Dec 11, 2025

1. 16:01

   Fix: pipeline.instantiate() modifies in-place, use pipeline directly after instantiate

   Thanh-Lam
   81c336f
2. 16:00

   Fix pipeline instantiation: properly handle default_parameters and instantiate

   Thanh-Lam
   71155dc
3. 15:55

   Add clear error message about all required gated models that need terms acceptance

   Thanh-Lam
   b503b1d
4. 15:14

   Upgrade to pyannote.audio 3.3.2 and add better error handling for gated models

   Thanh-Lam
   e169a58
5. 14:42

   Switch to pyannote/speaker-diarization (latest) to avoid deprecated plda parameter

   Thanh-Lam
   88a86dd
6. 14:36

   Improve plda workaround: instantiate SpeakerDiarization class directly with cleaned params

   Thanh-Lam
   29d2dcb
7. 14:04

   Add workaround for deprecated plda parameter in model config, downgrade to pyannote.audio 3.0.1

   Thanh-Lam
   c2b3996
8. 13:53

   Pin pyannote.audio to 3.1.1 for compatibility with speaker-diarization-3.1 model

   Thanh-Lam
   419fa8b
9. 13:50

   Fix Pipeline.from_pretrained() to support both token and use_auth_token parameters

   Thanh-Lam
   d2997d8
10. 12:10

    Update project for Hugging Face Space deployment

    Thanh-Lam
    4193ca8
11. 05:15

    .

    Thanh-Lam
    b924e1d