Computer Vision-aided Efficient 7B size Large Language and Vision Models. Let's enjoy it