Jax
PulseAugur coverage of Jax — every cluster mentioning Jax across labs, papers, and developer communities, ranked by signal.
- used by Flax 90%
- developed alphaXiv 90%
- used by reinforcement learning 90%
- used by Fortran 90%
- used by graphics processing unit 70%
- used by NumPy 70%
- used by vLLM 70%
- used by central processing unit 70%
- used by Orbax Distributed Checkpointing With Jax 70%
- used by NVIDIA H100 70%
- used by Hugging Face Transformers 70%
- used by AI accelerator 70%
10 day(s) with sentiment data
-
New JAX framework vidax optimizes video generation for Cloud TPUs
Researchers have developed vidax, a new open-source framework built with JAX and Flax designed to optimize video generative models for Cloud TPU pods. This engine includes a zero-copy translator for PyTorch weights, ena…
-
New research explores advanced training for cooperative drone swarms · 2 papers
Two new research papers explore advanced methods for training and deploying cooperative drone swarms. The first paper introduces AeroWeaver, an embodied-agent harness that connects large language model decisions to exec…
-
New method calibrates ocean models with uncertainty quantification
Researchers have developed a new method for calibrating single-column ocean models using simulation-based inference (SBI). This approach addresses the limitation of previous methods by quantifying the uncertainty associ…
-
New JAX library accelerates AI ad hoc teamwork research
Researchers have developed JaxAHT, a new open-source library built with JAX, designed to accelerate and standardize the research process for ad hoc teamwork (AHT) in artificial intelligence. This library aims to overcom…
-
Sakana AI proposes layer-local training method for 1000-layer networks
Researchers at Sakana AI have developed a novel training method called Augmented Lagrangian Predictive Coding (PC-ALM), which offers a layer-local alternative to traditional backpropagation. This new approach allows for…
-
JAX3D enables hierarchical NeRF for advanced 3D rendering and reconstruction
Researchers have developed a method for creating hierarchical Neural Radiance Fields (NeRFs) using JAX and the jax3d library. This approach enables volumetric rendering, novel-view synthesis, and 3D reconstruction. The …
-
New agent HORIZON enhances multi-agent navigation with hierarchical belief modeling
Researchers have developed HORIZON, a hierarchical agent designed for the Lux AI Season 3 competition, which demands adaptation in partially observable multi-agent navigation scenarios. This agent employs a multi-facete…
-
China launches AI4S scientific computing platform with typhoon tracking system
Tai Chu Yuan Qi has launched its AI4S computing platform, designed for scientific research in fields like meteorology and quantum mechanics. The platform utilizes proprietary heterogeneous AI chips and supports various …
-
New framework uses AI to optimize chemical transport processes
Researchers have developed a new differentiable hybrid modeling framework designed to improve the accuracy and optimization of chemical transport processes. This framework integrates a JAX finite volume solver with neur…
-
GRADSOLVE library accelerates ODE gradient computation on GPUs
A new open-source JAX library called GRADSOLVE has been developed to accelerate the computation of exact gradients for ordinary differential equation (ODE) ensembles on NVIDIA GPUs. This library addresses a performance …
-
New method extracts biomechanical pose from 3D body models
Researchers have developed a new method to extract biomechanically accurate joint angles from single RGB images, addressing a limitation in current 3D body recovery techniques. This approach extends the SAM 3D Body foun…
-
New EMR-HyperNEAT method accelerates neuroevolution with tensorization
Researchers have developed EMR-HyperNEAT, a novel approach to neuroevolution that significantly accelerates the process of evolving large-scale neural network substrates. This new method overcomes limitations of previou…
-
Google Cloud enables vLLM on TPUs for Qwen3 long-context embeddings
Google Cloud has introduced native vLLM support for Tensor Processing Units (TPUs) optimized for embedding inference, aiming for production retrieval systems. The update focuses on enhancing long-context and multimodal …
-
New JAX implementation tackles ES-HyperNEAT scaling bottleneck
Researchers have developed JAX-ESHN, a new JAX-based implementation designed to parallelize ES-HyperNEAT on GPUs. This implementation addresses the quadtree bottleneck that previously limited scalability in coordinate-b…
-
Anyscale boosts Ray Data with GPU-native operators for AI workloads
Anyscale has enhanced its Ray Data engine with GPU-native operators, collaborating with NVIDIA to integrate cuDF and RapidsMPF. These updates allow data processing tasks to execute directly on GPUs, offering significant…
-
New research explores neural emulators for partial differential equations · 2 sources tracked
Two new research papers explore the relationship between numerical solvers for partial differential equations (PDEs) and neural emulators. The first paper, "From Numerical Simulators of PDEs to Neural Emulators and Back…
-
Gemma 4 E2B deployment issues highlight TPU serving stack limitations
A developer encountered issues deploying Google's Gemma 4 E2B model with quantized checkpoints on a vLLM serving stack using Google Cloud TPUs. Both the int4 and dequantized QAT variants failed to load due to discrepanc…
-
New research reveals surprising generalization in tabular foundation models
New research explores the surprising generalization capabilities of tabular foundation models (TFMs), suggesting that strong transfer learning can be achieved even from self-supervised pre-training on a single real tabl…
-
Google's TPU v6e-1 offers memory upgrade but at a higher cost
A technical analysis reveals that Google's new Cloud TPU v6e-1 (Trillium) offers a performance increase over the v5e-1, but its higher cost makes it less cost-effective for certain workloads. The v6e-1 provides double t…
-
Google details DiffusionGemma text-to-image model in technical report
Google has released a technical report detailing DiffusionGemma, a new text-to-image model. The report outlines the model's architecture, which incorporates elements like U-Net and LoRA+, and discusses its performance u…