Johannes

@hf.co Computer Vision Hangout on Wednesday, 16th April at 7.15pm CEST(UTC+2)

I'll be talking about what is new in Computer Vision at Hugging Face, we'll have a presentation by a community member about his work and a fun Computer Vision Quiz!

Join us on the HF discord server: HF.co/join/discord

PEP 751 – A file format to record Python dependencies for installation reproducibility | peps.python.org This PEP proposes a new file format for specifying dependencies to enable reproducible installation in a Python environment. The format is designed to be human-readable and machine-generated. Installe...

peps.python.org/pep-0751/

🐍"PEP 751 – A file format to record Python dependencies for installation reproducibility" has been accepted 🥳

The 31st March should be marked a special day for all Pythonistas, commemorating this (long overdue) achievement

Amazingly Simple Graphic Design Software – Canva Amazingly Simple Graphic Design Software – Canva

ChatGPT image generation seems to make a lot of headlines recently.

So I collected some use cased and experimented a bit myself. As a result I created this (not very serious) presentation, called "The End of Graphic Designers":

www.canva.com/design/DAGjR...

With the new image generation capability in ChatGPT, the path to AGI is closer than ever! We can even skip steps 3 and 4 on the way 😉

(this post may contain sarcasm)

Hallucinations in Multimodal Large Language Models

📣 Slides, Slides, who wants some slides? 📣

Anyway, here are mine from my presentation on "Hallucinations in MLLMs" on the @hf.co discord.

They are very similar to the ones I've used before for my PyCon Africa talk, just with a few additions.

speakerdeck.com/johko/halluc...

📚 ML Reading Group - 25th, Jan at 8.15pm CET 📚

Tomorrow I'm hosting my first Reading Group session of the year on the @hf.co discord server: discord.com/events/87954...

We will look at Hallucinations in MLLMs, roughly following this paper: arxiv.org/abs/2404.18930

But hey, at least the library name Triton will finally make sense.

a statue of a mermaid holding a trident in the water ALT: a statue of a mermaid holding a trident in the water

Can't wait for the Stargate follow-up project: Stargate Atlantis

"As we chose to worsen the climate crisis in favour of our batshit crazy AGI idea, we accidentally flooded some countries. But we figured out how to use sea water for cooling our servers, so AGI is pretty close now"

Realizing this comic speaks the truth took a lot of anxiety out of the whole Code Review process for me.

People coming over from Threads?

I guess it is Starter Pack time again!

Here is my contribution ;)

go.bsky.app/Nik64nt

🌟 Happening Today 🌟

at 8:30pm CET on 20th Dec we will have our End-of-Year Computer Vision Hangout on the @hf.co discord server: tinyurl.com/5t5ep7xc

We'll have a look on what happened in 2024 at HF including datasets, new transformers integrations and much more.

+ we'll have a QUIZ!!! 🎉

Screenshot of Table of Contents (Part 1) Contents 1 Introduction 217 2 Positionality 221 3 Overview of Risks and Harms Associated with Computer Vision Systems and Proposed Mitigation Strategies 223 3.1 Representational Harms . . . . . . . . . . . . . . . . . . . 223 3.2 Quality-of-Service and Allocative Harms . . . . . . . . . . 229 3.3 Interpersonal Harms . . . . . . . . . . . . . . . . . . . . . 237 3.4 Societal Harms: System Destabilization and Exacerbating Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . 245 4 Frameworks and Principles for Computer Vision Researchers 266 4.1 Guidelines for Responsible Data and Model Development . 267 4.2 Measurement Modeling . . . . . . . . . . . . . . . . . . . 271 4.3 Reflexivity . . . . . . . . . . . . . . . . . . . . . . . . . . 273 5 Reorientations of Computer Vision Research 276 5.1 Grounded in Historical Context and Considering Power Dynamics . . . . . . . . . . . . . . . . . . . . . . . 276 5.2 Small, Task Specific . . . . . . . . . . . . . . . . . . . . . 279 5.3 Community-Rooted . . . . . . . . . . . . . . . . . . . . . 280

Screenshot of Table of Contents (Part 2) 6 Systemic Change 285 6.1 Collective Action and Whistleblowing . . . . . . . . . . . . 285 6.2 Refusal/The Right not to Build Something . . . . . . . . . 287 6.3 Independent Funding Outside of Military and Multinational Corporations . . . . . . . . . . . . . . . . . . . . . . . . . 289 7 Conclusion 291 References 293

Dear computer vision researchers, students & practitioners🔇🔇🔇

Remi Denton & I have written what I consider to be a comprehensive paper on the harms of computer vision systems reported to date & how people have proposed addressing them, from different angles.

PDF: cdn.sanity.io/files/wc2kmx...

Feels like the Deepdream times were the slightly weird but innocent days of AI art

Sundays are for books 📖

And while I like to read technical books, I also love to read novels.

I'm a huge fan of the works by Sir Terry Pratchett and his timeless discworld series 🧙‍♂️

I feel Hex, thought up as an analogy to computers, also works great as an analogy for modern AI 🐜🐜🐜

nds - Neddersass’sch - Low German Join and contribute to the dataset nds - Neddersass’sch - Low German

The sprint for crowd sourced annotations with argilla is in full swing over at data-is-better-together-fineweb-c.hf.space

I've just contributed 100 examples to this dataset:
data-is-better-together-fineweb-c.hf.space/share-your-p...

Big thanks to @dvilasuero.hf.co, @nataliaelv.hf.co and team 🙌

Help shape the future of multilingual Open Source AI!

Join the FineWeb 2 Community Annotation Sprint to create an open training dataset with full transparency and human validation in many languages.

Review datasets in your language and help identify the best sources for training.

The Lichess database of games, puzzles, and engine evaluations is now on @hf.co - https://huggingface.co/Lichess. Billions of chess data points to download, query, and stream and we're excited to see what you'll build with it! ♟️ 🤗

Congrats 🎉

But now I'm sad to not have you around in the HF Fellows Slack Channel anymore 😢

Robot Holmes is back 🕵️

Last year I gave the talk "Robot Holmes and the Vision-Language Murder Mysteries", an intro to VLM models and tasks, at @pydatalondon.bsky.social and @europython.bsky.social.

Now I started to actually turn it into a story here: johko.github.io/robot-holmes...

GitHub - johko/computer-vision-course: This repo is the homebase of a community driven course on Computer Vision with Neural Networks. Feel free to join us on the Hugging Face discord: hf.co/join/disc... This repo is the homebase of a community driven course on Computer Vision with Neural Networks. Feel free to join us on the Hugging Face discord: hf.co/join/discord - johko/computer-vision-course

🌟 500! 🌟

Our Community Computer Vision Course Repo just reached 500 stars on GitHub: github.com/johko/comput... 🤩

I'm really proud of all the amazing content people from the community have contributed here and that they still keep on adding very cool and helpful material 💪

No, it’s that when the rightfully-angry-but-insufficiently-informed masses start swarming with pitchforks & torches, then even the well-intentioned do-gooders have to lay low and do their work out of sight.
It imposes a tax on open collaboration. Which actually then tends to privilege existing power

You know what's cool?
Discussing data scraping and related GDPR and TOS breaches.

You know what's not cool?
Going off the rails and insulting people who meant well and did a mistake.

I would love to see some more constructive discussion than on that former bird site, but maybe that is my utopia.

Julia REPL which activates the Bluesky project, and calls a function which simulates creating a new post on the julialanguage Bluesky account!

Hello @bsky.app from the #JuliaLang community!

We're excited to start posting here on Bluesky in addition to our other social media, all of which can be found here: julialang.org

We'll keep you updated on our fantastic community, our conferences, #JuliaLang updates, and more on this profile!

Btw, I put together this nice little starter pack of AI/ML people I like to read from 😉 (and myself 🙃)

go.bsky.app/Nik64nt

Still one of my favourite books about data and statistics.

Packed with illustrative examples, it is a good casual evening read and a nice refresher about general statistics.