121 karma · joined October 18, 2011
Dont have the code up for sharing but I documented xeil (along with a few other tools) in my blog: https://paul.mou.dev/posts/2026-04-28-software-for-one/#xeil
MusicCaps [1] is a dataset containing pairs of music audio and natural language description of the clip; the reason why the result is good imo is because the trained model was able to generate a description with features of the ground truth
Love to hear what you think of it!
Sloth is a modern day copy and paste manager that allows you to search snippets by both content and context. Context here being what program you copied from and the type of text (whehter it's an ip, python, or bash script, etc.).
Right now I am working on a custom machine learning model to do better text classification.
It's built on electron and I am adding slack integration to create a shared clipboard. Essentially, slack would act as a cloud storage.
Ohsloth is an open source copy and paste manager that aims to provide contextual actions based on snippet content.
It is still super early but it definitely could use some design love!
For image processing (CNN) https://cs231n.github.io/
For natural language processing (RNN) http://cs224d.stanford.edu/syllabus.html
I also found the following coursera helpful https://www.coursera.org/course/neuralnets https://www.coursera.org/learn/machine-learning
I researched geodb a while back and I only found PostGIS but I recently found out that elastic search also supports geo features.
Shameless plug - This is the simple API I built on top of PostGIS https://github.com/moo-mou/GeoGo
You should see an output.png if it run successfully in the same folder.
Sorry, no easy access to a windows right now, but I will try it out on when I get a chance.
In order to learn more (I admittedly know very little about anything other than the phong model - the other friend is much more knowledgable), I presume either a textbook or a course notes like this https://www.student.cs.uwaterloo.ca/~cs488/notes.pdf will probably help.
My gut feeling says it will work but I am curious if the model need to be trained separately for those.
Another idea I was thinking about is altering the music the programmer is playing in some way instead of generating music.
I deliberately chose to release as early as possible to find out if what people think of the project since I tend to have a tendency of taking forever to get feedback.
But in theory, any soundfont file will do. There is a bunch of sound fonts here:
http://soundfonts.homemusician.net/piano_soundfonts/page1/in...
Please let me know what you guys think!
It should not be too much work to add in external speech recognition library later on.
But I also believe that solving personal problem is a good place to start. After all, if one person has a problem, it is likely that there are at least 5 other person who shares the pain.