In addition to Muon, NVIDIA also supports many other optimizers for the research community to explore, including: The ultimate form of orthogonalized optimizer MOP (Momentum Orthogonalized by Polar decomposition) An advanced SOAP variant that updates eigen basis per step with eigen decomposition plus KL correction in REKLS
Advancing Emerging Optimizers for Accelerated LLM Training with NVIDIA Megatron | NVIDIA Technical Blog…git clone https://github.com/NVIDIA/nemoclaw-community.git cd nemoclaw-community/examples/personal-community-sentiment-triage curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/main/install.sh | OPENSHELL_VERSION=v0.0…
…As AI factories scale to support increasingly distributed and autonomous workloads, network communication becomes a critical attack surface. DOCA Flow enables security policies to be enforced directly within the infrastructure layer, preventing…
…community contributions expand it regularly. Best Practice 6: Design models with deployment in mind. When choosing architectures, evaluate the deployment cost of exotic operations early. Sometimes a functionally equivalent but better-supported…
…Community Support Jetson AI Lab Jetson AI Lab is an interactive platform for learning and experimenting with AI on NVIDIA Jetson, offering hands-on projects, tutorials, and tools tailored for edge AI…
…Connect With the Community Engage with the Isaac Sim community by joining the NVIDIA forum. Find answers for troubleshooting and post your own questions. Extend Isaac Sim Workflows NVIDIA Cosmos NVIDIA Cosmos…
…It supports open source inference engines including SGLang, TensorRT™ LLM, and vLLM and simplifies the complexities of distributed serving by disaggregating the various phases of inference across different GPUs, intelligently routing requests…
…Tensor parallelism is extended to support sharded quantized weights, with scaling factors handled per shard to preserve numerical correctness. Megatron fused kernels and communication patterns are retained, ensuring that ternary quantization delivers…
…This optimization fuses computation and communication operations into a single kernel to normalize query and key together. The kernel can better overlap computation and communication, reducing kernel launch and memory read/write…
…GPUs, CPUs, DPUs, and internode communication. Scale AI Applications to the Data Center and Cloud with NVIDIA Nsight Systems GTC Demo Video Feature Spotlight on Python support in Nsight Developer Tools Optimize…
…Developers across the AI community have contributed to Dynamo and broadened its capabilities. Mooncake and Alibaba extended the Dynamo AIConfigurator with SGLang support; Microsoft tested and hardened Dynamo on Azure Kubernetes Service…