In my dream, the process of matrix multiplication was depicted as stone pillars falling from sky... And I told Text2Image AI to draw 'matrix multiplication depicted as stone pillar...(blah)' and it gave me back the images of stone dicks.

The beats in the music are continually morphing into another using the neural network model 'AutoEncoder'.

'5D라 Radio VOL.1' Full Video
https://youtu.be/BIlJEICBZOo

SPA(Stone Pillar Anthem)

Recording of a RAVE model trained on a selection of Martsman tracks. Real time latent space manipulation via nn~ in Pure Data using only the decoder. 

RAVE is a realtime variational autoencoder for fast and high-quality neural audio synthesis by Antoine Caillon and Philippe Esling.
https://arxiv.org/abs/2111.05011
https://github.com/acids-ircam/RAVE

RAVE /// Rec 20230426 decoder var complex (norm) --- model martsparse V1 200K

Recording of a RAVE-Latent Diffusion model trained on a selection of Martsman tracks using seed interpolation and a latent length of 8192.

RAVE-Latent Diffusion is a denoising diffusion model designed to generate new RAVE latent codes with a large context window, written by  Moisés Horta Valenzuela / 𝔥𝔢𝔵𝔬𝔯𝔠𝔦𝔰𝔪𝔬𝔰
github.com/moiseshorta/RAVE-Latent-Diffusion

RAVE is a realtime variational autoencoder for fast and high-quality neural audio synthesis by Antoine Caillon and Philippe Esling.
arxiv.org/abs/2111.05011
github.com/acids-ircam/RAVE

RAVE Latent Diffusion /// seed 91827536 Slerp --- model martssel V1 400k Cap96

First submission for Weekly Beats 2016. 4+ minutes of noise from an untrained autoencoder, imported into LSDJ.

grueling noise composition #1

Three superimposed songs generated by an autoencoding Deep Belief Network, trained on the MIDI guitar parts from the Metallica albums Master of Puppets and Ride the Lightning.

Master of Neural Puppets

Three superimposed songs generated by an autoencoding Deep Belief Network, trained on the MIDI guitar parts from the Metallica albums Master of Puppets and Ride the Lightning.

Finetuning during training did better this time. And during sampling I tried varying the threshold to get parts that varied in how many notes are being played.


Master of Neural Puppets II

Martin Heinze is a sound artist, composer and electronic musician working in the field of experimental electronic dance music and algorithmic composition.

Label = Weevil Neighbourhood @weevilorchestra

Popular tracks tagged #autoencoder