Metadata-Version: 2.1
Name: optimized_transducer
Version: 0.9
Summary: UNKNOWN
Home-page: https://github.com/csukuangfj/optimized_transducer
Author: Fangjun Kuang
Author-email: csukuangfj@gmail.com
License: Apache licensed, as found in the LICENSE file
Platform: UNKNOWN
Classifier: Programming Language :: C++
Classifier: Programming Language :: Python
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Description-Content-Type: text/markdown
License-File: LICENSE

## Introduction

Still working in progress.

This project implements the optimization techniques proposed in
[Improving RNN Transducer Modeling for End-to-End Speech Recognition](https://arxiv.org/abs/1909.12415)
to reduce the memory consumption for computing transducer loss.

During the implementation, we use [torchaudio](https://github.com/pytorch/audio) as a reference.
We ensure that the transducer loss computed using this project is identical to the one computed
using torchaudio when giving the same inputs.


