/adaptive-softmax

Implements an efficient softmax approximation as described in the paper "Efficient softmax approximation for GPUs" (http://arxiv.org/abs/1609.04309)

Primary LanguageLuaOtherNOASSERTION

Adaptive Softmax

The adaptive-softmax project is a Torch implementation of the efficient softmax approximation for graphical processing units (GPU), described in the paper "Efficient softmax approximation for GPUs" (http://arxiv.org/abs/1609.04309).

This module has been developped by Facebook Research. Base repository is at https://github.com/facebookresearch/adaptive-softmax.

Installation

To use this module, clone this repository in OpenNMT/onmt/modules directory.

License

adaptive-softmax is BSD-licensed. Facebook also provides an additional patent grant - check LICENSE and PATENT file in this diretory.