mobiusml / aana_sdk

Aana SDK is a powerful framework for building AI enabled multimodal applications.
https://www.mobiuslabs.com/
Apache License 2.0
26 stars 3 forks source link

Updates for audio related issues #89

Closed Jiltseb closed 5 months ago

Jiltseb commented 5 months ago

Changes include:

  1. Updated faster-whisper version.
  2. Modified deployment file (whisper).
  3. Add silence audio in test cases [https://github.com/mobiusml/aana_sdk/issues/77]
  4. Test audio data class with audio URL [https://github.com/mobiusml/aana_sdk/issues/64]
  5. Gracefully show empty results for videos without audio track [https://github.com/mobiusml/aana_sdk/issues/36]
  6. Adding separate parameter list for batched whisper. [https://github.com/mobiusml/aana_sdk/issues/79]
  7. Added vad finegrained checks [ Finegrained checks are not possible for vad deployment. #78]
  8. Updated cache and test files.