Pylearn2 is a machine learning library that has been designed to facilitate research projects. While it is admittedly not very easy to use and demands a good grasp of ML from the user, on the upside, it provides great flexibility to a researcher and is quite fast.
Listed here – in no particular order – are top resources to learn Pylearn2:-
1| FastML’s Pylearn2 in Practice
About: While there are many resources to learn Pylearn2, this blog focuses on aspects of the library that are difficult to pick up on – getting your data in and making predictions out. To get data in, you need to write a Python wrapper class for your dataset, which it provides for, and which can further be used with multiclass sets. The blog has also provided for a hack in order to get predictions produced by Pylearn2.
For more, click here.
2| Pylearn2 Tutorial: Softmax Regression
About: This tutorial is part of Pylearn2’s official documentation to help users navigate this better. It is meant for those who are completely unfamiliar with this library. Led by Ian Goodfellow, who has written other literature on this subject, it will cover the basics of softmax regression, and how it is done in Pylearn2. Divided into two parts, it covers various concepts that will greatly benefit researchers.
For more, click here.
3| YAML For Pylearn2
About: Another tutorial from Pylearn2’s official documentation, the extensive use of YAML by the library makes this a useful resource for anyone interested in exploring its features. YAML is a human-readable dataset serialization scripting language that allows you to completely specify an experiment without the need to write Python code.
For more, click here.
4| Pylearn2: A Machine Learning Research Library
About: A paper released by a group of research scientists including Ian Goodfellow, it carries a summary of the library’s architecture, its brief history, its design philosophy, and an explanation on how the Pylearn2 community functions socially, as well as its goals. It will help you understand if Pylearn2 will help you in your research.
For more, click here.
5| Integrating Pylearn2 And Hyperopt
About: Condensed into a 20-minute video, this tutorial is led by David Warde-Farley, a senior research scientist at DeepMind. It offers a crash course into Pylearn2 and several deep learning concepts and will help you understand how the library can facilitate machine learning research. Unlike other libraries where users do not know how the underlying algorithms work, this is targeted at researchers for whom it is important to understand this in order to accomplish basic data analysis tasks.
For more, click here.
6| Neural Network Example Using Pylearn2
About: Penned by Gustav Arngården, a senior software engineer based out of Sweden, this blog details his experience of using Pylearn2 for a neural network project, and can be a great resource for you to understand the entire lifecycle of an application. Although the example used is a simple one that involves creating a neural network to solve the XOR problem, it will help you understand how to use Pylearn2 as a standalone library.
For more, click here.
