Skip to content

Latest commit

 

History

History
93 lines (66 loc) · 3.59 KB

LEARN.md

File metadata and controls

93 lines (66 loc) · 3.59 KB
Raw

GitHub issues GitHub forks GitHub stars GitHub license

Ai Captions

Create captions for videos using AI.

Description

This small project uses OpenAI's whisper AI to generate captions for videos.

Getting Started

Dependencies

  • FFmpeg
  • Python

Installing

  • Clone the repo
  • Install python from here if you have not already installed it.
  • You can install FFmpeg from here or use your package manager to install it.
  • Install the requirements by running py setup.py inside of the project directory.

Executing program

  • If you have not already, run the setup.py file to install the required packages.
python setup.py
  • Then run the main.py with your own arguments.
python main.py -v <video path> -m <whisper model name> -p (-p is optional)
  • Example
python main.py -v video.mp4 -m tiny.en -p

  • Arguments

    • -v or --video - The path to the video file.
    • -m or --model - The name of the model to use. (The list of models can be found below)
    • -p or --preview - If you want to preview the generated video.

  • Models

Size Parameters English-only model Multilingual model Required VRAM Relative speed
tiny 39 M tiny.en tiny ~1 GB ~32x
base 74 M base.en base ~1 GB ~16x
small 244 M small.en small ~2 GB ~6x
medium 769 M medium.en medium ~5 GB ~2x
large 1550 M N/A large ~10 GB 1x

For English-only applications, the .en models tend to perform better, especially for the tiny.en and base.en models. These models are provided by OpenAI. More Info

It may take some time to run for the first time since the AI model needs to be downloaded.

The output of the program will be in the output folder. There will be a video called output.mp4 inside of the folder. There will also be a transcript file called captions.vtt and an audio file called audio.mp3.

Help

Make an Issue to report a bug or request a feature.

Authors

@GodModed

Version History

  • 0.3
    • Using ffmpeg instead of moviepy
    • Increased speed by so much
  • 0.2
    • Added the ability to preview the generated video.
    • Added better argument parsing.
  • 0.1.1
    • Fixed video getting cut off
  • 0.1.0
    • Initial Release

License

This project is licensed under the GNU General Public License v3.0 License.

Acknowledgments