Blog

Blog

Accelerating Generative AI Part III: Diffusion, Fast

This post is the third part of a multi-series blog focused on how to accelerate…

Sayak Paul and Patrick von Platen (Hugging Face 🤗)January 3, 2024

Blog

Understanding GPU Memory 2: Finding and Removing Reference Cycles

This is part 2 of the Understanding GPU Memory blog series. Our first post Understanding GPU…

Aaron Shi, Zachary DeVitoDecember 19, 2023

Blog

Training Production AI Models with PyTorch 2.0

1. Introduction PyTorch 2.0 (abbreviated as PT2) can significantly improve the training and inference performance of…

CK Luk, Daohang Shi, Yuzhen Huang, Jackie (Jiaqi) Xu, Jade Nie, Zhou Wang, Lu Fang, Flavio Sales Truzzi, Devashish Shankar, Dima Ivashchenko, Chunzhi Yang, Nicolas Macchioni, David Berard, Yu Guo, Xiaodong Wang, Bert Maher, Yanbo Liang, Edward Yang, Brian Hirsh, Michael Voznesensky, Animesh Jain, Michael AndersonDecember 18, 2023

Blog

Empowering Models with Performance: The Art of Generalized Model Transformation Approach

Introduction PyTorch 2.0 (PT2) offers a compiled execution mode which rewrites Python bytecode to extract sequences…

Jackie (Jiaqi) Xu, Yanbo Liang, Jason Ansel, Chunzhi Yang, Jade Nie, Yuzhen Huang, CK Luk, Xiaodong Wang, Lu Fang, Menglu Yu, Jinwon Lee, Daohang Shi, Flavio Sales TruzziDecember 15, 2023

Blog

Understanding GPU Memory 1: Visualizing All Allocations over Time

During your time with PyTorch on GPUs, you may be familiar with this common error…

Aaron Shi, Zachary DeVitoDecember 14, 2023

Blog

From PyTorch Conference 2023: From Dinosaurs to Seismic Imaging with Intel

Lightning Talk 1: Seismic Data to Subsurface Models with OpenFWI Speaker: Benjamin Consolvo, AI Software…

Ramya Ravi, Susan Kahler at IntelDecember 12, 2023

Community

PyPose: A Library for Robot Learning with Physics-based Optimization

We are excited to share our new open-source library PyPose. It is a PyTorch-based robotics-oriented…

PyPoseDecember 6, 2023

Blog

Accelerating Generative AI with PyTorch II: GPT, Fast

This post is the second part of a multi-series blog focused on how to accelerate…

PyTorch FoundationNovember 30, 2023

Blog

PyTorch 2.1 Contains New Performance Features for AI Developers

We are excited to see the release of PyTorch 2.1. In this blog, we discuss…

IntelNovember 29, 2023

Blog

Accelerating Generative AI with PyTorch: Segment Anything, Fast

This post is the first part of a multi-series blog focused on how to accelerate…

PyTorch FoundationNovember 16, 2023

Community

How Activation Checkpointing enables scaling up training deep learning models

Activation checkpointing is a technique used for reducing the memory footprint at the cost of…

PyTorch FoundationNovember 9, 2023

Blog

PyTorch compile to speed up inference on Llama 2

In this blog, we discuss how to improve the inference latencies of the Llama 2…

IBM Research: Antoni Viros i Martin, Brian Vaughan, Davis Wertheimer, Joshua Rosenkranz, Mudhakar Srivatsa, Nelson Mimura Gonzalez, Raghu Ganti, Supriyo Chakraborty, Zhuoran Liu Meta: Geeta Chauhan, Hamid ShojanazeriNovember 7, 2023

Blog

High-Performance Llama 2 Training and Inference with PyTorch/XLA on Cloud TPUs

In a landscape where AI innovation is accelerating at an unprecedented pace, Meta’s Llama family of open…

Jiewen Tan, Jon Bolin, Yeounoh Chung, Liyang Lu, Siyuan Liu, Wonjoo Lee, Manfei Bai, Meghan Cowan, Jack Cao, Milad Mohammadi, Shauheen Zahirazami, Alex SpiridonovNovember 6, 2023

Blog

Accelerating Inference on x86-64 Machines with oneDNN Graph

Supported in PyTorch 2.0 as a beta feature, oneDNN Graph leverages aggressive fusion patterns to…

IntelNovember 2, 2023

Blog

AMD Extends Support for PyTorch Machine Learning Development on Select RDNA™ 3 GPUs with ROCm™ 5.7

Researchers and developers working with Machine Learning (ML) models and algorithms using PyTorch can now…

AMDOctober 31, 2023

Community

torch.compile, explained

Have you ever felt overwhelmed by the complexities of torch.compile? Diving into its workings can…

Kaichao YouOctober 26, 2023

Blog

Compiling NumPy code into C++ or CUDA via torch.compile

Quansight engineers have implemented support for tracing through NumPy code via torch.compile in PyTorch 2.1. This feature…

Evgeni Burovski, Ralf Gommers and Mario LezcanoOctober 17, 2023

Blog

Flash-Decoding for long-context inference

Motivation Large language models (LLM) such as ChatGPT or Llama have received unprecedented attention lately.…

Tri Dao, Daniel Haziza, Francisco Massa, Grigory SizovOctober 13, 2023

Blog

ML Model Server Resource Saving – Transition From High-Cost GPUs to Intel CPUs and oneAPI powered Software with performance

Reviewers: Yunsang Ju(Naver GplaceAI Leader), Min Jean Cho(Intel), Jing Xu(Intel), Mark Saroufim(Meta) Intro Here, We will…

Sangjune Park(Naver GplaceAI MLOps), Jooyoung Lee(Naver GplaceAI MLE), Junho Min(Naver GplaceAI MLE)October 11, 2023

Blog

Real-time Audio-visual Speech Recognition

Audio-Visual Speech Recognition (AV-ASR, or AVSR) is the task of transcribing text from audio and…

PyTorch FoundationOctober 10, 2023

Accelerating Generative AI Part III: Diffusion, Fast

Understanding GPU Memory 2: Finding and Removing Reference Cycles

Training Production AI Models with PyTorch 2.0

Empowering Models with Performance: The Art of Generalized Model Transformation Approach

Understanding GPU Memory 1: Visualizing All Allocations over Time

From PyTorch Conference 2023: From Dinosaurs to Seismic Imaging with Intel

PyPose: A Library for Robot Learning with Physics-based Optimization

Accelerating Generative AI with PyTorch II: GPT, Fast

PyTorch 2.1 Contains New Performance Features for AI Developers

Accelerating Generative AI with PyTorch: Segment Anything, Fast

How Activation Checkpointing enables scaling up training deep learning models

PyTorch compile to speed up inference on Llama 2

High-Performance Llama 2 Training and Inference with PyTorch/XLA on Cloud TPUs

Accelerating Inference on x86-64 Machines with oneDNN Graph

AMD Extends Support for PyTorch Machine Learning Development on Select RDNA™ 3 GPUs with ROCm™ 5.7

torch.compile, explained

Compiling NumPy code into C++ or CUDA via torch.compile

Flash-Decoding for long-context inference

ML Model Server Resource Saving – Transition From High-Cost GPUs to Intel CPUs and oneAPI powered Software with performance

Real-time Audio-visual Speech Recognition

Docs

Tutorials

Resources

Stay in touch for updates, event info, and the latest news