Expected System Specifications and Hosting Qwen2-VL-7B-Instruct on EC2

[shrinesabu2000]

I am planning to host the Qwen2-VL-7B-Instruct model as a server on an EC2 instance. I would like to know the recommended system specifications for running this model efficiently. Additionally, I am exploring whether using vLLMs is the best approach for deploying this model in a production environment. I plan to use this model primarily for OCR tasks and multimodal inference, processing both images and text in real-time.

Also, would you recommend vLLMs for deploying Qwen2-VL-7B-Instruct as a server for inference?