Initially, I have a neural network that take a 1d-tensor of 2500 elements, I run this network 20 times in a for loop to get the 20 output 1d-tensors. Now I want to run this network only one time but with a Batch size of 20. My question is
- Do I need to change anything in my Neural network to take the account of the changes that I have done?
- I tried to get the results without making any changes in the neural network, it is still giving me results but there is one problem, all 20 of them have same weights and biases, but I want them to be independent from each other.
Please help me to clarify my doubts.