Nodes/ComfyUI-Krea2T-Enhancer/Krea2T Character LoRA — Image Only
ComfyUI Node

Krea2T Character LoRA — Image Only

Applies a Krea2 character LoRA directly to image tokens and skips unsupported weights. Use in place of the regular loader for that character. Other LoRAs remain active. Text and image still interact through the model's normal attention.

By capitan01R·Created 3 months ago·Updated about 8 hours ago· 236
Krea2T Character LoRA — Image Only
  • model
  • model
  • report
◄lora_name▾►
◄strength_model1.00►
◄enabledtrue►
CategoryKrea2/loaders

Inputs (4)

NameTypeDefaultDescription
modelMODEL—
lora_nameCOMBOCharacter LoRA. Loads canonical A/B pairs for Krea2's shared DiT attention and MLP projections; other keys are skipped.
strength_modelFLOAT1.00-100–100Strength of this character adapter on image-token rows. Text-token rows receive no direct contribution from this adapter.
enabledBOOLEANtrue—

Outputs (2)

NameTypeDescription
modelMODEL—
reportSTRING—