Vystadial 2013 is a dataset of telephone conversations in English and Czech, developed for training acoustic models for automatic speech recognition in spoken dialogue systems. It ships in three parts: Czech data, English data, and scripts.
The data comprise over 41 hours of speech in English and over 15 hours in Czech, plus orthographic transcriptions. The scripts implement data pre-processing and building acoustic models using the HTK and Kaldi toolkits.
This is the scripts part of the dataset. and This research was funded by the Ministry of
Education, Youth and Sports of the Czech Republic under the grant agreement
LK11221.
A tool used to build multilingual corpora from wikipedia. Download the web pages, convert them to plain text, identify language, etc.
A set of 120 corpora collected using this tool is available at https://ufal-point.mff.cuni.cz/xmlui/handle/11858/00-097C-0000-0022-6133-9
WaveSurfer is an Open Source tool for sound visualization and manipulation. It has been designed to suit both novice and advanced users. WaveSurfer has a simple and logical user interface that provides functionality in an intuitive way and which can be adapted to different tasks. It can be used as a stand-alone tool for a wide range of tasks in speech research and education. Typical applications are speech/sound analysis and sound annotation/transcription. WaveSurfer can also serve as a platform for more advanced/specialized applications. This is accomplished either through extending the WaveSurfer application with new custom plug-ins or by embedding WaveSurfer visualization components in other applications.
WikiSpeech is a content management system for the web-based creation of speech databases for the development of spoken language technology and basic research. Its main features are full support for the typical recording, annotation and project administration workflow, easy editing of the speech content, plus a fully localizable user interface. For the creation of a new speech database, it is only necessary to open a new project within WikiSpeech, provide a link to any static project information pages and upload the prompt material to be presented to the speakers. Recordings and annotation are performed via the WWW in a platform independent manner on any Java compatible computer. WikiSpeech currently has been localized to four languages: German, English, Romanian and Russian.
Wmatrix is a corpus comparison and annotation tool. It is web based and incorporates the CLAWS POS tagger and the USAS semantic tagger for English. It also generates frequency lists, concordances, key words and key semantic domains by comparative frequency profiling.
Marian NMT model for Catalan to Occitan translation. It is a multi-task model, producing also a phonemic transcription of the Catalan source. The model was submitted to WMT'21 Shared Task on Multilingual Low-Resource Translation for Indo-European Languages as a CUNI-Contrastive system for Catalan to Occitan.
Marian NMT model for Catalan to Occitan translation. Primary CUNI submission for WMT21 Multilingual
Low-Resource Translation for Indo-European Languages Shared Task.