What is PuzzleLib?

PuzzleLib is a high level Deep Learning framework with CPU (Intel/AMD) and GPU (NVIDIA/AMD) support. The library is written in Python, it is modular and dynamic.

Documentation

The library documentation is divided into two main sections:

• General info - here you can find introductory information about the library and instructions for installing it;
• Base - this section contains information about all the elements necessary for designing and training your neural network.

Tutorials

The Tutorials section contains tutorials on working with the library. The easiest way to start learning Puzzle lab is with these tutorials.

Available tutorials:

1. Getting started

   1. PuzzleLib in a nutshell
   2. What's the theory behind - automatic differentiation
      • Introduction
      • Restoring a linear function with a single neuron
      • Restoring a linear function by a neural network
      • Restoration of a nonlinear function by a neural network with activation
   3. Training the MNIST classifier
   4. Pausing and resuming network training
   5. Network optimization

2. Application areas

   1. Computer vision
      • Training the CIFAR classifier
   2. Speech processing
      • Speech recognition with Wav2Letter

3. Additional options

   1. Interaction with other frameworks
      • Switching from TensorFlow
      • Switching from PyTorch
      • Converting to TensorRT engine
   2. Contents of the PuzzleLib module on the example of a linear layer
   3. Use of different backends
   4. Parallelize your computing