JFPuget

@jfpuget
4 Followers
24 Following
26 Posts
Machine learning at NVIDIA. CPMP on Kaggle. ex ENS ULM, ILOG CPLEX, IBM.

I am migrating to sigmoid.social. All my followers should move automatically. I'll follow again from the new account, and will follow more people as well.

Call 911 if you don't get any news from me by tonight...

Kidding apart I am following these great instructions: https://blog.djnavarro.net/posts/2022-11-03_what-i-know-about-mastodon/#how-do-i-move-my-account-to-a-new-server

Notes from a data witch - Everything I know about Mastodon

A hastily written guide for data science folks trying to navigate the fediverse.

Notes from a data witch

I just wrote a one liner that replaces a one page function.

I guess I will never be hired in a company owned by Musk.

To those leaving twitter, don't delete your account. Put where you went in your twitter bio so that others can find you there.
I will migrate to signoid.social, mastodon.social is too slow.

Dorsey has coined enough from so many people losing jobs overnight, he can go to hell with his bluesky and musk can put that $8 label on his forehead.

I am going to support @Gargron, the founder and lead developer of Mastodon 

Using lines of code to measure a developer contribution is pure nonsense.

I would value someone who actually decreases the number of loc via code factorization and redesign much higher than someone who duplicates code all over the place with copy and paste.

I tried to configure a twitter mastodon cross poster...
I share how to use pytorch DistributedDataParallel with Kaggle 2 T4 kernels there https://www.kaggle.com/competitions/feedback-prize-english-language-learning/discussion/363111
2x batch size in 0.5x the running time :D
Feedback Prize - English Language Learning

Evaluating language knowledge of ELL students from grades 8-12

The most tedious step when editing a notebook to turn it in a script I can use with torch DistributedDataParallel is to get rid of global variables...

Fun with #animation and patch scaling in #Matplotlib.

Here's the code in case you'd like to play around with it.
https://github.com/brohrer/how-to-train-your-robot/blob/main/chapter_4/patch_scaling.py

how-to-train-your-robot/patch_scaling.py at main · brohrer/how-to-train-your-robot

Content, code, and resources for the book How to Train Your Robot. - how-to-train-your-robot/patch_scaling.py at main · brohrer/how-to-train-your-robot

GitHub