Neural Networking

AI

I worked through another Python AI video. It was only about 12 minutes. It took me more like 4 hours. I had to download a 1.2 GB dataset. This time it involved music. The data set contained 10 different genres with 100, 30 second songs. I also had to download and install some additional software, librosa which is an audio/music analyzer, Tensorflow an open source machine learning framework and Kersa a Python Deep Learning Library.  I had to change up the code a little to save the analysis to files. It took several minutes to process the files. Since I was typing in the examples, I knew there would be typos.  I used something called pickle. Pickle can take and object and output it to a file. Once I got the code working it was interesting. The video did not really get in depth about how a neural network works. I did think the librosa output to matplot looked interesting. Here are some examples.

Matplot of Librosa analysis of Disco
Matplot of Librosa analysis of some classical music.
Matplot of Librosa analysis of hiphop
Matplot of Librosa analysis of the intro to I Want you Back by the Jackson 5.

 

Worklight Update

Here is a photo of the worklight I built last week for Angel.

Finished Work light setup and working

 

Art

Continued with the next lesson on Synfig

Nuntio

watched Great Courses- Writing Great Fiction.