An MLHub package for OpenAI

This MLHUB package provides a command line tool based on whispher for transcribing speech from an audio file. See the MLHub Desktop Survival Guide for details.

Usage

• To install mlhub (Ubuntu 23.10 LTS)

  pip3 install mlhub
  ml configure

• To install and configure the package

  ml install gjwgit/openai@main
  ml configure openai
  ml readme openai
  ml commands openai

• Command line tools

  ml supported openai --iso
  ml identify openai myspeech.wav
  ml transcribe openai myspeech.wav
  ml translate openai myspeech.wav

• Quick test

  wget https://github.com/realpython/python-speech-recognition/raw/master/audio_files/harvard.wav
  ml transcribe openai harvard.wav

to see

The stale smell of old beer lingers. 
It takes heat to bring out the odor. 
A cold dip restores health and zest. 
A salt pickle tastes fine with ham. 
Tacos al pastor are my favorite. 
A zestful food is the hot cross bun.

Options for transcribe/translate

• -l, --lang: Specify the language of the source audio.
  This will speed up the transcribe/translate process.
  See tokenizer.py for the list of all available languages.

• -f, --format: Specify the output format.
  Supported formats are txt, json, srt, tsv, and vtt. Read more about formatting on MLHub Survival Guide.
  For example, -f json will format the output text to the json format.

• -o, --output: Specify the output file name and format.
  For example, -o output.txt will save the output text as the txt format to a file named output.txt in the same directory.

Without -o, then by default, the output text will be printed in the console.

Format:
ml transcribe openai [-l LANGUAGE] [-f FORMAT] [FILENAME]
ml translate openai [-l LANGUAGE] [-o OUTPUT_FILENAME_AND_FORMAT] [FILENAME]

Examples:

ml transcribe openai -l en -f txt myspeech.wav 
ml translate openai --output jokowi.srt --lang id jokowi.wav