ELLAAdapter<T>

Models & Types · AiDotNet.Diffusion.Guidance

ELLA (Efficient Large Language Model Adapter) guidance adapter for enhanced text understanding.

Implements: IGuidanceMethod<T>

For Beginners

ELLA makes the AI better at understanding complex prompts. Instead of relying only on CLIP’s text understanding, it leverages a larger language model to capture nuances like spatial relationships and attributes.

How It Works

ELLA bridges large language models with diffusion models by providing richer text embeddings through an adapter network. It enhances prompt understanding for complex, compositional descriptions without retraining the diffusion model.

Reference: Hu et al., “ELLA: Equip Diffusion Models with LLM for Enhanced Semantic Alignment”, 2024

Constructors

Constructor Summary
ELLAAdapter(double adapterWeight) Initializes a new ELLA adapter.

Properties

Property Summary
GuidanceType GuidanceType { get; }

Methods

Method Summary
Tensor<T> Apply(Tensor<T> unconditional, Tensor<T> conditional, double scale, double timestep)