Ever wonder how Hugging Face moves so fast? Here’s one reason: 0 hours of meetings on my June calendar. An entire month to build, think, and ship. And it's not just me, it's our culture.

CVPR 2025 papers pt. 1 - Gaze-LLE Gaze-LLE simplifies gaze target estimation by building on top of a frozen DINOv2 visual foundation model; SOTA performance; open source code and model more papers: github.com/SkalskiP/top-c… ↓ more

CVPR 2025 papers pt. 2 - SAMWISE SAMWISE adds language understanding and temporal reasoning to SAM2; you can segment and track objects in videos just by describing them more papers: github.com/SkalskiP/top-c… ↓ more

DINOv2 meets text at #CVPR 2025! Why choose between high-quality DINO features and CLIP-style vision-language alignment? Pick both with dino.txt 🦖📖 We align frozen DINOv2 features with text captions, obtaining both image-level and patch-level alignment at a minimal cost. [1/N]

new blogpost on writing in the ~glorious~ age of LLMs

New video, starting to look at Diffusion Language Models. This one introduces some ideas, then shows how I turn ModernBERT into a LLaDA-style generative model. Lots of avenues to explore from here! Join me in playing with this? Project ideas in thread :) youtube.com/watch?v=Ds_cTc…

Andrej Karpathy's (Andrej Karpathy) keynote yesterday at AI Startup School in San Francisco.

New Moondream 2B release! 🌛 visual reasoning (!!) 🌜better object detection 🌛 improved UI understanding 🌜40% faster text generation

Phew, it's been a while. timm 1.0.16 released today to provide the image encoder for Gemma 3n. Additions kept on stacking so I haven't had chance to finalize a release since the last one to support SigLIP-2 backbones. Lots of stuff in there: * Gemma 3n encoder (via a

I compiled @bclaive 's opening lesson of our mini RAG optimization & eval course into these high density notes. You can consume the 30 minute talk in a few minutes in this format! 🏎️ Do ya'll like this format? LMK! I personally prefer this! hamel.dev/notes/llm/rag/…

Using a 12B model for image classification feels like overkill. You could just fine-tune a small DINOv2 or SigLIP 2 model using Transformers or Timm and you'll get great accuracy with very low inference costs

My NaFlexVit implementation is getting more flexy. In final verification stages of ROPE support, which means all timm ViT models based on the EVA model lineage (EVA, EVA02, Meta PE, Naver ROPE-ViT) can be loaded into NaFlexVit w/ support for native aspect, dynamic & variable

Today we restarted the Hamel Husain / Shreya Shankar evals course and to accompany this first week's class I'm publishing the first part of a series of annotated posts to accompany the course textbook. (Link in the 🧵) The aim was to give more examples from the ZenML LLMOps

🚀 Introducing GLiClass‑V3 – a leap forward in zero-shot classification! Matches or beats cross-encoder accuracy, while being up to 50× faster. Real-time inference is now possible on edge hardware. huggingface.co/collections/kn… #TextClassification #NLP #ZeroShot #GLiClass

At Answer.AI, we integrate Stripe into lots of projects. Every time, I found myself doing the same dance: create product, create price, create checkout session. Then hunting docs for parameters for each. So we built FastStripe, a self-documenting Stripe SDK that's easy to

Don't use any ML tracking solution for your training runs before reading this!

MetaCLIP-2 'worldwide' (multi-lingual) models are now supported in OpenCLIP and encoders in timm (main branch so far). See the collection on Hugging Face huggingface.co/collections/ti…

In order to celebrate the release of the print version for the Ultra-Scale Playbook (of which I have no affiliation with and love deeply), I'm going to be giving away 5 copies! To enter, simply like + retweet this tweet. Winners will be selected at random 10AM EST on the 13th

The 3rd edition of my book Deep Learning with Python is being printed right now, and will be in bookstores within 2 weeks. You can order it now from Amazon or from Manning. This time, we're also releasing the whole thing as a 100% free website. I don't care if it reduces book

Moondream 3 Preview adds frontier-level visual reasoning while retaining Moondream 2's blazing speed. No compromises.