Tag
56 articles
Learn how NVIDIA's NOOA framework simplifies AI agent development by letting developers create complex agents in just one Python class.
Learn how NVIDIA's new Alpamayo 2 Super AI model helps self-driving cars see, understand, and act on road information, and why this open-source tool is a big step toward safer autonomous vehicles.
Cursor Research has open-sourced Mixture-of-Kittens (MoK), a deterministic MoE training megakernel designed for high-performance computing environments like GB300 NVL72 racks, delivering up to 2.37x faster performance than public baselines.
NVIDIA has released Molt, a PyTorch-native agentic reinforcement learning framework designed to simplify and optimize RL development by integrating key components into a single asynchronous loop.
Jensen Huang has donated $75 million to Vanderbilt University's new art school in San Francisco, sparking over $100 million in additional funding. The gift underscores a growing trend among tech leaders to invest in arts and culture.
NVIDIA's srt-slurm framework simplifies distributed LLM serving benchmarking using SLURM, enabling reproducible workflows and advanced performance analysis.
NVIDIA has released Cosmos 3 Edge, a 4-billion-parameter open world model designed to run entirely on-device, enabling robots to reason and act autonomously in real-time environments.
NVIDIA has released Nemotron 3 Embed, an open embedding collection featuring three models, with the 8B checkpoint ranking #1 on the RTEB benchmark.
Learn to build and simulate an AI-powered radio access network platform similar to Nokia's AI-RAN, using NVIDIA's Aerial system and PyTorch for spectrum optimization.
NVIDIA introduces Nemotron-Labs-3-Puzzle-75B-A9B, a compressed hybrid MoE LLM delivering 2.03x server throughput, leveraging hardware-aware compression and knowledge distillation.
Learn how NVIDIA's new AI model Nemotron-Labs-3-Puzzle-75B-A9B uses compression and smart design to work faster and more efficiently than previous versions, without sacrificing quality.
NVIDIA's Cosmos framework tutorial demonstrates how to build a Colab-friendly, miniature version of Cosmos 3 world models using omnimodal Mixture-of-Transformers. The approach enables developers to experiment with multimodal prediction using synthetic data and shared cross-modal attention.