Blog

• Some of my open-source contributions

  A non-exhaustive, continually growing list.

• Performance and scikit-learn (4/4)

  Pairwise Distances Reductions: extra notes on technical details, benchmarks and further work.

• Performance and scikit-learn (3/4)

  Pairwise Distances Reductions: Abstracting the k-nn search pattern

• Performance and scikit-learn (2/4)

  Hardware scalability issue: the k-neighbors search example

• Performance and scikit-learn (1/4)

  Context: the current state of scikit-learn performance

• Performance and scikit-learn (0/4)

  A series of blog-posts.

• Academic Curriculum

• Vectorised implementation of splines

  All you need is numpy.einsum.

• Web Craftmanship

  « Une expérience bien conçue et bien conduite procure un plaisir comparable à celui qu’éprouve l’artisan devant un travail bien fait. »

• Understanding Multilayer Perceptron in Depth

  It’s all building blocks.

Page 1 / 1