An open API service indexing awesome lists of open source software.

https://github.com/rhasspy/ar_kaldi-rhasspy

Kaldi profile for Arabic trained from open speech data
https://github.com/rhasspy/ar_kaldi-rhasspy

Last synced: about 2 months ago
JSON representation

Kaldi profile for Arabic trained from open speech data

Awesome Lists containing this project

README

        

# Arabic Kaldi Profile

A [Rhasspy](https://github.com/rhasspy/rhasspy) profile for Arabic (`ar`).

Includes:

* A [Kaldi nnet3](https://kaldi-asr.org/doc/dnn3.html) speech to text model
* See files in `acoustic_model/`
* Recipe created with [ipa2kaldi](https://github.com/rhasspy/ipa2kaldi)
* Word error rate: 7.77%
* Trained on:
* [Common Voice](https://commonvoice.mozilla.org) (77.3 hours)
* An [IPA](https://en.wikipedia.org/wiki/International_Phonetic_Alphabet) pronunciation lexicon from [ipa-dict](https://github.com/open-dict-data/ipa-dict)
* See `base_dictionary.txt.gz`
* A [phonetisaurus](https://github.com/AdolfVonKleist/Phonetisaurus) grapheme to phoneme model for predicting word pronunciations
* See `g2p.fst.gz`
* Trained on `base_dictionary.txt.gz`
* A tri-gram [ARPA language model](https://cmusphinx.github.io/wiki/arpaformat/)
* See `base_language_model.txt.gz`
* Text from audio transcriptions, [Common Voice](https://github.com/mozilla/common-voice/tree/master/server/data/ar), and [Universal Dependencies](https://universaldependencies.org/)