VLMEvalKit

Overview

VLMEvalKit is an open-source evaluation toolkit designed for large multi-modality models, supporting over 220 models and 80 benchmarks. It is ideal for researchers and developers looking to assess the performance of various AI models in a comprehensive manner.