ComfyUI Node
MOSS i2i Reference
Feed the page to MOSS-Image as a VAE reference latent (RoPE r = 1).
MOSS i2i Reference
- model
- vae
- image
- MODEL
- LATENT
◄resolution1024►
◄ref_scale2►
CategoryMOSS-Image/i2i
Inputs (5)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | — | |
| vae | VAE | — | |
| image | IMAGE | — | |
| resolution | INT | 1024256–2048 | Target area ≈ resolution², native aspect, sides multiples of 16 (the training cache's target_size). Keep it equal on the Text Encode and Reference nodes so the vision tower and the ref see the same page. |
| ref_scale | INT | 21–8 | Ref side = target side / ref_scale. The colorize ref LoRA was trained at 2 (1/4 area). |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| MODEL | MODEL | — |
| LATENT | LATENT | — |