PyTorch

PyTorch is a machine learning library based on the

PyTorch
Torch library,[4][5][6] used for applications such as
computer vision and natural language processing,[7]
originally developed by Meta AI and now part of the
Linux Foundation umbrella.[8][9][10][11] It is one of the
most popular deep learning frameworks, alongside Original author(s) Adam Paszke
others such as TensorFlow and PaddlePaddle,[12][13] Sam Gross
offering free and open-source software released under Soumith Chintala
the modified BSD license. Although the Python Gregory Chanan
interface is more polished and the primary focus of
Developer(s) Meta AI
development, PyTorch also has a C++ interface.[14]
Initial release September 2016[1]
A number of pieces of deep learning software are built Stable release 2.6.0[2] / 29 January
on top of PyTorch, including Tesla Autopilot,[15] 2025
Uber's Pyro,[16] Hugging Face's Transformers,[17]
Repository github.com/pytorch/pytorch
PyTorch Lightning,[18][19] and Catalyst.[20][21]
(https://github.com/pytorch/
PyTorch provides two high-level features:[22] pytorch)
Written in Python
Tensor computing (like NumPy) with strong
C++
acceleration via graphics processing units
(GPU) CUDA
Deep neural networks built on a tape-based Operating system Linux
automatic differentiation system
macOS
Windows
History
Platform IA-32, x86-64, ARM64
Meta (formerly known as Facebook) operates both Available in English
PyTorch and Convolutional Architecture for Fast Type
Library for machine
Feature Embedding (Caffe2), but models defined by
learning and deep learning
the two frameworks were mutually incompatible. The
BSD-3[3]
Open Neural Network Exchange (ONNX) project was License
created by Meta and Microsoft in September 2017 for Website pytorch.org (https://pytorc
converting models between frameworks. Caffe2 was h.org/)
merged into PyTorch at the end of March 2018.[23] In
September 2022, Meta announced that PyTorch would be governed by the independent PyTorch
Foundation, a newly created subsidiary of the Linux Foundation.[24]
PyTorch 2.0 was released on 15 March 2023, introducing TorchDynamo, a Python-level compiler that
makes code run up to 2x faster, along with significant improvements in training and inference
performance across major cloud platforms.[25][26]

PyTorch tensors
PyTorch defines a class called Tensor (torch.Tensor) to store and operate on homogeneous
multidimensional rectangular arrays of numbers. PyTorch Tensors are similar to NumPy Arrays, but can
also be operated on a CUDA-capable NVIDIA GPU. PyTorch has also been developing support for other
GPU platforms, for example, AMD's ROCm[27] and Apple's Metal Framework.[28]

PyTorch supports various sub-types of Tensors.[29]

Note that the term "tensor" here does not carry the same meaning as tensor in mathematics or physics.
The meaning of the word in machine learning is only superficially related to its original meaning as a
certain kind of object in linear algebra. Tensors in PyTorch are simply multi-dimensional arrays.

PyTorch neural networks

PyTorch defines a module called nn (torch.nn) to describe neural networks and to support training.
This module offers a comprehensive collection of building blocks for neural networks, including various
layers and activation functions, enabling the construction of complex models. Networks are built by
inheriting from the torch.nn module and defining the sequence of operations in the forward()
function.

Example
The following program shows the low-level functionality of the library with a simple example.

1 import torch
2 dtype = torch.float
3 device = torch.device("cpu") # Execute all calculations on the CPU
4 # device = torch.device("cuda:0") # Executes all calculations on the GPU
5
6 # Create a tensor and fill it with random numbers
7 a = torch.randn(2, 3, device=device, dtype=dtype)
8 print(a)
9 # Output: tensor([[-1.1884, 0.8498, -1.7129],
10 # [-0.8816, 0.1944, 0.5847]])
11
12 b = torch.randn(2, 3, device=device, dtype=dtype)
13 print(b)
14 # Output: tensor([[ 0.7178, -0.8453, -1.3403],
15 # [ 1.3262, 1.1512, -1.7070]])
16
17 print(a * b)
18 # Output: tensor([[-0.8530, -0.7183, 2.58],
19 # [-1.1692, 0.2238, -0.9981]])
20
21 print(a.sum())
22 # Output: tensor(-2.1540)
23
24 print(a[1,2]) # Output of the element in the third column of the second row (zero based)
 25 # Output: tensor(0.5847)
26
27 print(a.max())
28 # Output: tensor(0.8498)

The following code-block defines a neural network with linear layers using the nn module.

1 import torch
2 from torch import nn # Import the nn sub-module from PyTorch
3
4 class NeuralNetwork(nn.Module): # Neural networks are defined as classes
5 def __init__(self): # Layers and variables are defined in the __init__ method
6 super().__init__() # Must be in every network.
7 self.flatten = nn.Flatten() # Construct a flattening layer.
8 self.linear_relu_stack = nn.Sequential( # Construct a stack of layers.
9 nn.Linear(28*28, 512), # Linear Layers have an input and output shape
10 nn.ReLU(), # ReLU is one of many activation functions provided by nn
11 nn.Linear(512, 512),
12 nn.ReLU(),
13 nn.Linear(512, 10),
14 )
15
16 def forward(self, x): # This function defines the forward pass.
17 x = self.flatten(x)
18 logits = self.linear_relu_stack(x)
19 return logits

See also
Free and open-
source software
portal

Comparison of deep learning software

Differentiable programming
DeepSpeed

References
1. Chintala, Soumith (1 September 2016). "PyTorch Alpha-1 release" (https://github.com/pytorc
h/pytorch/releases/tag/v0.1.1). GitHub.
2. "PyTorch 2.6.0 Release" (https://github.com/pytorch/pytorch/releases/tag/v2.6.0). 29
January 2025. Retrieved 2 February 2025.
3. Claburn, Thomas (12 September 2022). "PyTorch gets lit under The Linux Foundation" (http
s://www.theregister.com/2022/09/12/pytorch_meta_linux_foundation/). The Register.
4. Yegulalp, Serdar (19 January 2017). "Facebook brings GPU-powered machine learning to
Python" (https://www.infoworld.com/article/3159120/artificial-intelligence/facebook-brings-gp
u-powered-machine-learning-to-python.html). InfoWorld. Retrieved 11 December 2017.
5. Lorica, Ben (3 August 2017). "Why AI and machine learning researchers are beginning to
embrace PyTorch" (https://www.oreilly.com/ideas/why-ai-and-machine-learning-researchers-
are-beginning-to-embrace-pytorch). O'Reilly Media. Retrieved 11 December 2017.
6. Ketkar, Nikhil (2017). "Introduction to PyTorch". Deep Learning with Python. Apress,
Berkeley, CA. pp. 195–208. doi:10.1007/978-1-4842-2766-4_12 (https://doi.org/10.1007%2F
978-1-4842-2766-4_12). ISBN 9781484227657.
 7. Moez Ali (Jun 2023). "NLP with PyTorch: A Comprehensive Guide" (https://www.datacamp.c
om/tutorial/nlp-with-pytorch-a-comprehensive-guide). datacamp.com. Retrieved 2024-04-01.
8. Patel, Mo (2017-12-07). "When two trends fuse: PyTorch and recommender systems" (http
s://www.oreilly.com/ideas/when-two-trends-fuse-pytorch-and-recommender-systems).
O'Reilly Media. Retrieved 2017-12-18.
9. Mannes, John. "Facebook and Microsoft collaborate to simplify conversions from PyTorch to
Caffe2" (https://techcrunch.com/2017/09/07/facebook-and-microsoft-collaborate-to-simplify-
conversions-from-pytorch-to-caffe2/). TechCrunch. Retrieved 2017-12-18. "FAIR is
accustomed to working with PyTorch – a deep learning framework optimized for achieving
state of the art results in research, regardless of resource constraints. Unfortunately in the
real world, most of us are limited by the computational capabilities of our smartphones and
computers."
10. Arakelyan, Sophia (2017-11-29). "Tech giants are using open source frameworks to
dominate the AI community" (https://venturebeat.com/2017/11/29/tech-giants-are-using-ope
n-source-frameworks-to-dominate-the-ai-community/). VentureBeat. Retrieved 2017-12-18.
11. "PyTorch strengthens its governance by joining the Linux Foundation" (https://pytorch.org/bl
og/PyTorchfoundation/). pytorch.org. Retrieved 2022-09-13.
12. "Top 30 Open Source Projects" (https://github.com/cncf/velocity). Open Source Project
Velocity by CNCF. Retrieved 2023-10-12.
13. "Welcome to the PaddlePaddle GitHub" (https://github.com/PaddlePaddle/Paddle).
PaddlePaddle Official Github Repo. Retrieved 2024-10-28.
14. "The C++ Frontend" (https://pytorch.org/cppdocs/frontend.html). PyTorch Master
Documentation. Retrieved 2019-07-29.
15. Karpathy, Andrej (6 November 2019). "PyTorch at Tesla - Andrej Karpathy, Tesla" (https://ww
w.youtube.com/watch?v=oBklltKXtDE). YouTube.
16. "Uber AI Labs Open Sources Pyro, a Deep Probabilistic Programming Language" (https://en
g.uber.com/pyro/). Uber Engineering Blog. 2017-11-03. Retrieved 2017-12-18.
17. PYTORCH-TRANSFORMERS: PyTorch implementations of popular NLP Transformers (http
s://pytorch.org/hub/huggingface_pytorch-transformers/), PyTorch Hub, 2019-12-01, retrieved
2019-12-01
18. PYTORCH-Lightning: The lightweight PyTorch wrapper for ML researchers. Scale your
models. Write less boilerplate (https://github.com/PyTorchLightning/pytorch-lightning/),
Lightning-Team, 2020-06-18, retrieved 2020-06-18
19. "Ecosystem Tools" (https://pytorch.org/ecosystem/). pytorch.org. Retrieved 2020-06-18.
20. GitHub - catalyst-team/catalyst: Accelerated DL & RL (https://github.com/catalyst-team/catal
yst), Catalyst-Team, 2019-12-05, retrieved 2019-12-05
21. "Ecosystem Tools" (https://pytorch.org/ecosystem/). pytorch.org. Retrieved 2020-04-04.
22. "PyTorch – About" (https://web.archive.org/web/20180615190804/https://pytorch.org/about/).
pytorch.org. Archived from the original (https://pytorch.org/about/) on 2018-06-15. Retrieved
2018-06-11.
23. "Caffe2 Merges With PyTorch" (https://medium.com/@Synced/caffe2-merges-with-pytorch-a
89c70ad9eb7). 2018-04-02.
24. Edwards, Benj (2022-09-12). "Meta spins off PyTorch Foundation to make AI framework
vendor neutral" (https://arstechnica.com/information-technology/2022/09/meta-spins-off-pyto
rch-foundation-to-make-ai-framework-vendor-neutral/). Ars Technica.
25. "Dynamo Overview" (https://pytorch.org/docs/stable/torch.compiler_dynamo_overview.html).
26. "PyTorch 2.0 brings new fire to open-source machine learning" (https://venturebeat.com/ai/p
ytorch-2-0-brings-new-fire-to-open-source-machine-learning/). VentureBeat. 15 March 2023.
Retrieved 16 March 2023.
27. "Installing PyTorch for ROCm" (https://rocm.docs.amd.com/projects/install-on-linux/en/latest/
how-to/3rd-party/pytorch-install.html). rocm.docs.amd.com. 2024-02-09.
28. "Introducing Accelerated PyTorch Training on Mac" (https://pytorch.org/blog/introducing-acc
elerated-pytorch-training-on-mac/). pytorch.org. Retrieved 2022-06-04.
29. "An Introduction to PyTorch – A Simple yet Powerful Deep Learning Library" (https://www.an
alyticsvidhya.com/blog/2018/02/pytorch-tutorial/). analyticsvidhya.com. 2018-02-22.
Retrieved 2018-06-11.

External links
Official website (https://pytorch.org)

Retrieved from "https://en.wikipedia.org/w/index.php?title=PyTorch&oldid=1259992896"