Blocks and Fuel: Frameworks for deep learning

TL;DR

This paper presents Blocks and Fuel, two Python frameworks designed to simplify training neural networks on large datasets by providing tools for model construction, training, dataset management, and preprocessing.

Contribution

The paper introduces two new frameworks, Blocks and Fuel, that enhance deep learning workflows with modular design, dataset handling, and training utilities based on Theano.

Findings

Facilitates training of complex neural networks with new utilities.

Provides a standard dataset format for large-scale machine learning.

Supports extensive pre-processing and monitoring during training.

Abstract

We introduce two Python frameworks to train neural networks on large datasets: Blocks and Fuel. Blocks is based on Theano, a linear algebra compiler with CUDA-support. It facilitates the training of complex neural network models by providing parametrized Theano operations, attaching metadata to Theano's symbolic computational graph, and providing an extensive set of utilities to assist training the networks, e.g. training algorithms, logging, monitoring, visualization, and serialization. Fuel provides a standard format for machine learning datasets. It allows the user to easily iterate over large datasets, performing many types of pre-processing on the fly.