WaveGrad by @cnxhk et al is a clever new neural net for generating speech. It starts with random noise, and iteratively denoises, essentially hallucinating a voice in the noise. Listen to the voice emerge, as the number of iterations increases with time 🔉
the video shows mel-spectrogram features that are used to guide the denoising and specify the speech to produce. These could be the output of a text-to-speech system
I managed to make a pretty good 3d model of the Julia set ring. Is anyone interested in helping me 3d print it?
I first created code to generate the perimeter of a Julia set in mathematica, by repeatedly applying the inverse map to (1, -1, i, -i). Then I wrote some python code to generate a .obj 3d model, and rendered in blender