More Related Content
Introduction To TensorFlow | Deep Learning Using TensorFlow | TensorFlow Tuto...
GPU and Deep learning best practices
"Quantizing Deep Networks for Efficient Inference at the Edge," a Presentatio...
PyTorch Python Tutorial | Deep Learning Using PyTorch | Image Classifier Usin...
KubeFlow + GPU + Keras/TensorFlow 2.0 + TF Extended (TFX) + Kubernetes + PyTo...
Clean architectures with fast api pycones
What's hot
20分くらいでわかった気分になれるC++20コルーチン
物体検出フレームワークMMDetectionで快適な開発
Python Course | Python Programming | Python Tutorial | Python Training | Edureka
わかる！metadata.managedFields / Kubernetes Meetup Tokyo 48
Grafana LokiではじめるKubernetesロギングハンズオン（NTT Tech Conference #4 ハンズオン資料）
2015年度GPGPU実践プログラミング 第5回 GPUのメモリ階層
Interpreter, Compiler, JIT from scratch
Docker Networking with New Ipvlan and Macvlan Drivers
Python Basics | Python Tutorial | Edureka
組み込み関数（intrinsic）によるSIMD入門
Similar to PyTorch 2 Internals
Pytorch for tf_developers
pytdddddddddddddddddddddddddddddddddorch.pdf
Julien Simon - Deep Dive: Compiling Deep Learning Models
Essentials of PyTorch Tutorial .pptx
Soumith Chintala - Increasing the Impact of AI Through Better Software
PyTorch Deep Learning Framework | USDSI®
PyTorch Tutorial for NTU Machine Learing Course 2017
Machine learning with py torch
Simple, fast, and scalable torch7 tutorial
PyTorch for Deep Learning Practitioners
2Wisjshsbebe pehele isienew Dorene isksnwnw
pytorch_tutorial_follow_this_to_start.pptx
[Update] PyTorch Tutorial for NTU Machine Learing Course 2017
Torch: a scientific computing framework for machine-learning practitioners
"PyTorch Deep Learning Framework: Status and Directions," a Presentation from...
Python - Deep Learning library-PyTorch.pptx
More from Christian Perone
Gradient-based optimization for Deep Learning: a short introduction
Bayesian modelling for COVID-19 seroprevalence studies
Uncertainty Estimation in Deep Learning
Word Embeddings - Introduction
Apache Spark - Intro to Large-scale recommendations with Apache Spark and Python
Deep Learning - Convolutional Neural Networks - Architectural Zoo
Deep Learning - Convolutional Neural Networks
Machine Learning com Python e Scikit-learn
Python - Introdução Básica
C++0x :: Introduction to some amazing features
Recently uploaded
VIBE CHECK: Accelerating Project Prototyping with AI-Driven Vibe Coding
ScriptRipple: Accelerating Digital Creation Marketplace
Abolena Mobile OS fullstack rendering software benchmark
Are Your Workplace Safety Videos Actually Driving Behavior Change
Advanced MERN Stack Development for Modern Web Applications
MailsDaddy PST Attachment Extractor: Efficient Bulk Email Attachment Management
ServiceNow to Freshservice Migration Case Study
Executive Assistant That Runs Your Calendar, Inbox & Scheduling - Consul
PyTorch 2 Internals
- 1.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch PyTorch 2 internals A not so short guide to recent PyTorch innovations Christian S. Perone (christian.perone@gmail.com) http://blog.christianperone.com London, UK, Dec 2023
- 2.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Who Am I ▸ Christian S. Perone ▸ ML Research Engineer in London/UK ▸ Blog at ▸ blog.christianperone.com ▸ Open-source projects at ▸ https://github.com/perone ▸ Twitter @tarantulae
- 3.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Disclaimer PyTorch development pace is so fast that no man ever steps in PyTorch code twice, for it’s not the same code and he’s not the same man. —Heraclitus, 500 BC
- 4.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Section I ; Tensors <
- 5.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors Simply put, tensors are a generalization of vectors and matrices. In PyTorch, they are a multi-dimensional matrix containing elements of a single data type.
- 6.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors Simply put, tensors are a generalization of vectors and matrices. In PyTorch, they are a multi-dimensional matrix containing elements of a single data type. >>> import torch >>> t = torch.tensor([[1., -1.], [1., -1.]]) >>> t tensor([[ 1., -1.] [ 1., -1.]])
- 7.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors Simply put, tensors are a generalization of vectors and matrices. In PyTorch, they are a multi-dimensional matrix containing elements of a single data type. >>> import torch >>> t = torch.tensor([[1., -1.], [1., -1.]]) >>> t tensor([[ 1., -1.] [ 1., -1.]]) >>> t.dtype # They have a type torch.float32
- 8.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors Simply put, tensors are a generalization of vectors and matrices. In PyTorch, they are a multi-dimensional matrix containing elements of a single data type. >>> import torch >>> t = torch.tensor([[1., -1.], [1., -1.]]) >>> t tensor([[ 1., -1.] [ 1., -1.]]) >>> t.dtype # They have a type torch.float32 >>> t.shape # a shape torch.Size([2, 2])
- 9.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors Simply put, tensors are a generalization of vectors and matrices. In PyTorch, they are a multi-dimensional matrix containing elements of a single data type. >>> import torch >>> t = torch.tensor([[1., -1.], [1., -1.]]) >>> t tensor([[ 1., -1.] [ 1., -1.]]) >>> t.dtype # They have a type torch.float32 >>> t.shape # a shape torch.Size([2, 2]) >>> t.device # and live in some device device(type='cpu')
- 10.
PyTorch 2 internals- Christian S. Perone (2023) Tensors JIT Dynamo Inductor Torch Export ExecuTorch Tensors ▸ Although PyTorch has an elegant python first design, all PyTorch heavy work is actually implemented in C++. ▸ In Python, the integration of C++ code is (usually) done using what is called an extension;