Hugging Face's logo

hpcai-tech
/
OpenSora-VAE-v1.2

Transformers
Safetensors
VideoAutoencoderPipeline
Inference Endpoints
Model card Files Community

Open-Sora VAE-v1.2 Weights

This repository stores the weights of the VAE released by the Open-Sora team. You can visit our project at:

The weights are released together with Open-Sora v1.2.

We recommend you to use this weights in the Open-Sora codebase. If you want to use VAE in your own project, you may use the following sample code.

  1. Install opensora
pip install git+https://github.com/hpcaitech/Open-Sora.git
  1. Use STDiT3 in your own code
from opensora.models.vae.vae import VideoAutoencoderPipeline
vae = VideoAutoencoderPipeline.from_pretrained("hpcai-tech/OpenSora-VAE-v1.2")
Downloads last month
33,741
Safetensors
Model size
393M params
Tensor type
F32
·
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Spaces using hpcai-tech/OpenSora-VAE-v1.2 10

Collection including hpcai-tech/OpenSora-VAE-v1.2