EVALALIGN: Supervised Fine-Tuning Multimodal LLMs with Human-Aligned Data for Evaluating Text-to-Image Models
This addresses the problem of evaluating text-to-image models for researchers and developers, but it is incremental as it builds on existing MLLM capabilities.
The authors tackled the lack of accurate evaluation metrics for text-to-image models by proposing EvalAlign, a metric that uses supervised fine-tuning of multimodal LLMs on human-aligned data, resulting in superior stability and closer alignment with human preferences across 24 models.
The recent advancements in text-to-image generative models have been remarkable. Yet, the field suffers from a lack of evaluation metrics that accurately reflect the performance of these models, particularly lacking fine-grained metrics that can guide the optimization of the models. In this paper, we propose EvalAlign, a metric characterized by its accuracy, stability, and fine granularity. Our approach leverages the capabilities of Multimodal Large Language Models (MLLMs) pre-trained on extensive data. We develop evaluation protocols that focus on two key dimensions: image faithfulness and text-image alignment. Each protocol comprises a set of detailed, fine-grained instructions linked to specific scoring options, enabling precise manual scoring of the generated images. We supervised fine-tune (SFT) the MLLM to align with human evaluative judgments, resulting in a robust evaluation model. Our evaluation across 24 text-to-image generation models demonstrate that EvalAlign not only provides superior metric stability but also aligns more closely with human preferences than existing metrics, confirming its effectiveness and utility in model assessment.