ur_audio_sub made generating captions easy for any audio files & youtube video using OpenAI Whisper. Multiple languages support.
You can skip all of these introductions by viewing the Video demo here. Let's start generating audio transcription effortlessly.
You can find the demo Jupiter notebook here, make a copy and play with this package on Google Colab to generate your first transcription with only 3 steps and 5 lines of code literally.
Demonstration
You will need to install OpenAI whisper package from source using pip:
Let's runt this command in the terminal fist: !pip install git+https://github.com/openai/whisper.git -q
Install ffmpeg (optional)
In case ffmpeg-python doesn't work, you will need ffmpeg be installed separately to start generating captions:
- on Ubuntu or Debian, or Google Colab:
sudo apt update && sudo apt install ffmpeg - on MacOS using Homebrew:
brew install ffmpeg - on Windows using Chocolatey:
choco install ffmpeg - on Windows using Scoop:
scoop install ffmpeg
Go to the Google Colab menu: Select Runtime > Change runtime type and make sure that GPU has been chosen. You can run this AI model way faster with GPU on Google Colab than the normal CPU or your personal computer.
- Using pip to installed pre-builded package on Pypip
pip install ur_audio_sub - If you want to use the latest pydata_master version instead of the stable one, you can install it from source with the following command:
pip install git+https://github.com/thinh-vu/ur_audio_sub.git@main
(*) You might need to insert a ! before your command when running terminal commands on Google Colab.
Please find it in the same folder where you store the audio file. On Google Colab, you can find it in the root folder when generating subtitles for Youtube videos.
-
Run this command first to import the package to your working environment before using any functions below
from ur_audio_sub import * -
ytSub(link='YOUR_YOUTUBE_URL', model='medium', language='', translate=False, lmt=LMT, basepath=ROOT_DIR)orytSub('YOUR_YOUTUBE_URL')in short -
subGen_path(file_path, model='medium', language='', translate=False)orsubGen_path('YOUR_FILE_PATH')in short -
In case of any issue that happens, you can't generate the Youtube Video caption with
ytSubfunction; try thesubGen_pathfunction with the audio file downloaded path to solving that issue. You can also try this terminal command instead!whisper 'PATH_TO_YOUR_AUDIO_FILE' --model medium.
- This package has been built on top of pytube and OpenAI Whisper:
- pytube is a genuine, lightweight, dependency-free Python library (and command-line utility) for downloading YouTube videos.
- whisper: Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multi-task model that can perform multilingual speech recognition as well as speech translation and language identification.
You can contact me at one of my social network profiles:
If you want to support my open-source projects, you can "buy me a coffee" via Patreon or Momo e-wallet (VN). Your support will help to maintain my blog hosting fee & to develop high-quality content.
