Using tensorflow grpc causes memory leaks when calling to server

[YHL04]

I took out the code for tensorflow grpc and for some reason the memory keeps increasing when the client makes a call to the server, the function takes in no parameters and returns tf.random.uniform tensors but the memory still accumulates over thousands of calls

[YHL04]

https://github.com/YHL04/tensorflow_grpc

simplest code to reproduce my issue can be found here (prints memory in MB)