-
stylegan2-pytorch
Simplest working implementation of Stylegan2, state of the art generative adversarial network, in Pytorch. Enabling everyone to experience disentanglement
-
InfluxDB
Power Real-Time Data Analytics at Scale. Get real-time insights from all types of time series data with InfluxDB. Ingest, query, and analyze billions of data points in real-time with unbounded cardinality.
Syllables should always have only one vowel (technically nucleus) in them, which means that the bogus "syllables" the website is producing can't be real syllables - that might be the rule that you can't articulate. However, the website actually does not use GPT-2 to produce the syllables, it uses the library "pyhyphen".
This library provides (misleadingly) a function claiming to create a list of syllables from a word, which does not actually do that and instead splits the word up by all possible hyphenation points - and not every syllable break gets or indeed should get a hyphenation point. (For example, you do not want to break up prefixes or suffixes during hyphenation.)
[0] https://github.com/turtlesoupy/this-word-does-not-exist/blob...
I'd suggest this code to follow: https://github.com/lucidrains/stylegan2-pytorch/blob/fc22408.... The standard practice is to interpolate between the source latents and target latents and feed the interpolated values into your GAN. There are various ways of doing interpolation (i.e. not always straight linear interpolation)