Extension of the NIM to handle convolution of subunit outputs with learned kernels.
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).