Lora Loader 🐍

Documentation

Class name: LoraLoader|pysssss
Category: loaders
Output node: False

This node specializes in loading and applying LoRA (Low-Rank Adaptation) adjustments to models and clips, enhancing their capabilities or altering their behavior based on specified LoRA configurations. It extends the functionality of a base loader to also handle image-specific LoRA configurations, making it versatile for various multimedia applications.

Input types

Required

model
- The model to which the LoRA adjustments will be applied. It's crucial for defining the base architecture that will be enhanced or modified through LoRA.
- Comfy dtype: MODEL
- Python dtype: torch.nn.Module
clip
- The CLIP model to which the LoRA adjustments will be applied, allowing for enhanced or altered multimodal understanding and representation.
- Comfy dtype: CLIP
- Python dtype: torch.nn.Module
lora_name
- Specifies the name of the LoRA configuration to be applied, determining the specific adjustments and enhancements to the model and clip.
- Comfy dtype: COMBO[STRING]
- Python dtype: str
strength_model
- Defines the intensity of the LoRA adjustments applied to the model, allowing for fine-tuned control over the modifications.
- Comfy dtype: FLOAT
- Python dtype: float
strength_clip
- Defines the intensity of the LoRA adjustments applied to the CLIP model, enabling precise control over the enhancements.
- Comfy dtype: FLOAT
- Python dtype: float

Output types

model
- Comfy dtype: MODEL
- The model with applied LoRA adjustments, reflecting enhanced or altered capabilities.
- Python dtype: torch.nn.Module
clip
- Comfy dtype: CLIP
- The CLIP model with applied LoRA adjustments, showcasing enhanced or altered multimodal understanding and representation.
- Python dtype: torch.nn.Module

Usage tips

Infra type: GPU
Common nodes:
- LoraLoader|pysssss
- ModelSamplingDiscrete
- Reroute
- CLIPTextEncode
- IPAdapterApply
- [Anything Everywhere](../../cg-use-everywhere/Nodes/Anything Everywhere.md)

Source code

class LoraLoaderWithImages(LoraLoader):
    @classmethod
    def INPUT_TYPES(s):
        types = super().INPUT_TYPES()
        names = types["required"]["lora_name"][0]
        populate_items(names, "loras")
        return types

    def load_lora(self, **kwargs):
        kwargs["lora_name"] = kwargs["lora_name"]["content"]
        return super().load_lora(**kwargs)

2.7 KiB Raw Permalink Blame History