EMEA/US Opening Talks
summary
Five short updates: a personal path from research to contribution and forum-led feature work that helps shape priorities; release process now runs quarterly with RCs, 1.8 ships ~41 reviewed features and tools like torch.fx graph transformations, numpy-aligned torch.linalg/fft, and torch.profiler + Kineto for GPU+TensorBoard profiling; distributed work adds pipeline parallelism, DDP communication hooks and beta AMD GPU support; partner integrations include TorchServe, Kubeflow pipelines and MLflow hooks; Disney’s Creative Genome runs an end-to-end Python pipeline with 270 models in production, multimodal models, optuna and torchdrift.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:1609:10: Piotr Bialecki - Technical Lead, PyTorch @NVIDIA
- 9:1626:34: Joe Spisak - PyTorch Product Lead @Facebook AI
- 26:4138:45: Geeta Chauhan - AI/PyTorch Partner Engineering Head @Facebook AI
- 38:5248:18: Miquel Farré - Senior Technology Manager @Disney
- 48:2554:34: Suraj Subramanian - PyTorch Developer Advocate @Faceook AI
related talks
48mFireside Chat
12mLightning Talk: State of PyTorch
43mPyTorch Technical Deep Dive
11mBuilding Frontier Intelligence in the Open
47mGet to Know PyTorch - A Cutting Edge Open Source AI Framework for Deep Learning
30mOur Journey With TorchTitan
11mUnlock the Future of Generative AI: TorchTitan's Latest...
22mIncreasing the Impact of AI Through Better Software
