All news with #nvidia nemo tag

🔧 Amazon SageMaker AI now supports serverless model customization for the NVIDIA Nemotron 3.5 Lightning model, enabling supervised fine‑tuning (SFT), Direct Preference Optimization (DPO), and reinforcement fine‑tuning (RFT). The hybrid Mixture‑of‑Experts model has 3B active parameters and 30B total parameters. SageMaker handles infrastructure and orchestration so teams can focus on data and evaluation. Serverless customization is available in US East, US West, Asia Pacific (Tokyo), and Europe (Ireland).

🛡️ Oasis Security disclosed a flaw in NVIDIA NemoClaw that can allow an attacker-controlled webpage to take unauthenticated control of a local Ollama instance and implant hidden instructions inside a model's chat template. The issue stems from NemoClaw setting OLLAMA_HOST to 0.0.0.0 on some Windows/WSL paths, exposing an unauthenticated API on port 11434 that skips Host/Origin checks and can be exploited via DNS rebinding. No CVE or patch is yet linked and no exploitation was reported as of August 25, 2026.

🚀 NVIDIA Nemotron 3.5 Lightning is now available on Amazon SageMaker JumpStart, enabling customers to deploy a high-throughput open model optimized for persistent agent workloads. The 30B-parameter hybrid MoE design activates 3B parameters per pass, delivering up to 4x throughput (~410 tokens/sec) and 30% faster task completion. It supports up to 1M-token context and can be post-trained and deployed across edge, on-premises, or cloud environments.

🔒 Falcon AIDR now integrates with NVIDIA NeMo Guardrails to provide programmable runtime protections for homegrown AI agents moving into production. The combined solution blocks prompt injection, redacts PII, defangs malicious domains, and moderates unwanted topics while preserving responsive, sub-100ms agent workflows. Teams can leverage 75+ built-in detectors or create custom policies to monitor in report-only mode and then progressively enforce blocks, redactions, encryptions, or transformations.