TheanoLM swMATH ID: 26594 Software Authors: Seppo Enarvi, Mikko Kurimo Description: TheanoLM - An Extensible Toolkit for Neural Network Language Modeling. We present a new tool for training neural network language models (NNLMs), scoring sentences, and generating text. The tool has been written using Python library Theano, which allows researcher to easily extend it and tune any aspect of the training process. Regardless of the flexibility, Theano is able to generate extremely fast native code that can utilize a GPU or multiple CPU cores in order to parallelize the heavy numerical computations. The tool has been evaluated in difficult Finnish and English conversational speech recognition tasks, and significant improvement was obtained over our best back-off n-gram models. The results that we obtained in the Finnish task were compared to those from existing RNNLM and RWTHLM toolkits, and found to be as good or better, while training times were an order of magnitude shorter. Homepage: https://arxiv.org/abs/1605.00942 Source Code: https://github.com/senarvi/theanolm Keywords: Computation and Language; arXiv_cs.CL; Neural and Evolutionary Computing; arXiv_cs.NE; language modeling; artificial neural networks; automatic speech recognition; conversational language Related Software: CSLM; Kaldi; Adam; RNNLM; rwthlm; Theano Cited in: 0 Documents Standard Articles 1 Publication describing the Software Year TheanoLM - An Extensible Toolkit for Neural Network Language Modeling Seppo Enarvi, Mikko Kurimo 2016