Awesome

Compiled Kaldi ASR

This image contains a compiled version of Kaldi-ASR in /opt/kaldi. To shorten the compilation time, only certain shared libraries have been compiled (see Dockerfile for details). Furthermore, no executables have been compiled. To save disk space, all files but .h and .so files have been removed from /opt/kaldi.

This image can be used as a base image to install py-kaldi-asr on top of it.

The image is part of Zamia Speech.

See /opt/kaldi/current-git-commit.txt for the commit in Kaldi's git repo that this image is based on.

Sample STT Service based on Kaldi ASR

Check out docker-py-kaldi-asr-and-model for a sample Docker image that can be used right away for transcription.