Tilde ASR Training platform
Develop and dynamically update your custom ASR/speech-to-text models for voice-enabled applications and IT solutions!
Build custom ASR/speech-to-text models that perfectly meet your needs!
Engineers, researchers, data scientists, and product developers of voice-enabled IT solutions and products always strive to ensure the best speech recognition quality. Overcoming this technological challenge requires custom development and/or domain or use-case adaptation of speech recognition tools by relying on large amounts of training data and complex model training methods.
Tilde ASR Training platform is built to help in solving this problem.
ASR/speech-to-text data collection, processing and model training in a secure cloud platform
- Secure cloud hosting
- Multiple language support
- Easy-to-use data uploading and ASR model training
- Data labeling and curation in the web-based user interface
- Highly scalable cloud service
- Privacy by design
- Multiple ASR/speech-to-text model training recipes
- API for integration in your solutions
This solution is based on R&D results of the COMPRISE project
The COMPRISE project (Cost-effective, Multilingual, Privacy-driven Voice-enabled Services) is a research and innovation action funded by the European Union Horizon 2020 program. It aims to develop the next generation voice interaction technology that will be more affordable, inclusive and, above all, secure.
For more information, see https://www.compriseh2020.eu/