffusion (SVD, 14 frames)

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Stable Video Diffusion (SVD, 14 frames) is an open-weights video model from Stability AI, released November 2023. Its published size is 1.52B parameters. Licence: stable-video-diffusion-community.

Catalogue identity notes

Image-to-video model requiring a starting image; not text-to-video. Original release, distinct from SVD XT 1.1. Place a starting image at input.png. CUDA GPU plus system RAM required for CPU offload. 8GB VRAM is a planning estimate based on the documented chunking/offload setup; not a local measurement.

Model information sources

· Parameter count ↗

Video UNet only: HF Hub safetensors total for the official repo (pinned revision), 1,524,623,082 = unet/diffusion_pytorch_model.safetensors. The CLIP image encoder (632.1M) and the VAE (97.7M) are not included. Observed 2026-10-07.

No comparable score yet. No reviewed benchmark results yet.

$ how it compares

Scored video models released closest to Stable Video Diffusion (SVD, 14 frames), out of 139 with a score.

#modelscorewhereRAMreleased
118VideoCrafter2 Tencent AI Lab34±2local~9.4GB2024-01
120Pika 1.0 Pika32±3comparison only2023-11
127Hotshot-XL Hotshot27±2local~10GB2023-10
133VideoCrafter1 T2V Tencent AI Lab / HKUST26±2local~8.1GB2023-10
134Show-1 Show Lab / NUS25±2local~83GB2023-10

see all video models · find the ones that fit your machine

$ where to run it

Other ways to run it

  1. Run it as published · needs ~8 GB, NVIDIA only · from huggingface.co/docs
    uv run --with torch --with diffusers --with transformers --with accelerate --with imageio-ffmpeg python -c "import torch; from diffusers import StableVideoDiffusionPipeline; from diffusers.utils import load_image, export_to_video; p=StableVideoDiffusionPipeline.from_pretrained('stabilityai/stable-video-diffusion-img2vid', torch_dtype=torch.float16, variant='fp16'); p.enable_model_cpu_offload(); p.unet.enable_forward_chunking(); image=load_image('input.png').resize((1024,576)); frames=p(image, num_frames=14, decode_chunk_size=2).frames[0]; export_to_video(frames, 'video.mp4', fps=7)"

Official weights and documentation: huggingface.co/stabilityai ↗

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json