Inspiration
For this project, the main thing that I was inspired by was this video: https://www.youtube.com/watch?v=x_rLw6SCSmE
that was released about a year ago of people singing music. Personally, I don't know music that well, however, I think I know some things about computers. This is why I personally decided to try and create something fun and difficult while also something that can be used in the real world
What it does
In theory, if the machine learning worked, the code is supposed to generate a new piece of music based on already existing pieces.
Challenges we ran into
The biggest challenge that I ran into was that the dependencies were not as robust as I previously thought. Something in the first dependency I used was causing an error with TensorFlow compatibility and not allowing the classes to import/run. I am still working on how to solve this.
Accomplishments that we're proud of
The biggest thing I am proud of with this project is that I came in not knowing a single thing about generating music with Machine Learning and even if it did not work in the end, I was still able to put code on an editor and test it out.
What we learned
During this Hackathon, personally, I learned a lot about encoder/decoder machine learning. The biggest hurdle that I faced during this hackathon was definitely writing the code itself, however, I was able to do a lot of research and get a lot of good example code and projects for inspiration.
What's next for Music generation
If I had more time I would really have liked to fix me errors in the dependencies I used. This is the main problem with my code that I was unable to fix. Alongside that, I generally think I need to work more with TensorFlow as I am not very fluent with it yet. Overall, I hope to work on this project more in the future to try and get a robust code up and running.
(Not all of the code in the github is mine, most of them are dependencies Also, got a lot of help from https://gist.github.com/llSourcell https://replit.com/join/lhabdudmaw-maryan24 - website link )
Log in or sign up for Devpost to join the conversation.