PyTorch @PyTorch

At PyTorch Conference North America 2026 in San Jose, @RedHat engineers Jewel Muraledharan and Subin George Malana will present on Cross-Repository CI Relay (CRCR) to explore how downstream repositories can seamlessly integrate into upstream PyTorch CI. Their session will detail how CRCR enables a tiered onboarding model alongside real-time visibility for downstream repositories in HUD. Join us in San Jose on October 20 and 21 to learn how CRCR streamlines testing workflows across the ecosystem. Register today: hubs.la/Q04v4SL60 #PyTorchCon

At PyTorch Conference North America, explore how new optimizers like Muon and TorchJD work alongside tools like fastsafetensors and LMCache to drive multi-loss efficiency and faster LLM serving. Featured speakers include: @awscloud : Paulo Aragao @Arm : Thomas Cottenier @Google : Claudio Basile, Aleksey Vlasenko, Ankita Luthra, Trinadh Kotturu @huggingface : Ben Burtenshaw, Aritra Roy Gosthipaty, Suvaditya Mukherjee @Huawei : Jiahao Chen, Jiahao Tan @IBM : Takeshi Yoshimura, Prasanth Chatarasi, Bardia Mahjour, Viji Srinivasan, David Grove, Antoni Viros Martin, Avery Blanchard @intel : Panagiotis Kourdis, Tanima Dey @Meta : Felipe Mello, Jiani Wang, Will Constable, Natalia Gimelshein, Driss Guessous, Sanket Jayant Purandare, Aditya Venkataraman, Nicolas Hug, Scott Schneider, Edward Yang @nvidia : Ian Stenbit, Christine Cheng, Dylan Doblar Annapurna Labs (AWS subsidiary): Yahav Biran @argonne National Laboratory: Sam Foreman Beijing Academy of Artificial Intelligence: Yonghua Lin Bird of Paradise AI: Jennifer Wei @BytedanceTalk : Neiwen Ling @Harvard University: Vijay Janapa Reddi @RedHat Hat: Subin George, Jewel K M SimplexLab: Valérian Rey; Khush Patel @tensormesh : Kuntai Du University of Genoa: Andrea Mattia Garavagno University of Virginia: Tianle Zhong Join us in San Jose, October 20-21: hubs.la/Q04tBgv_0 Read the Open Research, Tooling & Optimization Sessions Guide here: bit.ly/4gXxn4r #PyTorchCon

Muon has attracted a lot of attention for fast convergence, but getting those optimizers to work in a real training stack is a different challenge. At PyTorch Conference North America, @JenniferWe17599 will walk through Muon, Dion, and Dion3 and look at what happens when you add the messy details: sharding, communication, schedulers, and everything else that can go wrong. If you are interested in optimizers and training dynamics, see you at PyTorchCon NA for Jen's talk on Beyond AdamW: A Practical PyTorch Walkthrough of Muon, Dion, and Orthogonalized Optimizer Variants. Get your ticket: hubs.la/Q04v4SL60

@Jaival_Suthar Recordings of the sessions will be made available after the conference on the PyTorch Foundation youtube channel.

Engineering teams from across the ecosystem are presenting major breakthroughs in torch.compile, custom kernel authoring, and scaling efficiency at the upcoming PyTorch Conference North America. Deepening performance across the compiler pipeline requires tackling bottlenecks at every layer, from front-end tracing and shape checking, to back-end execution and distributed orchestration. A few of the speakers sharing their insights include: @AMD: Liz Li, Prachi Gupta @huggingface: Sayak Paul @Huawei: Yun Zhao; Haonan Zhang @IBM, @IBMResearch: Olivier Tardieu, Matthew Arnold, Burkhard Ringlein @Intel: Xiaogang Gu, Qun Yang, Whitney Tsang, Artur Fierka, Panagiotis Kourdis, Tanima Dey @Meta: William Wen, Steven Troxler, Avik Chaudhuri, Elias Ellison, Laith Sakka, Angel Li, Richard Zou, Yidi Wu, Oguz Ulgen, Dunfan Lu, Jason Ansel, Jongsok Choi, Ethan Che, Kaiming Cheng, Laura Wang, Driss Guessous, Simon Layton, Marius Eriksen, Wei Feng, Anshul Sinha, Ailing Zhang, Bob Ren, Aaron Orenstein, Chien-Chin Huang, Pian Pawakapan, Sanket Jayant Purandare, Francisco Massa, Tristan Rice, Kapil Sharma, Natalia Gimelshein, Ben Carver @NVIDIA: Daniel Galvez, Michael Goldfarb, Guray Ozen, Ke Wen, Sreeram Potluri, Artem Polyakov, Anjulie Agrusa, Ryan Spring, Bruce Zitelli @anyscalecompute : Masahiro Tanaka @CrusoeAI: Suman Debnath, JanakiRam Goteti @UCBerkeley: Yi Pan University of Washington (incl. Paul Allen School): Megan Frisella, Stephanie Wang PyTorch Conference is the open source AI community's town square. Where what's next gets decided. Register now to join the open source AI community in San Jose, October 20-21: bit.ly/464FTbp Read the Open Research, Tooling & Optimization Sessions Guide here: bit.ly/4gXxn4r #PyTorchCon

The first-ever PyTorch Day Japan comes to Tokyo on December 10, bringing the community together to explore PyTorch, @vllm_project, @DeepSpeedAI, @raydistributed, Helion, and Safetensors. Co-hosted by PyTorch Foundation, @huggingface, @IBM, and @ME_JP_official, the event will feature technical talks and interactive discussions across open source AI, including training, inference, responsible AI, physical and edge AI, open model development, and the broader PyTorch ecosystem. The call for proposals is open for session presentations and lightning talks through September 27 at 11:59 PM JST. Register today! 🔗 Read more: pytorch.org/blog/pytorch-d…

For supported biomolecular structure-prediction models, NVIDIA BioNeMo Inference Runtime keeps the familiar PyTorch workflow and lets developers construct supported models as torch.nn.Module objects or reuse selected modules in custom code. In a new NVIDIA Developer Blog post, the team details how BioIR uses optimized kernels and, where applicable, CUDA Graphs for model execution, with Ray replicas for processing large batches of independent inputs across GPUs in a single node. This post covers BioIR acceleration for Boltz-2, OpenFold2, and OpenFold3. Read more: developer.nvidia.com/blog/high-thro…

↘️ Deep technical content. Meaningful conversations with people solving the same problems. At #PyTorchCon North America, go deeper into #PyTorch features, optimization, deployment, and more then continue the conversation with developers and researchers doing the work. The talks matter. So do the conversations you can only have in the room. Join us Oct 20-21 in San Jose: bit.ly/4sh3DSw

At PyTorch Conference North America 2026, Dhritiman Das, Staff Software Engineer - Machine Learning Infrastructure (@LinkedIn), will present a torch-native retrieval engine that uses PyTorch as the primary runtime for retrieval, filtering, and ranking. The system combines a GPU-resident torch tensor index, tensor-based retrieval operations, custom CUDA kernels for attribute filtering, and TorchScript model execution within a unified serving architecture orchestrated through a Rust and tch-rs backend. Dhritiman will also share how the approach scales to critical use cases like feed and search at LinkedIn. Join us in San Jose on October 20-21: hubs.la/Q04v4SL60 View the poster sessions: events.linuxfoundation.org/pytorch-confer… #PyTorchCon

Live demos. Big ideas. Zero slide-deck suspense. 👀 See #PyTorch technology in action at the Demo Theater in the #PyTorchCon North America Community Expo, October 20-21 in San Jose. Explore: bit.ly/4hb0ekq Register: bit.ly/4sh3DSw

New from @Meta Engineering: FlashAttention-4 extended with MXFP8 support for @nvidia Blackwell—from forward and backward kernels to fused quantization and jagged cross-attention. The team developed an end-to-end jagged module with fused quantization, FP8 activation and compute which is being used internally at Meta for GEM training. On the latest gen hardware, LP FA4 kernel reaches 2.85 PFLOP/s forward and 2 PFLOP/s backward, with up to a 1.30× end-to-end module speedup. ✍️ Devashish Shankar, Santosh Mohan, Jiaqi Xu, Darren Liu, Han Xu Explore the design and open source implementation in our latest blog: pytorch.org/blog/low-preci…

More visibility. New audiences. One seriously intelligent week. 🧠 Bring your Bay Area event into the #OpenSourceAIWeek lineup, October 16-25, alongside flagship events #PyTorchCon and #AGNTCon + #MCPCon in San Jose. Submit by October 15: bit.ly/4ipYZ39

At #PyTorchCon North America 2026, Sudhanshu Shrivastava, Graduate Student Researcher, Electrical and Computer Engineering, @ucdavis, will present a federated learning case study in glaucoma imaging using 5,550 color fundus photographs from nine datasets across seven countries. Sudhanshu will discuss how federated learning enables collaborative medical imaging research without data centralization, with practical lessons on site-specific fine-tuning, cross-site generalizability, and multi-site learning. Join us in San Jose on October 20-21: hubs.la/Q04v4SL60 #PyTorchCon

New chips are shipping faster than ever before, but the ecosystem is being held back by having to rewrite and then re-debug the same code over & over again. @clattner_llvm CEO and Co-founder at @Modular and EVP of Advanced AI Software & Platforms at @Qualcomm, will deliver a keynote at PyTorch Conference North America about an open software platform for heterogeneous compute powered by Mojo and MAX. PyTorch has always been the place where the best models come together, and now there's a way to get those models onto all kinds of hardware. If you're interested in Al and compute, join us at the PyTorch Conference in San Jose, CA. Register now: hubs.la/Q04v4SL60 #PyTorchCon

🎤 Meet Keynote Speaker Colin Brace, VP of Annapurna Labs at @awscloud. At #PyTorchCon North America, Colin will present “Trainium’s Journey to Native PyTorch,” sharing how #PyTorch now runs natively on AWS Trainium with no code changes required. Hear how AWS enabled eager mode and torch.compile, integrated Trainium with TorchTitan, TorchAO, and Hugging Face Transformers v5, and contributed upstream to the PyTorch ecosystem. 📅 October 20-21 📍 San Jose, California Join the PyTorch community - register today: bit.ly/4sh3DSw

@UBS From Client Communications to Competing Risks: A PyTorch Transformer for Client Attrition Model by @witty_smart

Learn how PyTorch can extend language modeling beyond classification to practical time-to-event prediction. Witold Czubala of @UBS will share a PyTorch-based transformer model that predicts six-month attrition from longitudinal advisor–client email communications for wealth management during PyTorch Conference North America in San Jose, CA. Join us October 20-21: hubs.la/Q04v4SL60 #PyTorchCon

🍎 Fresh ideas. 💬 Relevant technical conversations. 💡 Insights from the people doing the work. That’s what attendees have come to expect from #PyTorchCon North America, taking place Oct 20-21 in San Jose. Join the open source AI community for keynotes, technical talks, hands-on sessions, and candid conversations about the work shaping AI today. bit.ly/3Ra1oUx Register: bit.ly/4sh3DSw

"It's really great to feel the energy and see all the collaboration that's happening, lots of conversations. We have people here who are just starting to learn about PyTorch, and we have people who are maintainers. It's super nice to see that energy and see how together we're going to be able to solve some of these very difficult problems for humanity." Ankit Patel, @nvidia, at PyTorch Conference Europe 2026 Register today to join the open source AI community at PyTorch Conference North America in San Jose, October 20–21: hubs.la/Q04tBgv_0 #PyTorchCon

The open source AI community thrives on collaboration, choice, and open contribution. As Joseph Groenenboom from @RedHat highlights, empowering developers to freely contribute and build options is essential for growing the entire AI ecosystem. PyTorch Conference North America 2026 is where that community comes together in person. Join us to collaborate with open source pioneers, expand your deployment capabilities, and directly contribute to shaping the future of machine learning. Register for PyTorch Conference North America today: hubs.la/Q04tBgv_0 #PyTorchCon