Horace He on Building ML Systems for Extreme-Scale Computation at Jane Street
Horace He from Meta's PyTorch team delivers a technical talk at Jane Street about building machine learning systems capable of operating at extreme scale — specifically systems designed to handle a trillion trillion floating point operations. The talk covers the intersection of compilers and machine learning, including innovations like torch.compile and FlexAttention, which enable significant performance improvements and flexible kernel design within the Python ecosystem. The content is a deep technical presentation aimed at an audience of quantitative trading engineers and computer scientists.
Key quotes
If you've used things like torch.compile, which is a thing in PyTorch that makes your model go like 2-4x faster in just one line of code.
Or if you've used FlexAttention, which is something that lets researchers design fast kernels for attention without leaving the world of Python.
He's the person responsible for both of those things.
From the article
You might also wanna read
Google's TorchTPU Enables Native PyTorch Execution on TPU Infrastructure
Google's TorchTPU is a new engineering stack that enables native, high-performance execution of PyTorch workloads on Google's TPU infrastruc
developers.googleblog.com·2mo ago
AI-Generated Metal Kernels Accelerate PyTorch Inference by 87% on Apple Devices
Researchers developed AI-generated Metal kernels that accelerate PyTorch inference on Apple devices by 87% across 215 modules. The study dem
gimletlabs.ai·10mo ago
mytorch: Python Automatic Differentiation Library Inspired by PyTorch
mytorch is an open-source Python library that implements automatic differentiation with a PyTorch-like API, using NumPy for computations. Th
github.com·6mo ago
Torchcomms: New Experimental Communication API for PyTorch Distributed Training at Scale
Torchcomms is a new experimental, lightweight communication API designed for PyTorch Distributed (PTD) that aims to enable large-scale model
pytorch.org·8mo ago
MegaTrain: Full Precision Training of 100B+ Parameter Large Language Models on a Single GPU
Lightning AI: Cloud IDE for Machine Learning and PyTorch Development
Lightning AI is a cloud-based integrated development environment (IDE) specifically designed for machine learning and PyTorch development. I
Product Hunt·8mo ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.