Rename new neural networks tutorial and add contents #6105

RalfG · 2025-05-17T01:07:43Z

Renames the tutorial to a more fitting title
Added metadata
Added tutorial contents

bebatut

Here is my review with some suggestions
Could you add the previous path in bin/check-url-persistence.sh here:
https://github.com/galaxyproject/training-material/blob/main/bin/check-url-persistence.sh#L32?
Thanks a lot

bebatut · 2025-05-19T12:09:54Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+- Discovering the basic training loop
+- Learning the concepts of an activation function
+- Training a simple fully-connected neural network
+time_estimation: 1H


Suggested change

time_estimation: 1H

time_estimation: 1H

redirect_from:

- /topics/statistics/tutorials/deep-learning-without-gai-with-python/tutorial.md

bebatut · 2025-05-19T12:10:13Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+layout: tutorial_hands_on
+title: "Practical deep learning with PyTorch"
+level: Intermediate
+draft: true


Suggested change

draft: true

bebatut · 2025-05-19T12:10:33Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+  - intro-to-ml-with-python
+  - neural-networks-with-python
+questions:
+- to do


Could you add a few questions covered in the tutorial?

bebatut · 2025-05-19T13:13:15Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+---
+
+
+# 1 Neural networks with PyTorch


Suggested change

# 1 Neural networks with PyTorch

# Neural networks with PyTorch

bebatut · 2025-05-19T13:13:23Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+_ = torch.manual_seed(42)
+```
+
+## 1.1 Working with tensors


Suggested change

## 1.1 Working with tensors

## Working with tensors

bebatut · 2025-05-19T13:42:49Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+
+Now the dataset can be accessed by index, just like a list. Each item is a tuple containing the features and targets as tensors.
+
+*Assignment: Try to access the first 10 samples of the dataset. What do you see?*


Suggested change

*Assignment: Try to access the first 10 samples of the dataset. What do you see?*

> <hands-on-title></hands-on-title>

> Try to access the first 10 samples of the dataset.

>

> > <question-title></question-title>

> > What do you see?

> {: .question}

{: .hands-on}

bebatut · 2025-05-19T13:43:31Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+
+### 1.4.2 Extending the model architecture
+
+*Assignment: Modify the `BreastCancerClassifier` class to include three hidden layers, each with a decreasing number of neurons (30, 15, 1). Use the ReLU activation function between each layer. The final output layer should have 1 neuron and use the sigmoid activation function.*


Suggested change

*Assignment: Modify the `BreastCancerClassifier` class to include three hidden layers, each with a decreasing number of neurons (30, 15, 1). Use the ReLU activation function between each layer. The final output layer should have 1 neuron and use the sigmoid activation function.*

> <hands-on-title></hands-on-title>

> 1. Modify the `BreastCancerClassifier` class to include three hidden layers, each with a decreasing number of neurons (30, 15, 1).

> 2. Use the ReLU activation function between each layer.

>

> The final output layer should have 1 neuron and use the sigmoid activation function.

{: .hands-on}

bebatut · 2025-05-19T13:44:25Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+
+### 1.4.3 Training the model
+
+*Assignment: Now repeat the training process by reusing your code from 1.3.3. How does the model perform? Is it better than the previous model?*


Suggested change

*Assignment: Now repeat the training process by reusing your code from 1.3.3. How does the model perform? Is it better than the previous model?*

> <hands-on-title></hands-on-title>

> Repeat the training process by reusing your code from 1.3.3.

>

> > <question-title></question-title>

> > 1. How does the model perform?

> > 2. Is it better than the previous model?

> {: .question}

{: .hands-on}

bebatut · 2025-05-19T13:45:09Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+
+All configuration of the training process is done in the [`Trainer`](https://pytorch-lightning.readthedocs.io/en/stable/common/trainer.html#trainer) class. This class will take care of the training and validation loops, as well as logging and checkpointing. We will create an instance of this class and pass it our model and the training and validation data loaders.
+
+*Assignment: Browse through the documentation for the trainer class and try to understand the different arguments of the class. Pay special attention to the `accelerator` and `devices` arguments, which are some of the most useful features of PyTorch Lightning.*


Suggested change

*Assignment: Browse through the documentation for the trainer class and try to understand the different arguments of the class. Pay special attention to the `accelerator` and `devices` arguments, which are some of the most useful features of PyTorch Lightning.*

> <hands-on-title></hands-on-title>

> Browse through the documentation for the trainer class and try to understand the different arguments of the class. Pay special attention to the `accelerator` and `devices` arguments, which are some of the most useful features of PyTorch Lightning.

{: .hands-on}

bebatut · 2025-05-19T13:45:58Z

topics/statistics/tutorials/practical-deep-learning-with-pytorch/tutorial.md

+  language: python
+  pyolite: true
+---
+


Suggested change

> <agenda-title></agenda-title>

>

> In this tutorial, we will cover:

>

> 1. TOC

> {:toc}

>

{: .agenda}

RalfG added 2 commits May 17, 2025 02:07

Rename new neural networks tutorial and add contents

47ea869

Add part 2

fb69920

github-actions bot added the statistics label May 17, 2025

bebatut reviewed May 19, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Rename new neural networks tutorial and add contents #6105

Rename new neural networks tutorial and add contents #6105

Uh oh!

RalfG commented May 17, 2025

Uh oh!

bebatut left a comment

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

bebatut May 19, 2025

Uh oh!

Uh oh!

	# 1 Neural networks with PyTorch
	# Neural networks with PyTorch


		Now the dataset can be accessed by index, just like a list. Each item is a tuple containing the features and targets as tensors.

		Assignment: Try to access the first 10 samples of the dataset. What do you see?

-*Assignment: Try to access the first 10 samples of the dataset. What do you see?*
+> <hands-on-title></hands-on-title>
+> Try to access the first 10 samples of the dataset.
+>
+> > <question-title></question-title>
+> > What do you see?
+> {: .question}
+{: .hands-on}


		### 1.4.2 Extending the model architecture

		Assignment: Modify the `BreastCancerClassifier` class to include three hidden layers, each with a decreasing number of neurons (30, 15, 1). Use the ReLU activation function between each layer. The final output layer should have 1 neuron and use the sigmoid activation function.


		### 1.4.3 Training the model

		Assignment: Now repeat the training process by reusing your code from 1.3.3. How does the model perform? Is it better than the previous model?


		All configuration of the training process is done in the [`Trainer`](https://pytorch-lightning.readthedocs.io/en/stable/common/trainer.html#trainer) class. This class will take care of the training and validation loops, as well as logging and checkpointing. We will create an instance of this class and pass it our model and the training and validation data loaders.

		Assignment: Browse through the documentation for the trainer class and try to understand the different arguments of the class. Pay special attention to the `accelerator` and `devices` arguments, which are some of the most useful features of PyTorch Lightning.

+> <agenda-title></agenda-title>
+>
+> In this tutorial, we will cover:
+>
+> 1. TOC
+> {:toc}
+>
+{: .agenda}

Rename new neural networks tutorial and add contents #6105

Are you sure you want to change the base?

Rename new neural networks tutorial and add contents #6105

Uh oh!

Conversation

RalfG commented May 17, 2025

Uh oh!

bebatut left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!