Google has released FunctionGemma, a specialized Gemma 3 270M model fine-tuned for function calling—turning natural-language commands into structured API/tool calls for local agents. Why this matters for edge AI engineers Local-first “action agents”: FunctionGemma is positioned as an on-device “action” model (or a lightweight controller that can route complex requests to larger models like Gemma […]
This article was originally published at Analog Devices’ website. It is reprinted here with the permission of Analog Devices. Why Smaller Models Matter Shrinking AI models isn’t just a nice-to-have—it’s a necessity for bringing powerful, real-time intelligence to edge devices. Whether it’s smartphones, wearables, or embedded systems, these platforms operate with strict memory, compute, and
97% Smaller, Just as Smart: Scaling Down Networks with Structured Pruning Read More +
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. Learn how to integrate vision language models into video analytics applications, from AI-powered search to fully automated video analysis. Today’s computer vision systems excel at identifying what happens in physical spaces and processes, but lack the abilities to explain the
AI On: 3 Ways to Bring Agentic AI to Computer Vision Applications Read More +
News Summary: The Nemotron 3 family of open models — in Nano, Super and Ultra sizes — introduces the most efficient family of open models with leading accuracy for building agentic AI applications. Nemotron 3 Nano delivers 4x higher throughput than Nemotron 2 Nano and delivers the most tokens per second for multi-agent systems at scale through a
NVIDIA Debuts Nemotron 3 Family of Open Models Read More +
This blog post was originally published at Arm’s website. It is reprinted here with the permission of Arm. NeurIPS 2025 provided Arm with a unique opportunity to share the latest technical trends and insights with the global AI research community. NeurIPS is one of the world’s leading AI research conferences, acting as a thriving global hub for
Arm at NeurIPS 2025: How AI research is shaping the future of intelligent computing Read More +
This blog post was originally published at Visidon’s website. It is reprinted here with the permission of Visidon. In the world of embedded vision—whether for mobile phones, surveillance systems, or smart edge devices—image quality in low-light conditions can make or break user experience. That’s where advanced AI-based denoising algorithms come into play. At our company, we
Low-Light Image Enhancement: YUV vs RAW – What’s the Difference? Read More +
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. The new Mistral 3 open model family delivers industry-leading accuracy, efficiency, and customization capabilities for developers and enterprises. Optimized from NVIDIA GB200 NVL72 to edge platforms, Mistral 3 includes: One large state-of-the-art sparse multimodal and multilingual mixture of
NVIDIA-Accelerated Mistral 3 Open Models Deliver Efficiency, Accuracy at Any Scale Read More +
This blog post was originally published at NVIDIA’s website. It is reprinted here with the permission of NVIDIA. NVIDIA releases new AI tools for speech, safety and autonomous driving — including NVIDIA DRIVE Alpamayo-R1, the world’s first open industry-scale reasoning vision language action model for mobility — and a new independent benchmark recognizes the openness and
NVIDIA Advances Open Model Development for Digital and Physical AI Read More +
This blog post was originally published at Quadric’s website. It is reprinted here with the permission of Quadric. The term “model zoo” first gained prominence in the world of AI / machine learning beginning in the 2016-2017 timeframe. Originally used to describe open-source public repositories of working AI models – the most prominent of which today
Let’s Visit the Zoo Read More +
This blog post was originally published at ENERZAi’ website. It is reprinted here with the permission of ENERZAi. Today, we’ll share results where we re-trained the original Whisper for optimal Korean ASR(Automatic Speech Recognition), applied Post-Training Quantization (PTQ), and provided a richer Pareto analysis so customers with different constraints and requirements can pick exactly what
Small Models, Big Heat — Conquering Korean ASR with Low-bit Whisper Read More +
Here you’ll find a wealth of practical technical insights and expert advice to help you bring AI and visual intelligence into your products without flying blind.
Berkeley Design Technology, Inc.
PO Box #4446
Walnut Creek, CA 94596
