Log inSign up
Image
user avatar
NVIDIA AI
NVIDIA
@NVIDIAAI
Teaching your AI new tricks.
Santa Clara, CA
developer.nvidia.com
Joined June 2016
890
Following
326.3K
Followers
RepliesArticlesMedia
  • Pinned
    user avatar
    NVIDIA AI
    NVIDIA
    @NVIDIAAI
    Jun 4
    Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model built for long-running agents. It delivers 5x faster inference and lowers the cost of complex agentic tasks by up to 30% versus other open frontier models.
    Image
    00:00
  • user avatar
    NVIDIA AI
    NVIDIA
    @NVIDIAAI
    12h
    A long-context model's serving speed is largely decided before training starts. Attention used to be a small part of a model's inference cost, but its share grows sharply as context windows expand. Once attention is the majority of the work, faster kernels stop being enough, and
    Three pie charts show changing ratios of elementwise ops, attention, and FFN for 4K, 32K, and 128K tokens.
  • user avatar
    NVIDIA AI
    NVIDIA
    @NVIDIAAI
    16h
    Hermes Agent 🤝 NeMo Relay
    user avatar
    Teknium 🪽
    Nous Research
    @Teknium
    Aug 2
    Hermes Agent is now dramatically more efficient, especially for smaller/weaker/local models! With the help of @nvidia's Nemo Relay and several other strategies Hermes was able to identify a ton of optimizations beyond just saving tool execution time or memory - but also turns
    Image
  • user avatar
    NVIDIA AI
    NVIDIA
    @NVIDIAAI
    Jul 31
    Humans count the boxes in this image, hidden ones included, with 82.1% accuracy. The best off-the-shelf multimodal model manages 17.7%. Spatial-IQ is a diagnostic benchmark from NVIDIA Research that breaks 3D object counting into nine perceptual and cognitive sub-tasks, from
    Image
    GIF