bsvogler/music-genre-recognition-pipeline

A tool to detect the music genre using machine learning with keras.

🎓 MSc Robotics, Cognition, Intelligence @ TU Munich

🎓 BSc Media Informatics @ Bauhaus University Weimar

My interest is in extending capabilities of reason with computers. My academic interest is Spiking Neural Networks. I did some game development in my early years.

More about my thoughts on my blog.

music-genre-recognition-pipeline's Issues

Model improvements

• The first column in the data contains a timestamp in ~11 ms increments. It could therefore be removed.
• GRU Cells could replace the LSTM cells.

Reevaluate 30s split

There might be no need to split files in 30s chunks. This was introduced because of the high memory footprint of Sonic Annotator when extracting.

Dockerize feature extraction

Installing the dependency might not work. A working dockerized image should be provided.

Server issue

Exception happened during processing of request from ('127.0.0.1', 47328)
Traceback (most recent call last):
  File "/usr/lib/python3.8/http/server.py", line 1179, in run_cgi
    os.execve(scriptfile, args, env)
FileNotFoundError: [Errno 2] No such file or directory: '/home/bsvogler/Dokumente/music-genre-recognition-pipeline/cgi-bin/genreclassifierWebAPI.py'

Check if can write files before calculating