Galaxy Zoo Challenge: code published

Reading time ~1 minute

Some two weeks ago I posted about my solution for the Galaxy Zoo challenge on Kaggle. Today I’ve published the code with documentation on GitHub.

Get it here:

git clone git://github.com/benanne/kaggle-galaxies.git

Have a look at the README file for instructions on how to generate the winning solution, or check out doc/documentation.pdf for more information.

The code is available under a BSD 3-clause licence. If you’ve found it useful, dropping me a line to let me know how you used it is appreciated. Have fun with it!

Read More

Diffusion language models

Diffusion models have completely taken over generative modelling of perceptual signals -- why is autoregression still the name of the game for language modelling? Can we do anything about that? Continue reading

Guidance: a cheat code for diffusion models

Published on May 26, 2022

Diffusion models are autoencoders

Published on January 31, 2022