A real-time object detector much faster and accurate than YOLO with Apache 2.0 license just landed to @huggingface transformers 🔥

D-FINE is the sota real-time object detector that runs on T4 (free Colab) 🤩
Keep reading for the paper explainer, notebooks & demo 👀 > Collection with all checkpoints and demo huggingface.co/collections/us…
Notebooks:
> Tracking github.com/qubvel/transfo…
> Inference github.com/qubvel/transfo…
> Fine-tuning github.com/qubvel/transfo…

h/t @qubvelx @ariG23498 pic.x.com/oI95hkb6Xr

DO NOT SLEEP ON THIS MODEL @Kimi_Moonshot 🤯

Kimi-VL-A3B-Thinking is the first ever capable open-source reasoning VLM with MIT license ❤️
> it has only 2.8B activated params 👏
> it's agentic 🔥
> surpasses gpt-4o

I've put it to test (see below ⤵️) Try it:
This model consists of a dynamic res handling MoonViT encoder, a projection layer and a 16B MoE decoder (with 2.8B active params)
the paper introduces an interesting pre-training pipeline to handle long context and the model saw 4.4T tokens huggingface.co/spaces/moonsho…

Microsoft released the most powerful vision language action model this week 🔥

MAGMA-8B can operate in both physical and digital world: embodied robots, web automation and more! 🤯 Model: huggingface.co/microsoft/Magm…

Demo for UI agents: huggingface.co/spaces/microso…

New gaming agent demo based on Magma: huggingface.co/spaces/microso…

people who are baffled by DeepSeek have been and still are sleeping on Qwen, InternLM, ByteDance and Tencent

here's couple of fan-favorite models from them 🪭 Qwen2VL: best open-source vision language model
here's demo for 7B, use 72B for longer context length and better performance
huggingface.co/spaces/Ganymed…

ByteDance just dropped SA2VA: a new family of vision LMs combining Qwen2VL/InternVL and SAM2 with MIT license 💗

The models are capable of tasks involving vision-language understanding and visual referrals (referring segmentation) for images and videos ⏯️
take a look 🧶 The models come in 1B, 4B and 8B and are based on InternVL2.5 for base architecture and Qwen2, Qwen2.5 and InternLM2 for language model part (depending on the checkpoint)

All of them are in this collection huggingface.co/collections/By…

NVIDIA solved physics and open-sourced it? Can we just build our own autonomous robots now? 🤯

They released Cosmos: new family of open world foundation models (WFMs) 🌌

Unwrapping the release and why it's so revolutionary 🧶 They have released 7 models:
- four autoregressive video2world models
- four diffusion-based video2world and text2world models
- prompt upsampler
- content safety model

Collection on Hugging Face: huggingface.co/collections/nv…
Tech report: research.nvidia.com/publication/20…
Try them here: (try here build.nvidia.com/nvidia/cosmos-…

supercharge your LLM apps with smolagents 🔥

however cool your LLM is, without being agentic it can only go so far

enter smolagents: a new agent library by @huggingface to make the LLM write code, do analysis and automate boring stuff! smolagents is a barebones library to unlock both native and traditional tool calling for language models

LLMs can already write code and do reasoning, so why bother yourself with writing the tool?

CodeAgent class is here for it! see it in action below

Microsoft released a groundbreaking model that can be used for web automation, with MIT license 🔥👏

OmniParser is a state-of-the-art UI parsing/understanding model that outperforms GPT4V in parsing. 👏 Model:
Interesting highlight for me was Mind2Web (a benchmark for web navigation) capabilities of the model, which unlocks agentic behavior for RPA agents.

no need for hefty web automation pipelines that get broken when the website/app design changes! Amazing work.huggingface.co/microsoft/Omni…

this is the BEST vision language model I have ever tried!

Aria is a new model by @rhymes_ai_: a 25.3B multimodal model that can take image/video inputs 🤩

They release the model with Apache-2.0 license and fine-tuning scripts as well 👏
I tested it extensively, keep reading to learn more 🧶 The model is open-sourced here: huggingface .co/rhymes-ai/Aria

The authors have released fine-tuning examples on RefCOCO, NextQA and NLVR and inference examples: github .com/rhymes-ai/Aria

Try the demo here: rhymes .ai

It's super nice that you can get started with this model using @huggingface transformers 🤗

I'm bullish on this foundation OCR model called GOT 📝 @eccvconf

This model can transcribe anything and it's Apache-2.0!

Keep reading to learn more 🧶 This model can take in screenshots of tables/LaTeX and output formatted text, music sheets, charts, literally anything to meaningful format!

Try it huggingface.co/spaces/stepfun…

I was giving consultancy about multimodal RAG to some external people, wanted to post it here

stop using OCR + LLMs. if you want to retrieve, use ColPali, if you want RAG from docs, use vision language models my tweet might come off a bit sensational, but I'm mostly talking about documents that aren't just only plain text. leaving these here, pdf parser will parse incorrectly and slowly

https://twitter.com/mervenoyann/status/1831416247344623632

it's more convoluted to build a demo to SAM with text to box prompting

one needs to find first frame where the object of interest appears, then give it to SAM2 as a mask and let it propagate
github.com/facebookresear… one technical problem I had was for each text prompt one needs to tune threshold for text -> box prediction for a given frame to make sure right mask is given

ofc it's not a problem, but also there's gradio not having a component to pick frames like in Meta's demo

Forget any document retrievers, use ColPali 💥💥

Document retrieval is done through OCR + layout detection, but it's overkill and doesn't work well! 🤓

ColPali uses a vision language model, which is better in doc understanding 📑
keep reading ⤋ ColPali: (mit license!)
Blog post:

The authors also released a new benchmark for document retrieval, ViDoRe Leaderboard, submit your model! huggingface.co/vidore/colpali
huggingface.co/blog/manu/colp…
huggingface.co/spaces/vidore/…

Real-time DEtection Transformer (RT-DETR) landed in @huggingface transformers 🤩 with Apache 2.0 license 😍

do DETRs Beat YOLOs on Real-time Object Detection?
keep reading 👀 @huggingface short answer, it does!

📖 notebook:
🔖 models:
🔖 demo:
📝 paper: github.com/merveenoyan/ex…
huggingface.co/models?search=…
huggingface.co/spaces/merve/R…
huggingface.co/papers/2304.08…

DocOwl 1.5 is the state-of-the-art document understanding model by Alibaba with Apache 2.0 license 😍📝
time to dive in and learn more 🧶 This model consists of a ViT-based visual encoder part that takes in crops of image and the original image itself

Then the outputs of the encoder goes through a convolution based model, after that the outputs are merged with text and then fed to LLM

SigLIP just got merged to 🤗transformers and it's super easy to use!
To celebrate this, I have created a repository on various SigLIP based projects 🥳
But what is it and how does it work?
SigLIP an vision-text pre-training technique based on contrastive learning. It jointly trains an image encoder and text encoder such that the dot product of embeddings are most similar for the appropriate text-image pairs
The image below is taken from CLIP, where this contrastive pre-training takes place with softmax, but SigLIP replaces softmax with sigmoid. 📎 Highlights✨
🖼️📝 Authors used medium sized B/16 ViT for image encoder and B-sized transformer for text encoder
😍 More performant than CLIP on zero-shot
🗣️ Authors trained a multilingual model too!
⚡️ Super efficient, sigmoid is enabling up to 1M items per batch, but the authors chose 32k (see saturation on perf below)

There are many known “foundation models” for chat models, but what about computer vision? 🧐

In this thread, we’ll talk about few of them 👇
🖼️ Segment Anything Model
🦉 OWLViT
💬 BLIP-2
🐕 IDEFICS
🧩 CLIP
🦖 Grounding DINO

Let’s go! ✨ How does this work for vision? 👀
Foundation models for vision are powerful models that can be out of box or fine-tuned (on your use case, if needed) to solve many problems!
All the models in this thread are open-source, can be found on Hugging Face Hub and can be used with 🤗 transformers with only few lines of code ♥️

AWS 🤝 Hugging Face for LLM Deployment 🤩
Hugging Face & AWS now offer LLM Containers to deploy any open-source LLM of your choice 🫡
See how @_philschmid deploys OpenAssistant pythia 12B model in this blog post 👉 huggingface.co/blog/sagemaker…
details in 🧶 @_philschmid There's a new open-source LLM every day, but how to deploy them is a question 🧐
📖 Having your LLM under your control enables your company to work in private, govern data and reduce costs,
📖 Hugging Face has put tons of know-hows on deploying these models into LLM container!

We had many great submissions for our Keras DreamBooth sprint, wanted to share a few of them in this 🧶 Some of the folks have fine-tuned on video games, hogwarts legacy and return to monkey island (which I really like 😍 )

PyCon SE started with the keynote of @julsimon 🤗 nothing says remote first more than meeting your colleague for the first time in a conference 😏

a 🧵 on how to make the best of model repositories on 🤗Hub this is how a model repository looks 👀
there's a model card which is good for documenting & communicating your model 📖 if it's a library with Hub integration you can get a widget that works out of the box 📦
also on bottom left you have links to Spaces that use your model ✨