ComfyUI Node
SeeThrough Generate Layers (Custom)
A ComfyUI node in SeeThrough with 8 inputs and 2 outputs.
SeeThrough Generate Layers (Custom)
- image
- layerdiff_model
- layers
- preview
◄seed42►
◄resolution1280►
◄num_inference_steps30►
◄enable_head_detailtrue►
◄auto_fillfalse►
◄min_alpha_coverage0.010►
CategorySeeThrough
Inputs (8)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | — | |
| layerdiff_model | SEETHROUGH_LAYERDIFF_MODEL | — | |
| seed | INT | 420–4294967295 | — |
| resolution | INT | 1280512–2048 | — |
| num_inference_steps | INT | 301–100 | — |
| enable_head_detail | BOOLEAN | true | v3 only: enable head detail stage (face, eyes, ears, etc). Disabling skips the 2nd inference pass and saves ~50% time. |
| auto_fill | BOOLEAN | false | Auto-fill missing layers: if enabled, automatically re-runs inference (up to 5 times) until all expected layers are generated. Expected: v3+head=24, v3 body=13, v2=19. |
| min_alpha_coverage | FLOAT | 0.0100.001–0.1 | Minimum alpha coverage ratio to consider a layer valid. Layers below this threshold are treated as missing. Only used when auto_fill is enabled. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| layers | SEETHROUGH_LAYERS | — |
| preview | IMAGE | — |