Paper Title
Music Genre Classification/Recognition

Abstract
In this paper we talk about the architecture of Neural networks, mostly of which includes CNN (convolutional neural network) and CRNN ( convolutional recurrent neural network) which we are planning to implement in our system for a custom music genre classification with our own data-set and GTZAN data-sets. Here we try to develop our system in the case of low computational level and data budget by which we will be able to train with a much larger data-sets. We use multiple strategies for tuning , initializing and optimizing. We also use a multiframe method by which we can almost analyze a n entire song or audio in detail. Basically which is used during training time for producing more samples it is also used during testing time to get an entire an overall summary of the entire song or audio. We finally at the end evaluate the results with both our handmade data-sets and GTZAN data-sets which are used in a lot of work. Keywords - Music Genre Classification, Neural Networks, Multiframes.