The VAE used by both Ming-Image-0.1-Design checkpoints, hosted separately so any Ming Image model - ours or a community finetune - can point at it instead of shipping its own copy. One file, BF16, about 242 MB.
Originally released by inclusionAI on Hugging Face under the MIT license, and repackaged as a single file for ComfyUI by Comfy-Org. All credit for the weights goes to the inclusionAI team.
One VAE for the whole series
Unlike the text encoders, which differ between the base and layer-decomposition checkpoints, the VAE is shared. The same file serves Ming Image Design and Ming Image Design Layer, and it is unchanged from the upstream release.
Where it goes
In ComfyUI, models/vae/. This is a component rather than a runnable model, so there is no Generate button and no on-site generation - if you want to generate with Ming, that is the Ming Image Design page, where the VAE is already handled for you. You only need this file if you are running Ming locally, or running a finetune whose author did not bundle it.
Links
- Source: inclusionAI/Ming-Image-0.1-Design
- Repackaged file: Comfy-Org/Ming-Image
- Code: github.com/inclusionAI/Ming-Image
- Text encoders: Ming Image Text Encoders