Manage Kubernetes Clusters for GPU Infrastructure with NVIDIA Kubernetes Engine
NVIDIA is announcing two open source projects for building and operating managed Kubernetes services on GPU infrastructure—NVIDIA Kubernetes Engine (NVKE) and...
en
Hardware y chips
633
74
Start and Restore GPU Workloads in Seconds with NVIDIA Snapshot
NVIDIA Snapshot, a new open-source component of the NVIDIA DSX AI Factory Platform, makes initialized GPU workloads reusable across inference...
en
Few-Step Streaming Video Generation via Condition Shifting
Few-step sampling reduces the cost of streaming video generation. Distribution Matching Distillation (DMD) supervises few-step generators with a separate fake-score network, which adds training memory and computation and must track the evolving generator. We extend APEX to streaming video by using condition shifting to estimate the velocity field of the generated distribution within the generator. The original text condition selects generation, while a shifted condition selects fake estimation using the same parameters. This removes the independent fake network while retaining S2-Avatar's frozen reference teacher, Self-Replay Forcing, and perceptual regularization. We evaluate APEX on 180-second videos using three backbone evaluations per block and a one-step spatial Refiner, with a 70-case human study and a 320-case StreamAV-Bench comparison. Compared with S2-Avatar, APEX achieves higher video instruction fulfillment on the Progressive track (3.856 versus 3.831) and higher visual aesthetics across both tracks (0.641898 versus 0.632606). Human raters favor APEX most strongly on instruction following, with +30.0% net preference, followed by consistency and video quality at +4.3% each. Motion quality favors S2-Avatar, audiovisual synchronization is predominantly tied, and usability is unchanged. Separate measurements under matched training settings show 4.4% lower peak allocated GPU memory per GPU and 9.2% shorter mean iteration time than DMD.
en
Understanding Mixture-of-Experts Language Models Beyond Expert Boundaries
In Mixture-of-Experts (MoE) models, the nature of expert specialization remains controversial. Reports of domain-associated expert preferences contrast with analyses attributing routing primarily to token identity or syntactic patterns. To clarify these differing accounts, we look beyond expert-level routing to investigate the neuron-level organization of domain-specialized computation using Routing-Aware Magnitude Profiling (RAMP), a measure developed for this analysis. Using mixed-domain calibration data, we identify domain-specialized neuron sets with 58-74% overlap across multi-source samples of the same domain, compared with 15-24% across domains. To understand how this organization evolves during training, we trace neuron importance throughout pre-training and post-training across three fully open MoE families. Neuron importance continues to reorganize between pre-training checkpoints despite comparatively stable expert-level importance rankings, while post-training largely preserves neuron rankings. Building on these neuron-level perspectives, we further introduce RAMP-based Pruning (RAMPr) to select neurons across experts for balanced multi-domain capability preservation. At 75% neuron retention, RAMPr preserves average 91-98% of original full performance across six models. At 50% retention, our pruning method can reduce resident GPU memory by 36-43% and improves batched prefill throughput by 17-39% effectively. Our code is available at https://github.com/ydyjya/Neuron-Level-Specialization-in-MoE
en
Mobile AI Factory: Building Specialized Language Models for Your Mobile Device
Small language models (SLMs) running on mobile devices can keep user data private and work without connectivity. But they perform well only when fine-tuned for the target task and sized for the target device, a process that today requires repeated expert tuning. We present {Mobile AI Factory}, an agentic system that builds a fine-tuned, quantized SLM for a given task and phone. Given a task specification and a target phone, an orchestrator LLM chooses whether to tune hyperparameters, mine real data, or generate synthetic examples, while fixed rules roll back changes that hurt accuracy. The factory searches from small to large models across base models and quantizations, adapts its accuracy goal as it searches, and uses real measurements of the target phone across inference engines and backends. We evaluate Mobile AI Factory on ten tasks and four Android phones. On an unseen testset, its on-device models outperform two five-shot frontier models on five tasks, by up to 0.34, and match them on three more. Its search comes within 0.03 of exhaustively training all 27 model variants on an entity extraction task, while using 19less GPU time. Phone measurements also matter: the best configuration changes across phones. Without phone measurements, Claude Opus 5.5, given the full hardware and model specifications, misses the app's requirements in 25\% of easy and moderate cases and in 60\% of hard ones. Code is available at: https://github.com/evanly-gh/SLM_Factory.
en
Placing LLM Lookup Tables by Per-Layer Prefetch Deadlines
DeepSeek-V4.1's 203 GB of Engram n-gram tables sit in host DRAM today, taking 190 of the 223 GiB a 4-GPU serving process uses. Row indices depend only on input tokens, so reads can start with each step and overlap earlier layers. V4.1's two Engram layers have windows that differ by 30x (0.3 vs. 9.3 ms). We show that this deadline, together with session reuse and contention on a shared device, decides which tier each table can use and which layers need cache, and we give a model that predicts the resulting slowdown from measured read times. Agent and chat traffic reuse rows within sessions, and static hot sets transfer poorly. We build SlackTier, a table tier for SGLang that issues all lookups at step start. It reads the short-window layer first and splits a small GPU row cache across layers. The path runs inside the decode CUDA graph and is bit-exact with the original gather. Layer 14's table can go to NVMe uncached with no measurable decode cost. With both tables on four NVMe disks, mean time per output token on agent and chat replays rises at most 1.4% over unmodified SGLang, and host memory falls from 212 to 34 GiB. Both tables need no cache when read over RDMA (remote direct memory access) from another node's memory; one copy serves two instances. On a KV-constrained agent workload, using the freed memory as host KV cache at the same memory budget raises median throughput by 36% and cuts median p90 time to first token from 38 to 24 s.
en
SteerDistill: Quantizing LLM Agents via Closed-Loop Teacher Steering
Low-bit quantization reduces the inference cost of LLM agents but can compromise task success. Examining failed rollouts, we find that quantized agents overlook task requirements or environmental conditions, producing environment transitions that diverge from successful full-precision references. Correcting these behavioral divergences remains difficult: offline data may miss the student's failure contexts, while its own rollouts may provide few correct responses in those contexts. We introduce SteerDistill, a quantization-aware distillation method in which a full-precision teacher guides the student to revise failed responses and iteratively refines its guidance using feedback from student execution. This closed-loop teacher steering produces locally verified behavioral corrections. We internalize these corrections by training the student without guidance to match the steered student's predictions at corrected turns and the full-precision teacher's at other turns, enabling unassisted inference. Across five agentic benchmarks and three models, SteerDistill exceeds the strongest quantized baseline by 1.67--2.18 points in average score under W4A16 and stays within one point of BF16. On τ^2-Bench, it outperforms QeRL by 1.84 percentage points in success rate using 67% fewer training GPU hours and produces shorter successful trajectories than the quantized baselines.
en
What could China get from Jensen Huang?
It's difficult to overstate Nvidia's power in today's AI landscape. The company's chips are the bedrock upon which advanced AI technology in the US is…
en
How a Solo Dev Built an AI for Dads, No GPU, No Team, Just $5
No GPU. No team. Just $5. This is how one solo developer used Runpod Serverless to build and deploy a working AI product, "AI for Dads", without writing.
en
Introducing Claude Fable 5.2 on AWS
Claude Fable 5.2 and Claude Mythos 5.2 are now available on Amazon Bedrock and Claude Platform on AWS. Learn what's new in both models, why Fable 5.2 matters for your technical workloads, and how to get started building with it on Amazon Bedrock.
en
NVIDIA Nemotron 3.5 Super VL is now available in Amazon SageMaker JumpStart
NVIDIA Nemotron 3.5 Super VL, a 120B-parameter multimodal vision-language model, is now available in Amazon SageMaker JumpStart. Deploy the NVFP4 or BF16 variant in a few steps, then run text, image, and video inference for document intelligence, computer-use agents, and multimodal agentic workloads.
en
Kolmogorov-Chapman Maps: Trajectory-Free Learning for Simulating Molecular Dynamics
Stochastic differential equations (SDEs) describe stochastic processes across a wide range of scientific domains, including physics, biology, and finance. Although individual numerical integration steps are inexpensive, observing meaningful events or transitions can require millions of sequential steps, resulting in substantial wall-clock simulation time and poor hardware utilization. Deep learning offers an alternative by amortizing many integration steps into a single neural network forward pass, potentially enabling much faster simulation. However, existing supervised approaches typically require long simulated trajectory data, which restricts their applicability. In this work, we introduce Kolmogorov-Chapman Maps (KC-Maps), a method for learning the transition kernel of an SDE that relies only on the trajectory-free marginal samples together with the known drift and diffusion coefficients. Given an initial time and state of a process, KC-Maps models its distribution integrated to any given time, corresponding to the weak solution of the SDE. To recover the correct dynamics, KC-Maps satisfies two key properties: (i) it satisfies the boundary conditions, which we incorporate directly into the neural network architecture, and (ii) it respects the Kolmogorov--Chapman equation, which provides an efficient training objective. We empirically demonstrate that KC-Maps learns to integrate the dynamics of peptide molecular systems and outperforms classical molecular dynamics in terms of GPU-hours.
en
How CrossnoKaye grounds physical AI with Amazon Bedrock AgentCore
CrossnoKaye brought physical AI to heavy industry with Amazon Bedrock AgentCore. Its ATLAS platform and AURA agentic system ground AI reasoning in live plant data and run verifiable Python in isolated AgentCore sandboxes, giving plant engineers checkable, math-backed answers instead of model guesses.
en
2D Gum-Net: Unsupervised Dense Elastic Image Registration via Symmetric Correlation and Spectral Pooling
Dense non-rigid registration, the recovery of a pixel-wise deformation field that warps one image into geometric agreement with another, is fundamental across medical imaging, structural biology, and correspondence estimation, yet learning-based approaches face a severe bottleneck: ground-truth deformation fields are notoriously difficult to obtain for real data. We present 2D Gum-Net, an end-to-end architecture that predicts dense elastic deformation fields fully unsupervised. We transfer the Geometric Unsupervised Matching Network (Gum-Net), originally developed for 3D subtomogram alignment in cryo-electron tomography, to 2D elastic registration: all three modules are redesigned in 2D, and the rigid transformation head is replaced by a Dense Spatial Transformer that regresses a sparse control-point grid, lifted to a full-resolution field Φ by bicubic interpolation. Downsampling uses DCT spectral pooling, implemented as frozen-basis matrix multiplications (50GPU speedup), preserving the fine, quasi-periodic structure that max and average pooling destroy. We evaluate on contact-based fingerprints, a challenging testbed for this task: locally self-similar ridge patterns make correspondence maximally ambiguous, skin pressed against a sensor generates authentic large-displacement elastic distortion, and minutiae act as precise landmarks through which an independent matcher externally verifies the topological correctness of predicted fields. Trained purely on synthetic impressions and evaluated zero-shot on real benchmarks, our method improves structural correlation by up to 15.7% on FVC2004 and raises genuine match scores under an independent matcher while leaving imposter scores essentially unchanged. This suggests that the predicted deformations restore true structure without fabricating correspondences.
en
Jensen Huang’s affair with alleged Chinese spy puts Nvidia’s China ambitions under scrutiny
Jensen Huang’s alleged affair with a suspected Chinese spy threatens to complicate Nvidia’s lobbying efforts to loosen U.S. chip export restrictions and maintain access to…
en
Large Language Models as a Universal Searcher for Open-World Automated Machine Learning
Developing effective machine learning models requires complex manual tuning, while current Automated Machine Learning (AutoML) frameworks remain domain-specific and isolate model architecture, training hyperparameter, and data ratio optimization. Here, we bridge this gap by proposing a universal, large language model (LLM)-based AutoML framework that jointly optimizes these configurations for the first time without the need for task-specific fine-tuning. We propose NCode, a unified numerical encoding scheme that translates diverse AutoML design choices into universal discrete sequences and effectively reframes the optimization process as a unified ranking task. Trained solely on optimization trajectories derived from a limited, single-domain vision task, the proposed LLM-based searcher demonstrates impressive cross-task, cross-modality and joint search-space optimization generalizability, achieving state-of-the-art performance across image classification, image segmentation, speaker verification and text-to-image benchmarks. Crucially, it exhibits superior scalability in the large-scale training and joint optimization of a billion-parameter Diffusion Transformer, requiring 896 GPU-days and attaining a state-of-the-art GenEval score of 0.91. By leveraging reinforcement learning to instill Chain-of-Thought reasoning, we provide a systematic way to interpret the LLM-based searcher's internal decision-making mechanisms, revealing its optimization strategy shifted from the early-stage diversity-driven exploration to the late-stage anchor exploitation. Our proposed framework provides a scalable, generalizable, and interpretable pathway toward general-purpose open-world machine learning optimization.
en
LogCDM: Interpretable Log Anomaly Detection via Behavioral Graph Drift and Adaptive Signal Fusion
Log-based anomaly detection is increasingly dominated by large pretrained models, from BERT-scale encoders to 8B-parameter language models, which require GPUs at inference time and are difficult to deploy on the monitoring nodes where logs are produced. We present LogCDM, a compact detector that scores each log session using three complementary signals: the next-event surprise of a small causal Transformer, the distance between a graph-attention embedding of the session's own transition graph and that of a normal reference graph, and a transition-drift statistic that we show is exactly a frequency-weighted chi-squared divergence between transition kernels. The divergence view yields an asymptotic null distribution, a length-standardized score, and a per-edge decomposition that explains each alarm. A sigmoid gate fuses the three signals; under a shared-covariance Gaussian model, this gate is the Bayes posterior, and its learned weights indicate which signal each dataset relies on. With 8.1-8.8M parameters and no GPU required at inference, LogCDM reaches F1 scores of 0.984 on HDFS, 0.870 on chronologically split BGL, and 0.850 on Thunderbird, within 0.013-0.12 of an 8.1B-parameter LLM detector evaluated under the same protocol and above all compact baselines on BGL. It scores a session in 5.5-10.5 ms on a CPU. Ablations show that the structural signals add 0.025-0.030 F1 on BGL and Thunderbird but almost nothing on HDFS, consistent with the learned fusion weights.
en
The future of AI hinges on… Jensen Huang’s sex life?
I would prefer that these people's sex lives were none of my business but I guess we have to talk about Jensen Huang and the…
en
Inference Validation for Regulated AI: Building a Decision Pipeline on NVIDIA AIPerf
Deploying large language models in regulated industries requires more than picking the fastest inference engine: it requires evidence that a configuration meets contractual and legal obligations before it serves a single customer request. We describe Runway, an internal validation harness built at MasterControl that builds a governed decision pipeline around NVIDIA AIPerf. The key insight is separating concerns: AIPerf provides the trusted measurement, while Runway provides the experiment design that comes before it and the compliance, cost, and audit decisions that come after. We describe the architecture and the storage model that makes caching and audit one mechanism. We report three findings from Runway sweeps: TensorRT-LLM and vLLM rankings invert with concurrency; a newer, pricier GPU instance can be the cheaper one per request; and speculative decoding behaves differently across engines under load, so engine configuration can outweigh engine choice.
en
To SIMD or Not to SIMD: Comparing RVV vs. CGRA SoCs for Non-Linear Functions in AI
Non-linear functions are an important class of operations in modern large language models. However, chip vendors building for AI have so far relied on traditional vector or scalar processing units for non-linear functions. The performance of these units has not kept pace with matrix multiplication units (MMUs), which can be found in both high-performance and edge-computing devices from almost every vendor. Reconfigurable units such as FPGAs and CGRAs have been proposed in the literature, but have not been systematically compared with traditional units backed by MMUs. In this paper, we build two SoCs: one using an RVV vector processor and the other using a latency-insensitive (elastic) CGRA, both coupled with an MMU and driven by a single-core CPU. We then propose mappings for common feed-forward networks on our SoCs and find that the CGRA reaches a maximum speedup in clock cycles of 7.12x over RVV on the activation stage of simple point-wise non-linear functions like SiLU, whereas the RVV reaches a maximum speedup of 9.8x over the CGRA on complex non-linear functions such as softmax. Over the enclosing feed-forward block the corresponding maxima are 3.16x and 4.24x.
en