ComfyUI Node
LC Anima LoRA Editor
Loads an Anima LoRA with per-block, attention / MLP and spike edits, applies it to the model and outputs the edited LoRA for LC LoRA Save. The heatmap shows each layer's strength before and after the edits.
LC Anima LoRA Editor
- model
- model
- edited_lora
◄lora_name▾►
◄strength1.00►
◄presetFull►
◄blocks1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1►
◄attention1.00►
◄mlp1.00►
◄outside_blocks1.00►
◄spike_softenoff►
◄spike_limit2.0►
◄blocks_in—►
CategoryLC ModelBuilder/Anima
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| model | MODEL | Anima model. The edited LoRA is applied to it, like a LoRA loader. | |
| lora_name | COMBO | LoRA to edit, from models/loras. | |
| strength | FLOAT | 1.00-4–4 | Overall LoRA strength, same as a LoRA loader. Baked into the saved LoRA. |
| preset | COMBO | Full | Full = the LoRA as it is. Each block group preset turns those blocks down to 0.3. Try each on a fixed seed batch and keep the one that frees the face. Dragging the graph switches to Custom. |
| blocks | STRING | 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 | 28 block multipliers, 0 to 2 (1 = unchanged). Drag the bars on the node. |
| attention | FLOAT | 1.000–2 | Multiplier for the attention layers (self and cross attention) in every block. |
| mlp | FLOAT | 1.000–2 | Multiplier for the MLP layers in every block. |
| outside_blocks | FLOAT | 1.000–2 | Multiplier for layers outside the 28 blocks (LLM adapter, final layer), if the LoRA has any. The heatmap note says how many. |
| spike_soften | COMBO | off | layers = pull any layer that is far stronger than the same layer in other blocks down to the limit. directions = inside each layer, flatten the one or two directions that dominate it (plain lora_A / lora_B files only). both = both. |
| spike_limit | FLOAT | 2.01–10 | How strong a spike may be before it is softened. layers: times the median of that layer across the 28 blocks. directions: times the layer's second strongest direction (well-behaved LoRAs sit around 2). Lower = more softening. |
| blocks_inopt | STRING | A 28-block curve from another node, such as LC Anima LoRA Block Scan's blocks output. When wired it replaces the preset and the bars. |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| model | MODEL | The model with the edited LoRA applied. |
| edited_lora | LC_LORA | The edited LoRA, for LC LoRA Save. |