Apskaties vai bauskas bibliotēkā ir neredzīgo grāmatas sākam kopīgi ierakstīt (Ēvalds ventspilī)
Uzlikt punctuation jauno modeli uz gramatins API
Summary / Title modelis balstoties uz ranking
Jaunie LV STT (itīpaši, kad būs Tet dati)
Subtitru tulkojumu timestamp padomāt idejas - dod ziņu evaldam
TTS API: https://github.com/asya-ai/tts-asya-api
DB: 194.213.101.153 u: tts_asya p: wOklXt8BOPkmvKOp3TdA
DB structure:
Multiple organizations with different API keys
App execution:
In background worker that takes recordings from DB and processes them by status
For future:
change tempo, emotion, etc.
Needed REST API functions:
add_speaker(api_key, speaker_name) -> speaker_uuid
add_voice_recording(api_key, speaker_uuid, audio_file)
list_speakers(api_key) -> [{ speaker_uuid, voice_recordings: [recording_uuid, recording_uuid, ...] }, ...]
delete_speaker(api_key, speaker_uuid)
delete_voice_recording(api_key, recording_uuid)
generate_audio(api_key, text, speaker_uuid, language) -> task_uuid
status_audio(api_key, task_uuid) -> status
download_audio(api_key, task_uuid) -> audio_file
Dataset:
Krišs atrada STT bugs grāmatas
Neredzīgo biedrību audiogrāmatas
Lielāks, PDFs
Summary
Punctuation uz grāmatiņa? Grāmatiņš uzlikt modeli
Lietuviešu, LT CER -
T5 + Lora
Laikā balstīti likumi
Ranking bulllet points