Last December, Google released PaliGemma 2: a new family of pre-trained (pt) PaliGemma vision language models (VLMs) based on SigLIP and Gemma 2. The models come in three different sizes (3B, 10B, 28B) and three different resolutions (224×224, 448×448, 896×896). Today, Google is releasing PaliGemma 2 mix: fine-tuned on a mix of vision language tasks, including OCR, long and short captioning and more. PaliGemma 2 pretrained (pt) variants are great vision language models to transfer on a given task at […]

SmolVLM2 represents a fundamental shift in how we think about video understanding – moving from massive models that require substantial computing resources to efficient models that can run anywhere. Our goal is simple: make video understanding accessible across all devices and use cases, from phones to servers. We are releasing models in three sizes (2.2B, 500M and 256M), MLX ready (Python and Swift APIs) from day zero. We’ve made all models and demos available in this collection. Want to try […]

Today Google releases a new and better family of multilingual vision-language encoders, SigLIP 2. The authors have extended the training objective of SigLIP (sigmoid loss) with additional objectives for improved semantic understanding, localization, and dense features. SigLIP 2 models outperform the older SigLIP ones at all model scales in core capabilities, including zero-shot classification, image-text retrieval, and transfer performance when extracting visual representations for Vision-Language Models (VLMs). A cherry on top is the dynamic resolution (naflex) variant. This is useful […]

(This post was authored by hlky and Sayak) When operating with latent-space diffusion models for high-resolution image and video synthesis, the VAE decoder can consume quite a bit more memory. This makes it hard for the users to run    

In the last few months, many new real-time speech models have been released and entire companies have been founded around both open and closed source models. To name a few milestones: OpenAI and Google released their live multimodal APIs    

The Indian Institute of Science IISc and ARTPARK partner with Hugging Face to enable developers across the globe to access Vaani, India’s most diverse open-source, multi-modal, multi-lingual dataset. Both organisations share a commitment to building inclusive, accessible, and state-of-the-art AI technologies that honor linguistic and cultural diversity. Partnership The partnership between Hugging Face and IISc/ARTPARK aims to increase the accessibility and improve usability of the Vaani dataset, encouraging the development of AI    

So, you’ve built your agent. It takes in inputs and tools, processes them, and generates responses. Maybe it’s making decisions, retrieving information, executing tasks autonomously, or all three. But now comes the big question – how effectively is it performing? And more importantly, how do you know? Building an agent is one thing; understanding its behavior is another. That’s where tracing and evaluations come in. Tracing allows you to see exactly what your agent is doing step by step—what inputs […]

With the release of the Aya Vision family, our new 8B and 32B parameter vision-language models (VLMs), we are addressing one of the biggest challenges in AI: bringing multilingual performance to multimodal models. Aya Vision is Cohere For AI‘s latest open-weight multilingual and multimodal model family, designed to be a strong foundation for language and vision understanding across 23 languages. It builds on the success of Aya Expanse, state-of-the-art multilingual language models, and extends it using a combination of advanced […]

TL;DR of L2D, the world’s largest self-driving dataset! 90+ TeraBytes of multimodal data (5000+ hours of driving) from 30 cities in Germany 6x surrounding HD cameras and complete vehicle state: Speed/Heading/GPS/IMU Continuous: Gas/Brake/Steering and discrete actions: Gear/Turn Signals    

Over the last few weeks, we have focused our efforts on reproducing the competitive programming (code reasoning) aspects of the DeepSeek-R1 recipe. In this post, we are excited to share: The construction of CodeForces-CoTs: a dataset of nearly 100k high-quality samples distilled from R1 to produce solutions in C++ and Python. The IOI benchmark: a new benchmark of challenging problems from the 2024 International Olympiad in Informatics (IOI). OlympicCoder: two fine-tuned 7B and 32B code models that outperform closed-source frontier […]