Measure AWS SageMaker latency for object detection

Please only apply to this project if you have experience deploying image object detection models in AWS SageMaker.

Deploy a Yolov7 Torch model in SageMaker.

Make sure to use a scalable deployment.
Only one instance is needed, but it must use a load balancer and anything else required for scale.
Demonstrate a python based client that sends images to that model.
It should be able to send one or a few jpeg images of any size.
Measure the time at the client it takes from the time the image is sent until the prediction result is seen back at the client.
Repeat the experiment a 100 times.
List the latency for each trial.

You need to deliver these:
Precise instructions on how to configure the SageMaker to deploy the model.
Python code for the client.
Latency measurements as described above

If you are unable to find a yolov7 model I can provide one.