Multilang ASR Captioner
A multilingual automatic speech recognition and video captioning tool using faster whisper.
Supports real-time translation to english. Runs on consumer grade cpu.
Requirements and Instalations
Docker (preferred)
You'll need to install docker.
Then, follow the steps below.
1. clone the repo
```{bash}
git clone [email protected]:marquesafonso/multilang-asr-captioner.git
```
2. Build and run the container using docker-compose
```{bash}
docker compose up
```
Check the landing page.
From there you will see the submit_video endpoint and the documentation
Tip: on Linux or Mac localhost will resolve directly to 0.0.0.0 but on windows you will need to change it to 127.0.0.1 or localhost
Local
To run this tool locally on your computer you will need the following sofware installed:
Once you are at your desired working directory, run the following commands on your terminal:
git clone [email protected]:marquesafonso/multilang-asr-captioner.git
pip install pipenv
pipenv install
Note that this assumes a proper Git installation and ssh key configuration.
Quick start (local)
API
A FastAPI API is available. This is the easiest way to use the program locally, akin to docker.
To start the API run:
pipenv run python main.py
Then check the landing page.
From there you will see the submit_video endpoint and the documentation
Tip: on Linux or Mac localhost will resolve directly to 0.0.0.0 but on windows you will need to change it to 127.0.0.1 or localhost