Deep Learning Adventures with PyTorch

Understand project’s goals and plan to get there

   •  Define the key objectives for this section

   •  Define steps to finish the project

   •  Get a quick overview of each step

How to quickly use PyTorch components to build classifier.

   •  Learn how to load a pre trained CNN network

   •  Learn how to prepare images for classification with CNN network

   •  Learn how to do classification using prepared images and CNN

Learn how to use ready to use CIFAR100 dataset from PyTorch.

   •  Import the dataset

   •  Learn how to prepare data for our custom CNN network

   •  Load both training and testing set using all of the prepared elements

Learn how build, train and test a simple CNN network for image classification.

   •  Learning how to build a simple CNN model

   •  Learning how to train CNN networks

   •  Learning how to evaluate the model accuracy using test set

Learn how to use a custom CNN model to classify a new image.

   •  Learn how prepare an image for classification

   •  Learn how to perform classification on a new image

   •  Explore ways to improve the accuracy of the model

Understand the project’s goals.

   •  Define the key project objectives

   •  Define the steps to get there

   •  Get a quick overview of each step

Understand our dataset’s format and steps to prepare it.

   •  Explore the dataset

   •  Learn how to read, filter, and clean out the data

   •  Use what you’ve learned to get both training and testing dataset ready

Learn how to build, train, and test a simple RNN network for character level classification.

   •  Build and configure an RNN network

   •  Train the network and understand the training metrics

   •  Evaluate the model using the test set

Learn how to prepare the new data and use the trained model for text classification.

   •  Read the data from the console and prepare it for classification

   •  Define and initialize the trained model from saved state

   •  Perform the classification and interpret the accuracy metrics

Understand what you will specifically learn in this section.

   •  Define project’s goals

   •  Define the process to finish the project

   •  Go quickly through each step

What dataset to use for language translation and what to do to make it ready for use with encoder-decoder networks.

   •  Open the data files and understand the data format

   •  Learn how to format and clean up the data

   •  The idea being vocabulary objects and why do you need them

Learn how to build, configure train and test a GRU-based encoder-decoder network.

   •  Build an encoder-decoder network for language translation

   •  Learn how to train the network

   •  Understand how to quickly evaluate model’s accuracy

How to use your models on a new data?

   •  Initialize and load your models from saved state files

   •  Read the data from console and prepare it to use with your models

   •  Do the translation from one language to the other

Understand the goals of this section.

   •  Define the main objectives for the project

   •  Define the main steps to achieve defined goals

   •  Go through each goal and steps to get there

What kind of images can we use for style transfer and what needs to be done to prepare them for further processing?

   •  Exploring our dataset

   •  Learn the key processing steps to prepare images

   •  Understand the data performance trade offs

How build, train and use an existing CNN network for style transfer.

   •  How to modify a pre trained CNN network to extract “style” features from its layers

   •  How to train this newly created architecture

   •  Understand the variables that influence model’s output and performance