3 ms·
Google has papers on device speech recognition, these are used in the keyboard & for live caption on Pixel devices.
by villgax 7y ago
Google has papers on device speech recognition, these are used in the keyboard & for live caption on Pixel devices.
- cjbassi 7y agoThis article from the Google AI blog about the Gboard speech recognition is really interesting: https://ai.googleblog.com/2019/03/an-all-neural-on-device-speech.html?m=1 https://ai.googleblog.com/2019/03/an-all-neural-on-device-sp...
- teapourer 7y agoThey are trained on a ton of non-public data though, and I’m not sure if pre-trained models are around.
- villgax 7y agoNope, they aren't available. CC YouTube videos with captions or radio broadcasts + transcripts could prove helpful for multiple languages as well as being able to create a multilingual ASR.