LoadersPartial Capture

Load LoRA

class_type: LoraLoader

Source repo

Applies a LoRA (Low-Rank Adaptation) weight file to an existing model and CLIP encoder. LoRAs modify the base model behavior—adding styles, concepts, or characters—without replacing the full checkpoint.

What It Does

LoraLoader takes a MODEL and CLIP from a checkpoint loader and applies a LoRA file with configurable strength for both the model and CLIP components independently. Multiple LoRA loaders can be chained to combine effects.

LoRA provenance is critical for reproducibility: the same prompt with different LoRA combinations produces dramatically different results. Each LoRA in the chain contributes to the final output, and the strength values determine the intensity of each effect.

ComfyUI records the LoRA filename and strength values in the workflow JSON, making LoRA usage fully traceable through Numonic extraction.

Inputs

modelMODEL

Base model to modify.

clipCLIP

CLIP encoder to modify.

lora_nameSTRING

LoRA filename from the models/loras directory.

strength_modelFLOAT

LoRA strength applied to the model (0.0–2.0).

strength_clipFLOAT

LoRA strength applied to CLIP (0.0–2.0).

Outputs

MODELMODEL

Modified model with LoRA applied.

CLIPCLIP

Modified CLIP with LoRA applied.

What Numonic Captures

  • LoRA filename
  • Model strength value
  • CLIP strength value
  • Position in LoRA chain (when multiple LoRAs used)

Known Gaps

  • LoRA file hash — filename only
  • LoRA training details (base model, trigger words)
  • Effective combined strength when multiple LoRAs interact

Related Nodes

Capture ComfyUI metadata automatically

Numonic extracts workflow metadata from every ComfyUI generation — models, samplers, seeds, prompts, and custom nodes. Track provenance, maintain compliance, and never lose a workflow.