PyTorch

The Startup Showcase is happening now at PyTorch Conference 2025, where AI startups are pitching live to top VCs. Startups presenting: Graphsignal, Runhouse, NimbleEdge, XOR, Snapshot AI, Node One, BrightOnLABS, Okahu AI, Curavoice, Seer Systems, KnowNext, and Backfield.ai. Judges: Chappy Asel (The GenAI Collective), Irving Hsu (Mayfield), Radhika Malik (Dell Technologies Capital), Kevin Crosby (GitHub), Denise T. (Gradient), and Simon Tiu (Vertex Ventures US). Sponsored by Dell Technologies Capital and Mayfield Stay tuned for more! 🔗 Learn more: https://hubs.la/Q03PCWKQ0 #PyTorchCon #PyTorch #StartupShowcase

The Reinforcement Learning track at PyTorch Conference 2025 highlights the latest research in RL algorithms and applications, with sessions on long-tail/MoE optimization and novel approaches to training. Who should attend: researchers and engineers exploring reinforcement learning. What you’ll learn: RL techniques in PyTorch for robotics, gaming, and optimization. Sessions include Chenyang Zhao (UCLA) on optimizing long-tail and MoE challenges in RL with SGLang, and Daniel Han (Unsloth) on maximizing luck in reinforcement learning. 📅 October 23, 2025 in San Francisco 🔗 Explore the sessions: https://hubs.la/Q03NC-lL0 #PyTorchCon #PyTorch #OpenSourceAI #AIInfrastructure

vLLM V1 is now fully supported on AMD GPUs! New blog from vLLM Team at IBM Research, vLLM Team at Red Hat, and vLLM Team at AMD: Learn how the teams at IBM, AMD and Red Hat built an optimized attention backend for vLLM V1 using Triton kernels. In this deep technical blog, we describe the optimizations we performed to ensure that the Triton backend achieves state-of-the-art performance on AMD. Check it out if you are interested in knowing more about the art of writing lighting fast Triton kernels for AI applications! 🔗 Read here: https://hubs.la/Q03PC6kh0 #vLLM #PyTorch #OpenSourceAI

The PyTorch in Academia track at PyTorch Conference 2025 highlights how PyTorch is advancing research and education across disciplines, from space science to climate modeling and beyond. Who should attend: researchers, students, and educators. What you’ll learn: cutting-edge research, teaching practices, and academic collaboration with PyTorch. Speakers include Vit Ruzicka (NASA JPL), Matt Archer (University of Cambridge), Rishit Dagli (University of Toronto), Shivay Lamba (Couchbase), Abin Shakya (Louisiana State University), and Vaclav Knapp (Smichov Technical High School). 📅 October 23, 2025 in San Francisco 🔗 Explore the sessions: https://hubs.la/Q03NCZYD0 #PyTorchCon #PyTorch #OpenSourceAI #AIInfrastructure

Live from the AI Infra Summit, co-located with PyTorch Conference 2025: Tianqi Chen of NVIDIA and Carnegie Mellon University shares insights on building shared foundations for modern machine learning systems. In Building ML Systems Foundations in the Age of AI, Chen explored interoperability across compilers, libraries, DSLs, and frameworks, and how AI can strengthen ML systems production and unify workloads across edge and cloud. 🔗 Full summit schedule: https://hubs.la/Q03PBqWW0 #AIInfraSummit #PyTorchCon #PyTorch #OpenSourceAI

Our LLMs track at PyTorch Conference 2025 dives into the latest breakthroughs in large language models, with sessions on architectures, training optimizations, and real-world applications. Who should attend: researchers, ML engineers, developers, and product teams. What you’ll learn: cutting-edge architectures, optimizations for scale, and applied use cases for LLMs. Speakers include Yineng Zhang (LMSYS Org), Evan Smothers (Meta), Sharbani R. (Arm), Chip Huyen (Tep Studio), Harry Kim (NVIDIA), Laikh Tewari (NVIDIA), Boyuan Feng (Meta), Xilun Wu (Meta), Chien-Chin Huang (Meta), Mert Toslali (IBM Research), Yu Chin Fabian Lim (IBM Research), Hongpeng Guo (ByteDance Seed), Ziheng Jiang (ByteDance Seed), Junchen Jiang (University of Chicago), Moein Khazraee (NVIDIA), Niels Bantilan (Union.ai) Sebastian Raschka, PhD, Yuxin Wu (Moonshot AI), and Philip Bontrager (Meta). 📅 October 22–23, 2025 in San Francisco 🔗 Explore the sessions: https://hubs.la/Q03NCSqR0 #PyTorchCon #PyTorch #OpenSourceAI #AIInfrastructure

Our Foundation Projects and Ecosystem track at PyTorch Conference 2025 delivers updates and roadmaps from PyTorch, vLLM, and DeepSpeed, plus integration insights from Hugging Face, Anyscale, Amazon, Intel, BAAI, and more. Who should attend: PyTorch users, contributors, and ecosystem developers. What you’ll learn: project roadmaps, integration strategies, and best practices for working across the PyTorch ecosystem. Sessions will include speakers Robert Nishihara (Anyscale), Simon Mo (vLLM), Yonghua Lin (BAAI), Lysandre Debut (Hugging Face), Arthur Zucker (Hugging Face), 🍃 Tristan Rice (Meta), Yahav Biran (Amazon), Fan Zhou (Intel Corporation), Eikan Wang (Intel), Daniel Vega-Myhre (Meta), Ke Wen (Meta), Natalia Gimelshein (Meta), Tunji Ruwase (Snowflake), and Aurick Qiao (Snowflake). 📅 October 22–23, 2025, San Francisco 🔗 Explore the sessions: https://hubs.la/Q03NCR190 #PyTorchCon #PyTorch #OpenSourceAI #AIInfrastructure

Happening now at PyTorch Conference 2025 — the PyTorch Associate Training. This in-person, instructor-led program equips participants with the foundational and practical skills aligned with the upcoming PyTorch Certified Associate exam. Through hands-on labs, real-world projects, and guided instruction, attendees are deepening their understanding of the #PyTorch ecosystem while preparing for certification. 🔗 Learn more about the training: https://hubs.la/Q03PxRq70 #PyTorchCon #PyTorch #OpenSourceAI

NVIDIA is on the ground at Open Source AI Week and #PyTorchCon, celebrating the innovation, collaboration, and community driving open source AI forward. PyTorch is powering AI worldwide—with over 2 million downloads in a single day and 65 million last month. Built on the NVIDIA CUDA platform, PyTorch continues to benefit from NVIDIA’s deep collaboration, including the addition of Python as a first-class CUDA language to make GPU acceleration more accessible than ever. 🔗 Read more: https://hubs.la/Q03PxwMV0 #PyTorch #OpenSourceAI #PyTorchCon

Noam Brown of OpenAI discusses reasoning, planning, & inference time scaling with Joseph Spisak of Meta at the Measuring Intelligence Summit, co-located with PyTorch Conference. #PyTorchCon #opensource #AI