The PyTorch Foundation is a neutral home for the deep learning community to collaborate on the open source PyTorch framework and ecosystem. The PyTorch Foundation is supported by leading contributors to the PyTorch open source project. The Foundation leverages resources provided by members and contributors to enable community discussions and collaboration.
Community collaboration is critical for the framework’s evolution as well as the development of associated projects that support using PyTorch in production and at scale. As part of The Linux Foundation, the PyTorch community will also collaborate on training, local and regional events, open source developer tooling, academic research, and guides to help new users and contributors have a productive experience.
The TorchData library provides new composable building blocks called DataPipes and a new DataLoader2. Together, they allow easy construction of flexible, reusable data pipeline and execution in various settings/backends. We will examine these new features, showcase examples, and discuss how they improve upon the data loading functionalities provided by PyTorch.View details
Learn about the what a production ready inference framework needs to do outside of model.forward() and how to make model.forward() run fast.View details
Get any questions you might have about symbolic shapes support in PyTorch.View details
Introduction to 2-D Parallelism (FSDP + Tensor Parallel) to train large scale ViT modelsView details
Learn more about TorchMultimodal - a PyTorch library for training SoTA multimodal multi-task models at scale.View details
In this talk, we will a provide an update on how we are developing new capabilities in PyTorch to both significantly increase the coverage of devices and use-cases as well as further reduce the friction in the research to production workflow.View details
Jan 25, 2023
PyTorch 2.0 Ask the Engineers Q&A Series: Deep Dive into TorchInductor and PT2 Backend Integration
Jan 24, 2023
PyTorch 2.0 Ask the Engineers Q&A Series: PT2 and Distributed (DDP/FSDP)
Jan 5, 2023