Generating Music with Neural Networks using TensorFlow
Submitted by Shauryadeep Chaudhuri (@xuryax) on Friday, 1 September 2017
Technical level: Intermediate
The talk will focus on Introduction to basic concepts of neural networks like CNN & RNN. Create a simple Neural Network which will train itself on MIDI samples of a particular artist. After training the neural network would generate a midi music file of 30 to 60 seconds.I would show two music samples from two different artist. The Music file would be generated and the neural network would be trained beforehand due to time constraints. Would make the code available and walkthrough the code.
Breaking Up the talk.
1. Introduction to Neural Networks, CNN , RNN (5 Mins)
2. Introduction to tools used like TensorFlow,numpy,pandas (5 Mins)
3. Doing code walkthrough and explaining the code part linking with theoretical part.(10 mins)
4. Showing different music samples generated by training data from different artist. (5 Mins)
5. Q & A Session of (5 Mins)
A basic idea about Machine Learning
I am Shauryadeep Chaudhuri, Worked as an engineer in Persistent Systems for Data Intensive Applications. Now work as an Engineer in Xoxzo Inc. which deals with SMS & Telephony APIs, and each employee works remotely. Took part in PyCon APAC 2017 held in Kuala Lumpur.
An Avid traveller likes to travel as much as possible.