• LoRA — A cheap way to fine-tune a model by training a small add-on layer instead of changing all of the model's weights.

Reinforcement learning with verifiable rewards on an open model is now a hosted, roughly five-minute setup producing a downloadable LoRA adapter — the infrastructure barrier that kept most teams from post-training is largely gone.

• articleEnhancing Goodput in Large-Scale LLM Training with Nonuniform Tensor Parallelism
• articleRestore LLM Inference Capacity in Seconds with Shadow Engine Recovery in NVIDIA Dynamo
• articleWhere Security Fits in an AI Agent Stack
• articleEvaluating AI Agent Skill Performance with NVIDIA SkillEvaluator

• articleDeveloping Nemotron 3.5 Lightning NVFP4 with QAD Using NVIDIA Model OptimizerTanya Lenz
• blogNVIDIA Nemotron 3.5 Lightning and NeMo Switchyard Deliver Faster, Smarter, More Efficient Agentic AIKari Briski
• videoNemotron 3.5 Lightning First Test – NVIDIA’s NEWEST Open Model!Bijan Bowen

Checking sign-in…

Loading comments…