3 ms·

For context, this is the Norwegian National Library [1], which is tasked with archiving and preserving everything that is publicly published or broadcast in Nor

For context, this is the Norwegian National Library [1], which is tasked with archiving and preserving everything that is publicly published or broadcast in Norway. Similar to the U.S. Library of Congress. [1] https://www.nb.no/en/ https://www.nb.no/en/

amelius 2y ago
Do they also perform speech recognition/transcription, while they are at it? If so, what tools are they using?
- torbjornbp 2y ago
  The re-archiving process was mainly a data migration. However, there is an initiative to use the in-house developed NB-whisper model [1] (based on OpenAIs Whisper) to do speech recognition/transcription. [1] https://huggingface.co/collections/NbAiLab/nb-whisper-65cb8322877f943912afcd9f https://huggingface.co/collections/NbAiLab/nb-whisper-65cb83...
  - magicalhippo 2y ago
    That is the model I've been playing with for my homebrew assistant, as mentioned here[1]. The medium and lower models wasn't quite up to the task, miss-interpreting some crucial words here and there, but so far I've been very pleased with the large Q5 model. It directly translates into English well enough that subsequent English LLMs understand the meaning with high degree of accuracy, at about 5-10x real-time on my 2080 Ti. [1]: https://news.ycombinator.com/item?id=41396649 https://news.ycombinator.com/item?id=41396649