pytorch · svekars · Jul 29, 2025 · Jul 29, 2025 · Jul 30, 2025 · Jul 30, 2025
diff --git a/docs/source/index.md b/docs/source/index.md
@@ -23,17 +23,11 @@ The APIs and performance characteristics of these features may change.
 :glob:
 :maxdepth: 2
 
+Install PyTorch <https://pytorch.org/get-started/locally/>
+user_guide/index
 pytorch-api
 notes
-```
-
-```{toctree}
-:glob:
-:hidden:
-:maxdepth: 2
-
 community/index
-C++ <https://docs.pytorch.org/cppdocs/>
 ```
 
 ## Indices and tables

diff --git a/docs/source/pytorch-api.md b/docs/source/pytorch-api.md
@@ -1,9 +1,16 @@
 (pytorch_api)=
-# Python API
+# Reference API
+
+```{toctree}
+:maxdepth: 1
+
+C++ <https://docs.pytorch.org/cppdocs/>
+```
 
 ```{toctree}
 :glob:
 :maxdepth: 1
+:caption: Python API
 
 torch
 nn

diff --git a/docs/source/user_guide/index.md b/docs/source/user_guide/index.md
@@ -0,0 +1,39 @@
+# User Guide
+
+PyTorch provides a flexible and efficient platform for building deep
+learning models, offering dynamic computation graphs and a rich
+ecosystem of tools and libraries. This guide will help you harness the power
+of PyTorch to create and deploy machine learning models effectively.
+
+```{note}
+This guide is a work in progress.
+```
+
+```{toctree}
+:maxdepth: 1
+:caption: Introduction
+
+Pytorch Overview <https://docs.pytorch.org/tutorials/beginner/deep_learning_60min_blitz.html>
+Get Started <https://pytorch.org/get-started/locally/>
+Learn the Basics <https://docs.pytorch.org/tutorials/beginner/basics/intro.html>
+```
+
+```{toctree}
+:maxdepth: 1
+:caption: Core Concepts
+
+pytorch_main_components
+```
+
+```{toctree}
+:maxdepth: 1
+:caption: Beyond the Basics
+
+```
+
+```{toctree}
+:maxdepth: 1
+:caption: Developer Notes
+
+../notes
+```
diff --git a/docs/source/user_guide/pytorch_main_components.md b/docs/source/user_guide/pytorch_main_components.md
@@ -0,0 +1,29 @@
+(pytorch_main_components)=
+# PyTorch Main Components
+
+PyTorch is a flexible and powerful library for deep learning that provides a comprehensive set of tools for building, training, and deploying machine learning models.
+
+## PyTorch Components for Basic Deep Learning
+
+Some of the basic PyTorch components include:
+
+* **Tensors** - N-dimensional arrays that serve as PyTorch's fundamental
+data structure. They support automatic differentiation, hardware acceleration, and provide a comprehensive API for mathematical operations.
+
+* **Autograd** - PyTorch's automatic differentiation engine
+that tracks operations performed on tensors and builds a computational
+graph dynamically to be able to compute gradients.
+
+* **Neural Network API** - A modular framework for building neural networks with pre-defined layers,
+activation functions, and loss functions. The {mod}`nn.Module` base class provides a clean interface
+for creating custom network architectures with parameter management.
+
+* **DataLoaders** - Tools for efficient data handling that provide
+features like batching, shuffling, and parallel data loading. They abstract away the complexities
+of data preprocessing and iteration, allowing for optimized training loops.
+
+
+## PyTorch Compiler
+
+The PyTorch compiler is a suite of tools that optimize model execution and
+reduce resource requirements. You can learn more about the PyTorch compiler [here](https://docs.pytorch.org/docs/stable/torch.compiler_get_started.html).
diff --git a/docs/source/user_guide/pytorch_overview.md b/docs/source/user_guide/pytorch_overview.md
@@ -0,0 +1,54 @@
+---
+file_format: mystnb
+kernelspec:
+  name: python3
+mystnb:
+  execution_timeout: 30
+  merge_streams: True
+---
+
+```{code-cell}
+:tags: [remove-cell]
+import torch
+```
+
+(what_is_pytorch)=
+
+# What is PyTorch?
+
+PyTorch, or torch, is an open-source machine learning library written in Python that
+provides a platform for deep learning. It features a dynamic computational graph
+that allows for flexible model building and debugging.
+
+At its core, PyTorch uses tensors (multidimensional arrays) that can run on GPUs
+for accelerated computation. For example, the gradient descent optimization
+process can be represented as:
+
+```{math}
+\theta_{\text{new}} = \theta_{\text{old}} - \alpha \nabla_{\theta} J(\theta)
+```
+
+* `θ_new` is the updated parameter
+* `θ_old` is the current parameter
+* `α` is the learning rate
+* `∇_θ J(θ)` is the gradient of the cost function with respect to `θ`
+
+PyTorch's autograd engine automatically computes the gradients needed for neural network training.
+
+For a quick tutorial on PyTorch, see the [Learn the Basics tutorial](https://pytorch.org/tutorials/beginner/basics/intro.html).
+
+# Beyond Basic Operations
+
+PyTorch extends far beyond basic arithmetic operations to provide a complete ecosystem for machine learning development:
+* {mod}`torch.nn` module: Offers pre-built neural network layers, activation functions, and loss functions for constructing complex architectures
+* {mod}`torch.utils.data`: Provides efficient data loading utilities with support for batching, shuffling, and parallel data loading
+* {mod}`torch.optim`: Contains a comprehensive suite of optimization algorithms including SGD, Adam, RMSprop, and many others
+* {func}`torch.compile`: Allows for compiling PyTorch models to improve execution speed and efficiency
+* {func}`torch.export`: Enables exporting models for deployment in various environments.
+* Distributed training: Facilitates training across multiple GPUs and nodes, making it suitable for large-scale machine learning tasks
+
+```{seealso}
+* {ref}`torch.compiler_overview`
+* {ref}`torch.export`
+* [Learn the basics](https://docs.pytorch.org/tutorials/beginner/basics/intro.html)
+```