Kandinsky LabKandinsky control models
Kandinsky 3 IP-Adapter
The 3.1 release includes an IP-Adapter trained to use an image as a generation condition.
Research release
Add image guidance to the model
The adapter is distributed separately from the image generator. The developer links its Kandinsky Diffusers implementation for use.Kandinsky 3: base, Flash and inpainting releases
Keep image prompting distinct from masked editing
Practical guidance. Use an image condition to guide the generated result. If only one region should change, choose a workflow that also provides an explicit edit mask.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | ai-forever/kandinsky3_ip_adapter |
| Card license | Apache 2.0ai-forever/kandinsky3_ip_adapter — model card and files |
Related models
- Kandinsky 3.0
Kandinsky Lab · Kandinsky 3
An image diffusion model with a larger text encoder and U-Net, trained with attention to Russian cultural subjects.
Launch date not confirmedDownloadable