Niels Rogge
@NielsRogge
ML Engineer @ML6team, part-time at @huggingface. @KU_Leuven grad. General interest in machine learning, deep learning. Making AI more accessible for everyone!
Today my Transformers-Tutorials repo hit 2,000 stars on @github! 🤩 Very greatful :) the repo contains many tutorial notebooks on inference + fine-tuning with custom data for Transformers on all kinds of data; text, images, scanned PDFs, videos ⭐ github.com/NielsRogge/Tra…
BREAKING: we've partnered with @metaai and @paperswithcode to build a successor to Papers with Code (which was sunsetted yesterday) PWC, founded by @rbstojnic and @rosstaylor90 has been an invaluable resource for AI scientists and engineers over the years (and an inspiration…
It’s fascinating how people, after learning that fossil fuels (carbon, gas, oil) heat up the planet, keep on burning more fossil fuels Lots of heatwaves this summer, even more soon?
OpenAI Stargate is applying the same strategy as @elonmusk‘s Colossus cluster and is currently installing 29 gas turbines. Each GE LM2500XPRESS gas turbines generates 34MW When all 29 will operational, it be enough to generate ~1000MW enough to 500k GB200 NVL72 chips in the…
Weird ChatGPT bug: I asked it an English question and it replied everything in Italian It also briefly showed <turn_image> and </turn_image> tokens before rendering images in its response ChatGPT can definitely still be improved a lot for location-based recommendations
Kinda cool that slowly but surely we get to know all the details on how models like Claude Sonnet 3.5 came about
Kimi K2 paper dropped! describes: - MuonClip optimizer - large-scale agentic data synthesis pipeline that systematically generates tool-use demonstrations via simulated and real-world environments - an RL framework that combines RLVR with a self- critique rubric reward mechanism…
Open-source audio scene is quite on 🔥 lately! - @kyutai_labs STT, TTS modules and Unmute fully open-sourced - @nvidia drops 3 models: Parakeet (beats Whisper), Audio Flamingo 3 and Canary-Qwen-2.5B (new SOTA on @huggingface leaderboard) - @MistralAI released 3B and 24B Voxtral
Wait wtf
BREAKING: Claude Code PMs Boris Cherny and Cat Wu have returned to Anthropic after a brief stint at Cursor.
My fridge is about to get smarter again
LG AI Research just dropped EXAONE 4.0 on @huggingface: a unified LLM that blends non-reasoning and reasoning modes. It features agentic tool use & expanded multilingual support (EN, KR, ES) across 1.2B & 32B sizes. Competitive even against frontier models.
So pumped to hear @NielsRogge speak at the Brussels meetup!
So the only way to get a decent search on @X is by paying for Grok 4? 🤨
And, you can now use Grok 4 to make advanced searches on 𝕏.
It feels really bad to me that while humanity should be solving sustainable energy and transportation, we're building ever larger datacenters so that we can ask "pls fix this" in Cursor or ask for "a cow on a beach" video It becomes harder and harder for me to justify it tbh
NEWS: Mark Zuckerberg has just announced that Meta will be spending hundreds of billions of dollars to build massive GPU compute clusters. Mark: "We're building several multi-GW clusters. We're calling the first one Prometheus and it's coming online in '26. We're also building…
Dario when he sees a Chinese start-up yet again releasing an open model with similar capabilities as his at much lower prices
Prior to the release of Kimi K2 the team of @Kimi_Moonshot sent me a doc on how to connect their model with Claude Code! Besides the awesome model release I love how approachable the Kimi folks are. Way to go!!
Very interesting - you can use Kimi with the Anthropic API. This means, perhaps most importantly, that you can now use Kimi with Claude Code! 🤯 x.com/op7418/status/…