APAC/ US Opening Talks

on YouTube ↗55m

" summary

Project updates and deployments focus on concrete tools and measured outcomes: releases ship quarterly (~3,000 commits, 800+ contributors) and the 1.8 cycle reviews ~41 features; new tooling includes torch.fx for Python IR transforms and graph-mode quantization, PyTorch profiler + Kineto with a TensorBoard plugin that shows CPU→GPU traces and cuts a sample step down to ~125 ms after fixing data‑loading, beta AMD GPU support, pipeline parallelism and a DDP communication hook (PowerSGD) for distributed scaling, and TorchServe/KFServing/MLflow integrations for versioned serving and canary rollouts.

AI-written summary · about

" chapters (4)

1. 0:00Start
2. 0:1625:33: Ritchie Ng - CEO @Hessain Matrix Research Scholar, NUS
3. 25:4042:57: Joe Spisak - PyTorch Product Lead @Facebook AI
4. 43:0355:08: Geeta Chauhan - AI/PyTorch Partner Engineering Head @Facebook AI

" related talks

• 12mLightning Talk: State of PyTorchAlban Desmaison · PyTorch Conference 2023
• 48mFireside ChatSoumith Chintala, Ibraham Haddad · Emerging OS Forum @ Open Source Summit Europe 2022
• 43mPyTorch Technical Deep DivePeng Wu, Alban Desmaison, Mark Saroufim, Edward Yang · PyTorch Conference 2025
• 13mKeynoteMike Schroepfer · PyTorch Developer Conference 2019
• 03mRecapPyTorch Developer Conference 2019
• 21mWhat's New for PyTorch Developer InfrastructureEli Uriegas, Omkar Salpekar · PyTorch Conference 2023
• 47mGet to Know PyTorch - A Cutting Edge Open Source AI Framework for Deep LearningSahdev P. Zala · Open AI + Data Forum @ Open Source Summit North America 2024
• 12mKeynote: How PyTorch Became the Foundation of the AI RevolutionJoe Spisak · PyTorch Conference 2023