Kandinsky LabKandinsky 2
Kandinsky 2.1 Image Prior
The separately trained diffusion prior supplies image embeddings for the 2.1 generation pipeline.
Pair the prior with its image decoder
The prior connects CLIP's text and image representation spaces. It does not directly output image pixels.ai-forever/Kandinsky_2.1 — model card and files
Specifications & access
| Detail | Value and conditions |
|---|---|
| Access | Downloadable |
| Model identifiers | kandinsky-community/kandinsky-2-1-priorai-forever/Kandinsky_2.1/prior_fp16.ckpt |
| Prior size | 1B parametersai-forever/Kandinsky_2.1 — model card and files |
| Weight license | Apache 2.0ai-forever/Kandinsky_2.1 — model card and files |
Related models
- Kandinsky 2.1
Kandinsky Lab · Kandinsky 2
An image model that combines an image prior and latent decoder with mixed text and image conditions.
Launch date not confirmedDownloadable