hello!
i have a 2080ti GPU in my desktop and has a parameter and graph(json format) about Lenet5. when i bring them like below code.
json_file = open("Lenet5_model.json","r")
loaded_model_json = json_file.read()
json_file.close()
Network = keras.models.model_from_json( loaded_model_json )
Network.load_weights("Weight.h5")
the gpu memory usage show like below
As you can see in the red box in the picture, the memory used by the GPU is high.
i can’t understand why they are pre-allocate gpu memory like that. Lenet5 model has really small paramter and simple network. why they pre allocate memory like that??