Blog

Blog Community

Hybrid Models Meet SGLang: More than Full Attention

Introduction Hybrid models that combine the capabilities of full attention layers with alternatives—such as Mamba…

SGLang TeamDecember 3, 2025

Blog

Efficient MoE Pre-training at Scale on 1K AMD GPUs with TorchTitan

Training massive Mixture-of-Experts (MoE) models like DeepSeek-V3 and Llama 4-Scout efficiently is one of the…

AMD Contributors: Liz Li, Yanyuan Qin, Yuankai Chen, Xinyu Kang, Xiaobo Chen, Zhen Huang, Shekhar Pandey, Zhenyu Gu, Andy Luo, Meta Contributors: Matthias Reso, Hamid Shojanazeri, Tianyu Liu, Jiani Wang, Howard Huang, Wei Feng, Special Thanks: Guru MP, Yao Fu, Nick Ni, Emad Barsoum, Ramine Roane, and the TensorWave team for providing MI325 clusterDecember 1, 2025

Blog Community

The Future of Inference: PyTorch ATX Event

On September 17, 2025, PyTorch ATX partnered with the vLLM community and Red Hat to…

Jason Meaux, ATX PyTorch leader and Stephen Watt, PyTorch Ambassador, Red HatNovember 26, 2025

Blog

OpenReg: A Self-Contained PyTorch Accelerator Simulator

Introduction The PyTorch community is actively working to build a growing ecosystem of specialized accelerators…

Jiahao Chen (Huawei) & Jiawei Li (Huawei) & Zesheng Zong (Huawei)November 21, 2025

Announcements Blog

PINA Joins the PyTorch Ecosystem: A Unified Framework for Scientific Machine Learning

Scientific Machine Learning (SciML) is reshaping how complex physical and scientific systems are modelled and…

Giovanni Canali, Dario Coscia, Nicola Demo, Filippo Olivo; PINA Team.November 18, 2025

Blog Community

Beyond Quantization: Bringing Sparse Inference to PyTorch

As developers, we all know the story: Large Language Models (LLMs) are revolutionary, but their…

Kira Selby & Varun Khare (NimbleEdge)November 13, 2025

Announcements Blog

Accelerating the Future of Open Source AI: PyTorch Conference 2025 Recap

PyTorch Conference 2025 brought together 3,432 developers, researchers, and innovators from 1,026 organizations across the…

PyTorch FoundationNovember 12, 2025

Blog

KernelFalcon: Autonomous GPU Kernel Generation via Deep Agents

Summary We introduce KernelFalcon, a deep agent architecture for generating GPU kernels that combines hierarchical…

Laura Wang and the PyTorch Team at MetaNovember 5, 2025

Blog

Hybrid Models as First-Class Citizens in vLLM

Introduction and Agenda Large language models are now running into the scaling limits of attention.…

vLLM Team at IBMNovember 5, 2025

Announcements Blog

Congratulations to the 2025 PyTorch Contributor Awardees

We are pleased to announce the awardees and nominees of our third annual 2025 Contributor…

PyTorch FoundationOctober 31, 2025

Announcements Blog

LMCache Joins the PyTorch Ecosystem: Accelerating the Future of AI, One Cache at a Time

We’re delighted to announce that LMCache has officially become a PyTorch Ecosystem project, joining the…

Nick Barcet, LMCacheOctober 30, 2025

Announcements Blog

PyTorch Foundation Welcomes Ray to Deliver a Unified Open Source AI Compute Stack

Ray joins leading open source AI projects including PyTorch and vLLM to minimize AI computing…

PyTorch FoundationOctober 22, 2025

Announcements Blog

Dell Technologies Joins the PyTorch Foundation as a Premier Member

The PyTorch Foundation, a community-driven hub supporting the open source PyTorch framework and a broader…

PyTorch FoundationOctober 22, 2025

Blog

Monarch + Lightning AI: Unlocking New Possibilities in Distributed Training

Introduction: Empowering the Next Generation of AI Builders We are excited to announce a partnership…

PyTorch Team at Meta: Alireza Shamsoshoara, Lucas Pasqualin, Peng Zhang, Hamid Shojanazeri, Ahmad Sharif, Kiuk Chung, Lightning AI: Lightning: Luca AntigaOctober 22, 2025

Blog

torchcomms: a modern PyTorch communications API

Introduction Torchcomms is a new experimental, lightweight communication API intended for use with PyTorch Distributed…

Team torchcomms at MetaOctober 22, 2025

Blog

Helion: A High-Level DSL for Performant and Portable ML Kernels

Introduction to Helion In modern machine learning, the demand for high-performance computation has led to…

PyTorch Team at MetaOctober 22, 2025

Blog

Introducing ExecuTorch 1.0: Powering the next generation of edge AI

TLDR ExecuTorch enables seamless, production-ready deployment of PyTorch models directly to edge devices (mobile, embedded,…

PyTorch Team at MetaOctober 22, 2025

Blog

Introducing PyTorch Monarch

We now live in a world where ML workflows (pre-training, post training, etc) are heterogeneous,…

The PyTorch Team at MetaOctober 22, 2025

Blog

Introducing torchforge – a PyTorch native library for scalable RL post-training and agentic development

In this post, we announce torchforge: A PyTorch-native agentic RL library that lets you focus…

The PyTorch Team at MetaOctober 22, 2025

Blog

Enabling vLLM V1 on AMD GPUs With Triton

What is vLLM V1? In January 2025, the vLLM team announced the alpha release of…

vLLM Team at IBM Research, vLLM Team at Red Hat, and vLLM Team at AMDOctober 21, 2025

Hybrid Models Meet SGLang: More than Full Attention

Efficient MoE Pre-training at Scale on 1K AMD GPUs with TorchTitan

The Future of Inference: PyTorch ATX Event

OpenReg: A Self-Contained PyTorch Accelerator Simulator

PINA Joins the PyTorch Ecosystem: A Unified Framework for Scientific Machine Learning

Beyond Quantization: Bringing Sparse Inference to PyTorch

Accelerating the Future of Open Source AI: PyTorch Conference 2025 Recap

KernelFalcon: Autonomous GPU Kernel Generation via Deep Agents

Hybrid Models as First-Class Citizens in vLLM

Congratulations to the 2025 PyTorch Contributor Awardees

LMCache Joins the PyTorch Ecosystem: Accelerating the Future of AI, One Cache at a Time

PyTorch Foundation Welcomes Ray to Deliver a Unified Open Source AI Compute Stack

Dell Technologies Joins the PyTorch Foundation as a Premier Member

Monarch + Lightning AI: Unlocking New Possibilities in Distributed Training

torchcomms: a modern PyTorch communications API

Helion: A High-Level DSL for Performant and Portable ML Kernels

Introducing ExecuTorch 1.0: Powering the next generation of edge AI

Introducing PyTorch Monarch

Introducing torchforge – a PyTorch native library for scalable RL post-training and agentic development

Enabling vLLM V1 on AMD GPUs With Triton

Docs

Tutorials

Resources

Stay in touch for updates, event info, and the latest news