Skip to content

Latest commit

 

History

History
12 lines (10 loc) · 846 Bytes

README.md

File metadata and controls

12 lines (10 loc) · 846 Bytes
Raw

speech-to-text

Totally hacked together machine learning speech-to-text using IBM's Watson and Python with speaker identification. The script is good to speed up occasional transcription jobs but the output still requires editing.

Full instructions from Caret Juice Marketing.

  1. Install json-to-text.py on your computer,
  2. Enter your Watson speech-to-text API key in console.txt,
  3. Edit (if necessary) the URL in console.txt to point to the region where Watson is running,
  4. Navigate your terminal instance to the folder that you installed json-to-text.py,
  5. Save your input.flac audio file to that same folder,
  6. Copy the console script to your terminal to transcribe the audio,
  7. When the transciption is done, run the Python script to transform Watson's JSON output to text.