When you build and train a PyTorch deep learning model, you can provide the training data in several different ways. Ultimately, a PyTorch model works like a function that takes a PyTorch tensor and returns you another tensor. You have a lot of freedom in how to get the input tensors. Probably the easiest is to prepare a large tensor of the entire dataset and extract a small batch from it in each training step. But you will see that […]

You can learn a lot about neural networks and deep learning models by observing their performance over time during training. For example, if you see the training accuracy went worse with training epochs, you know you have issue with the optimization. Probably your learning rate is too fast. In this post, you will discover how you can review and visualize the performance of PyTorch models over time during training. After completing this post, you will know: What metrics to collect […]

A large deep learning model can take a long time to train. You lose a lot of work if the training process interrupted in the middle. But sometimes, you actually want to interrupt the training process in the middle because you know going any further would not give you a better model. In this post, you will discover how to control the training loop in PyTorch such that you can resume an interrupted process, or early stop the training loop. […]

PyTorch is a deep learning library. You can build very sophisticated deep learning models with PyTorch. However, there are times you want to have a graphical representation of your model architecture. In this post, you will learn: How to save your PyTorch model in an exchange format How to use Netron to create a graphical representation. Kick-start your project with my book Deep Learning with PyTorch. It provides self-study tutorials with working code. Let’s get started. Visualizing a PyTorch ModelPhoto […]

Neural networks are built with layers connected to each other. There are many different kind of layers. For image related applications, you can always find convolutional layers. It is a layer with very few parameters but applied over a large sized input. It is powerful because it can preserve the spatial structure of the image. Therefore it is used to produce state-of-the-art results on computer vision neural networks. In this post, you will learn about the convolutional layer and the […]

A popular demonstration of the capability of deep learning techniques is object recognition in image data. The “hello world” of object recognition for machine learning and deep learning is the MNIST dataset for handwritten digit recognition. In this post, you will discover how to develop a deep learning model to achieve near state-of-the-art performance on the MNIST handwritten digit recognition task in PyTorch. After completing this chapter, you will know: How to load the MNIST dataset using torchvision How to […]

Long Short-Term Memory (LSTM) is a structure that can be used in neural network. It is a type of recurrent neural network (RNN) that expects the input in the form of a sequence of features. It is useful for data such as time series or string of text. In this post, you will learn about LSTM networks. In particular, What is LSTM and how they are different How to develop LSTM network for time series prediction How to train a […]

Recurrent neural network can be used for time series prediction. In which, a regression neural network is created. It can also be used as generative model, which usually is a classification neural network model. A generative model is to learn certain pattern from data, such that when it is presented with some prompt, it can create a complete output that in the same style as the learned pattern. In this post, you will discover how to build a generative model […]

A single-layer neural network, also known as a single-layer perceptron, is the simplest type of neural network. It consists of only one layer of neurons, which are connected to the input layer and the output layer. In case of an image classifier, the input layer would be an image and the output layer would be a class label. To build an image classifier using a single-layer neural network in PyTorch, you’ll first need to prepare your data. This typically involves […]

In the previous session of our PyTorch series, we demonstrated how badly initialized weights can impact the accuracy of a classification model when mean square error (MSE) loss is used. We noticed that the model didn’t converge during training and its accuracy was also significantly reduced. In the following, you will see what happens if you randomly initialize the weights and use cross-entropy as loss function for model training. This loss function fits logistic regression and other categorical classification problems […]