Repository files navigation

This image contains a demo STT service based onKaldi ASR andpy-kaldi-asr. Try it out by followingthese steps.

To start the STT service on your local machine, execute:

$ docker pull quay.io/mpuels/docker-py-kaldi-asr-and-model:kaldi-generic-en-tdnn_sp-r20180815$ docker run --rm -p 127.0.0.1:8080:80/tcp quay.io/mpuels/docker-py-kaldi-asr-and-model:kaldi-generic-en-tdnn_sp-r20180815

To transfer an audio file for transcription to the service, in a secondterminal, execute:

$ conda env create -f environment.yml$ source activate py-kaldi-asr-client$ ./asr_client.py asr.wav

For a list of available Kaldi models packaged in Docker containers, seehttps://quay.io/repository/mpuels/docker-py-kaldi-asr-and-model?tab=tags .

For a description of the available models, seehttps://github.com/gooofy/zamia-speech#asr-models .

Docker images are named according to the format

kaldi-generic-<LANG>-tdnn-<SIZE>-<RELEASEDATE>

1. <LANG>: There are models for English (en) and German (de).
2. <SIZE>: Kaldi models come in two sizes:sp (standard size) and250 (smaller size, suitable for realtime decoding on Raspberry Pi).
3. <RELEASEDATE>: Usually, models released later are trained on more data andhence have a lower word error rate.

The image is part ofZamia Speech.