Jul 25, 2017 · 1 min read
thanks for your question akshay, you’re asking 2 questions here.
First, the bag-of-words size: this won’t increase exponentially, recall it is a set of unique words throughout the corpus. Second, the classification function is a series of mathematical functions on tensors, the size of the corpus has no bearing on its speed. If you create a larger # of neurons and weights than this would impact performance.
