VLMInferSlow: Evaluating the Efficiency Robustness of Large Vision-Language Models as a Service

Open in new window