GRAN-TED: Generating Robust, Aligned, and Nuanced Text Embedding for Diffusion Models

TL;DR

GRAN-TED introduces a new benchmark and training paradigm for creating robust, aligned, and nuanced text embeddings that significantly improve the performance and efficiency of diffusion models in text-to-image and text-to-video generation.

Contribution

The paper presents TED-6K, a fast and reliable evaluation benchmark, and a novel two-stage training method for superior text encoders tailored for diffusion models.

Findings

TED-6K correlates strongly with downstream task performance.

Evaluating with TED-6K is approximately 750 times faster than end-to-end training.

GRAN-TED achieves state-of-the-art results on TED-6K and improves generation quality.

Abstract

The text encoder is a critical component of text-to-image and text-to-video diffusion models, fundamentally determining the semantic fidelity of the generated content. However, its development has been hindered by two major challenges: the lack of an efficient evaluation framework that reliably predicts downstream generation performance, and the difficulty of effectively adapting pretrained language models for visual synthesis. To address these issues, we introduce GRAN-TED, a paradigm to Generate Robust, Aligned, and Nuanced Text Embeddings for Diffusion models. Our contribution is twofold. First, we propose TED-6K, a novel text-only benchmark that enables efficient and robust assessment of an encoder's representational quality without requiring costly end-to-end model training. We demonstrate that performance on TED-6K, standardized via a lightweight, unified adapter, strongly correlates with an encoder's effectiveness in downstream generation tasks. Notably, under our experimental setup, compared with training a diffusion model from scratch, evaluating with TED-6K is about \textbf{750$\times$ faster}. Second, guided by this validated framework, we develop a superior text encoder using a novel two-stage training paradigm. This process involves an initial fine-tuning stage on a Multimodal Large Language Model for better visual representation, followed by a layer-wise weighting method to extract more nuanced and potent text features. Our experiments show that the resulting GRAN-TED encoder not only achieves state-of-the-art performance on TED-6K but also leads to demonstrable performance gains in text-to-image and text-to-video generation. Our TED-6K dataset and evaluation code are available at the following link: https://anonymous.4open.science/r/GRAN-TED-4FCC/.