Get to know PyTorch - a cutting edge open source AI framework for Machine Learning
summary
High‑level overview gives way to specifics: PyTorch now sits under the PyTorch Foundation and ships PyTorch 2.0; core features shown include Python/C++ APIs, built‑in datasets and DataLoader, and the automatic differentiation system autograd. Scaling notes include FSDP demonstrated on 500+ GPUs across 60+ nodes to cut memory footprint. IBM ties PyTorch into Watson X and points to torchvision torchaudio torchtext and torchbench and ongoing mobile/beta efforts.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:51Deep learning basics
- 6:30PyTorch as an open source framework
- 8:58Data handling in PyTorch
- 11:45Training process
- 15:00Distributed training
- 16:47Contributing to PyTorch
- 18:28PyTorch in Watson X
- 19:14Q&A
related talks
47mGet to Know PyTorch - A Cutting Edge Open Source AI Framework for Deep Learning
12mKeynote: How PyTorch Became the Foundation of the AI Revolution
19mIncreasing the Impact of AI Through Better Software (Pytorch Updates)
54mKeynote: PyTorch 2.1 Technical Deep Dive
06mKeynote: PyTorch Lightning: Powering the GenAI Revolution from Research to the...
12mLightning Talk: State of PyTorch
22mIncreasing the Impact of AI Through Better Software
43mPyTorch Technical Deep Dive