Skip to content

Latest commit

 

History

History
35 lines (30 loc) · 1.55 KB

README.md

File metadata and controls

35 lines (30 loc) · 1.55 KB
Raw

symscribe

A repository for transcribing audio files using Whisper. It provides guidelines for creating chapters at logical topic transitions, with concise headings and respect for timestamps. It performs transcription and export the results as chapters.csv and transcript.csv files.

Installation

See more info at the original Repository.

  1. Install symbolicai
pip install symbolicai[whisper]
  1. Use the builtin sympkg to install the package
sympkg i ExtensityAI/symscribe
  1. Create an alias for the symscribe command
symrun c symscribe ExtensityAI/symscribe

Usage

Supported features:

  • Whisper models: all models supported by symbolicai (default: "base")
    • To change the model, before you run the symscribe do export SPEECH_ENGINE_MODEL="...", where ... is the model name (available).
  • Language: "language=..." (default: "language=en")
  • Export directory: "export_dir=..." (default: "export_dir=.")
  • Transcription format: "transcript_only=..." (default: "transcript_only=True")
    • If transcript_only=True, the script will only export the transcript without creating chapters.
  • Bin size: "bin_size_s=..." (default: "bin_size_s=300")
    • The bin size is the duration of each audio file in seconds when splitting the audio file into smaller chunks.

Example:

symrun symscribe "path_to_file.mp3/mp4/..." "language=en" "export_dir=/tmp" "bin_size_s=300" "transcript_only=True"