ComfyUI Node
LC Krea2 Semantic Vector Tuner
Amplify or attenuate specific regions of a single Krea2 checkpoint's weights. Not a merge. Reuses Arthemy SDXL Suite's own tuning mechanism (bake existing patches, scale, re-patch the delta via ModelPatcher.add_patches) pointed at Krea2's real block/region prefixes instead of SDXL's UNet layout. 3 zone sliders cover the 28 uniform blocks (with an optional per-block override string for finer control); the 10 named, architecturally distinct regions each get their own slider.
LC Krea2 Semantic Vector Tuner
- model
- model
- info
◄modeSoft Value►
◄base_scale1.00►
◄low_blocks1.00►
◄mid_blocks1.00►
◄high_blocks1.00►
◄blocks_override►
◄first1.00►
◄tmlp1.00►
◄txtmlp1.00►
◄tproj1.00►
◄txtfusion_layer01.00►
◄txtfusion_layer11.00►
◄txtfusion_projector1.00►
◄txtfusion_refiner01.00►
◄txtfusion_refiner11.00►
◄last1.00►
CategoryLC ModelBuilder/Krea2
Inputs (17)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | Single checkpoint to amplify/attenuate by region. This is not a merge — no second model. | |
| mode | COMBO | Soft Value | Soft Value dampens the sliders so 0-2 stays safe (values above 1.0 land around 1.0-1.13x). Real Value applies each slider as a literal multiplier. |
| base_scale | FLOAT | 1.000–2 | Fallback multiplier for any weight that isn't one of the 28 blocks or the 10 named regions. |
| low_blocks | FLOAT | 1.000–2 | Scales blocks.0. - blocks.8. (low zone). |
| mid_blocks | FLOAT | 1.000–2 | Scales blocks.9. - blocks.17. (mid zone). |
| high_blocks | FLOAT | 1.000–2 | Scales blocks.18. - blocks.27. (high zone). |
| blocks_override | STRING | Optional comma list of exactly 28 multipliers, one per block. Overrides low_blocks/mid_blocks/high_blocks for the 28 blocks when present. Named regions below are unaffected — they always come from their own sliders. | |
| first | FLOAT | 1.000–2 | first. — earliest input/embedding weights, before block 0. |
| tmlp | FLOAT | 1.000–2 | tmlp. — timestep-embedding MLP (timestep conditioning). |
| txtmlp | FLOAT | 1.000–2 | txtmlp. — text-embedding MLP (text conditioning). |
| tproj | FLOAT | 1.000–2 | tproj. — timestep projection layer. |
| txtfusion_layer0 | FLOAT | 1.000–2 | txtfusion.layerwise_blocks.0. — first text/image fusion block. |
| txtfusion_layer1 | FLOAT | 1.000–2 | txtfusion.layerwise_blocks.1. — second text/image fusion block. |
| txtfusion_projector | FLOAT | 1.000–2 | txtfusion.projector. — projects fused text features into the main stream. |
| txtfusion_refiner0 | FLOAT | 1.000–2 | txtfusion.refiner_blocks.0. — first fusion refiner block. |
| txtfusion_refiner1 | FLOAT | 1.000–2 | txtfusion.refiner_blocks.1. — second fusion refiner block. |
| last | FLOAT | 1.000–2 | last. — final output layer, after block 27. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| model | MODEL | — |
| info | STRING | — |