Generating Sequences With Recurrent Neural Networks

Code

• github.com/szcom/rnnlib
  none★ 904
• github.com/grzego/handwriting-generation
  tf★ 586
• github.com/hardmaru/sketch-rnn-datasets
  tf★ 220
• github.com/cody2007/arcane_fortune
  none★ 28
• github.com/altsoph/paranoid_transformer
  pytorch★ 19
• github.com/robertknight/textgen
  pytorch★ 4
• github.com/CambridgeIIS/Gesture-Keyboard-Traj-Gen
  tf★ 3
• github.com/feiwu77777/Handwriting_generation
  tf★ 2
• gitlab.com/raymondhs/char-rnn-truecase
  torch★ 0
• github.com/sherjilozair/char-rnn-tensorflow
  tf★ 0

Abstract

This paper shows how Long Short-term Memory recurrent neural networks can be used to generate complex sequences with long-range structure, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwriting (where the data are real-valued). It is then extended to handwriting synthesis by allowing the network to condition its predictions on a text sequence. The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles.

Tasks

Benchmark Results

Reproductions