Assumption: If we **replace RNNs with max-pooling**, our network(NOT trainable) will be **same as SiMPLe**. Let's examine this!
Assumption:
If we replace RNNs with max-pooling, our network(NOT trainable) will be same as SiMPLe.
Let's examine this!