All pull requests

Pull requests

• fix(notebooks): fix automatic model training notebook on modern Colab runtimes

  #359·Oreozisbored opened Oct 21, 2026·

  ·
  11

• Read only the newest raw audio in the streaming melspectrogram

  #356·qwertey6 opened Oct 16, 2026·

  ·

• Count one false positive for a run inside the grouping window

  #353·SashaMIT opened Sep 27, 2026·

  ·

• Custom WakeWord training with recorded samples

  #352·puneetkumar-2005 opened Sep 24, 2026·

  ·

• Fix train.py CLI defaults, optional Piper, and validation batch size

  #351·zain1806481 opened Sep 21, 2026·3 tasks·

  ·
  1

• Change target phrase in model training config

  #347·devintobuisness-lgtm opened Sep 16, 2026·

  ·

• Add native browser port of openWakeWord with onDetection callback

  #339·kousthubraja opened Jun 30, 2026·

  ·

• Fix automatic_model_training.ipynb for Python 3.12 / torchaudio 2.x compatibility

  #326·Tinevagio opened Apr 4, 2026·

  ·
  8

• Fix evaluation of args.convert_to_tflite in train.py

  #308·gtjoseph opened Jan 8, 2026·

  ·
  1

• Add openwakeword/resources to .gitignore

  #307·gtjoseph opened Jan 8, 2026·

  ·

• corrected wrong wake word reference in hey_jarvis.md

  #278·HasselAssel opened Aug 26, 2025·

  ·
  1

• corrected link hey_mycroft.md performance details

  #244·therealadityashankar opened Feb 1, 2025·

  ·

• Make the detect_from_microphone.py example run on Windows

  #237·comodoro opened Jan 19, 2025·

  ·

• Adding LiteRT and more flexibility

  #216·stkr22 opened Nov 16, 2024·

  ·
  4

• fix: allow numpy to extend filesize for larger memmaps.

  #199·maciejmajek opened Aug 30, 2024·

  ·
  1

• Add config option for other language

  #183·AntorFr opened Jul 2, 2024·

  ·
  1

• feat: allow full tensorflow usage if available

  #178·mikejgray opened Jun 9, 2024Contributor·

  ·
  1

  4

• Catch edge-case in train.py when the model may be overfit

  #69·dscripka opened Oct 21, 2023Owner·

  ·
  1