PyTorch

We’re proud to announce the inaugural #PyTorchDayJapan, hosted by PyTorch Foundation, Hugging Face, IBM, and Mitsubishi Electric, December 10 in Tokyo. #PyTorchDayJapan will bring together PyTorch enthusiasts, machine learning engineers, AI researchers, and industry professionals for technical talks and discussions spanning training, inference, responsible AI, and more. Dec 10 is dedicated to open source AI and the impact of PyTorch Foundation projects, including PyTorch, vLLM, DeepSpeed.ai, Ray, Helion, and Safetensors. 🔗 Learn more: https://hubs.la/Q04vtQxJ0

In less than two weeks, Shanghai will welcome open source technologists from across China, Asia, and the global community. Engineers, maintainers, researchers, and technology leaders from Alibaba, Ant Group, ByteDance, Huawei, Tencent, NVIDIA, Google, Microsoft, Meta, Amazon, Apple, IBM, Intel, Red Hat, Samsung SDS, ZTE, and many more organizations are already registered for KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China. Join them September 7-9 to exchange ideas, build new connections, and explore the technologies shaping cloud native infrastructure, OpenInfra, machine learning, and open source AI. See who’s attending: https://bit.ly/45OMC9i Register and meet the community in Shanghai: https://bit.ly/3NG8hv1 #KubeCon #CloudNativeCon #OpenInfraSummit #PyTorchCon #Kubernetes #K8s #CloudNative #OpenInfra #PyTorch #PyTorchFoundation #FutureOfAI #AI #GenAI #MachineLearning #ML #DeepLearning #OpenSource #OpenSourceSoftware #OpenSourceDevelopment #OpenSourceCommunity #OSS #CNCF #CloudNativeComputingFoundation #events #linux #software

What began as experimentation with a Python framework evolved into PyTorch, the underlying framework for developers in the fast-moving AI space. At #PyTorchCon NA 2025, Jana van Greunen (Meta) reflected on that evolution and the scale of the AI work PyTorch powers today. #PyTorchCon North America returns October 20–21, 2026, in San Jose. 🔗 Register by September 4 to save on your ticket: https://hubs.la/Q04tBgv_0

The keynote lineup for PyTorch Conference North America is out and gives us a few hints about where the ecosystem is going next: a stronger core, greater support of diverse hardware, more efficient training and inference, and new approaches to building intelligent systems. We’ll begin with Alban Desmaison and Edward Yang presenting the latest PyTorch Updates. From there, the program explores how PyTorch is expanding across an increasingly diverse computing landscape. Bill Jia from Google will present “Workload Fungibility in the Age of Agents.” Maen S. of Amazon Web Services (AWS) will share “Trainium’s Journey to Native PyTorch,” and Jana van Greunen will present “PyTorch at Meta: From Internal Workloads to Multi-Silicon Ecosystem.” At the scale of frontier models, Zhipeng Wang, PhD will present “Scaling Large Frontier Models Training with DeepSpeed.” Yiwei Song will discuss an “AI Model Customization Platform,” while Mark Saroufim will return to a foundational challenge with “Linear Algebra for the Age of Research.” The program also looks beyond scale alone. Sara Hooker will present “Beyond Brute Force: The Era of Adaptive Intelligence,” and Joelle Pineau will explore how openness can support progress with “Open Science and AI Foundation Models: Allies in the Pursuit of Knowledge.” We’ll also welcome Brian Stevens discussing llm-d, Mazin Gilbert, MBA, Ph.D. Mazin Gilbert of the Agentic AI Foundation, Ujval Kapasi of NVIDIA, Chris Lattner of Modular, a Qualcomm company, and Simon Mo bringing insights on open source inference (& vLLM) to the keynote stage. I’m excited to open the conference and bring this group together with the PyTorch community in San Jose. Then we'll head to the wrap party with De La Soul! Join us October 20–21 at PyTorch Conference North America. Explore the keynote lineup: https://lnkd.in/g7JspPzh Register: https://lnkd.in/gvmTujpG #PyTorch #PyTorchCon #OpenSourceAI #vLLM

I’m excited to share that I’ll be speaking at #PyTorchCon North America, October 20–21 in San Jose! I’ll be talking about Elastic Expert Parallelism in vLLM. Elastic EP lets running MoE deployments add or remove GPUs as the workload changes, while continuing to serve traffic with minimal downtime. That same runtime reconfiguration is also a key building block for fault-tolerant serving, with NIXL EP supporting the communication side. Whether you’re a contributor, a user, or simply as excited about inference as I am, come join us at PyTorch Conference in San Jose! ➡️ Register: https://bit.ly/4pd5ffR 🔥 Schedule: https://bit.ly/4fBCG8P

How can teams transform powerful AI models into systems built for their specific needs? Meet Yiwei Song, Senior Director of AI Model Cycle at Crusoe and a keynote speaker at PyTorch Conference North America. In his session, “AI Model Customization Platform,” Yiwei will bring model customization to the keynote stage and explore an increasingly important part of the AI development lifecycle. Join the PyTorch community in San Jose to hear his perspective. 📅 October 20-21 📍 San Jose, CA Meet Yiwei: https://lnkd.in/grKb8wHM Register & join us! https://lnkd.in/gKzzmr_7 #PyTorchCon #PyTorch #PyTorchFoundation #FutureOfAI #AI #GenAI #MachineLearning #ML #DeepLearning #OpenSource #OpenSourceSoftware #OpenSourceDevelopment #OpenSourceCommunity #OSS #LinuxFoundation #events #linux

TRANSIT (TRANsparent Scale-In for multi-node Training) is a runtime that makes unified virtual memory practical for large-scale LLM training, enabling models to train on up to 50% fewer GPUs without requiring modification to existing PyTorch training code. Hyungyo Kim, Ph.D. Candidate in Electrical and Computer Engineering, University of Illinois at Urbana-Champaign will present on TRANSIT as a poster at PyTorch Conference North America with Apoorve Mohan from IBM Research. Learn how to train more with less at PyTorch Conference North America: https://hubs.la/Q04v4SL60

Attendees of PyTorch Conference Europe earlier this year share what it's like to attend #PyTorchCon. PyTorch Conference North America will be here before you know it. Don't miss out: https://lnkd.in/geC5_z9a

Leverage PyTorch-native building blocks for constructing GPU-accelerated materials simulation workflows with coding agents. NVIDIA follows an end-to-end NVIDIA ALCHEMI Toolkit workflow: what the researcher starts with, how they prompt the agent, what code and simulation pipeline it produces, and how the results are validated on NVIDIA H200 GPUs. Read more: https://lnkd.in/edSMJW63

Join Ujval Kapasi, VP of AI & HPC Frameworks & Libraries at NVIDIA, as he shares insights into the technologies powering the next generation of AI development. 📅 October 20-21 📍 San Jose, CA Schedule: https://lnkd.in/gDkRGz78 Register: https://lnkd.in/gKzzmr_7 #PyTorchCon #PyTorch #PyTorchFoundation #FutureOfAI #AI #GenAI #MachineLearning #ML #DeepLearning #OpenSource #OpenSourceSoftware #OpenSourceDevelopment #OpenSourceCommunity #OSS #LinuxFoundation #events #linux