I spend some time thinking about what could be the issue we were running into, accidently fix it by messing around with the settings and then randomly end up having a working model that I can comfortably release

The one where I dissect the results of the different models i’ve trained and the big problem surrounding sound synthesis that I am running into and come to a conclusion.

The one where we return from out haitus coffin and try to see what we have left to do and come to a compromise whilst discovering a major issue in our current model!

We have sound generated from our checkpoints, time to analyze them and think about our next steps!

Small update about how im moving the environment from the tiny powerless computer I have at home, to the more beefy powerful computers over at google.

Tickle.py dev diary #2, in this post we dive into the world of audio chunking scripts!

An introduction to Tickle.py, where I give a roadmap of what I want to do and how I intend to achieve it at this point in time!