Obtain a Hugging Face at https://huggingface.co/settings/tokens and add it to config.py for faster summarization inference
If you prefer not to use the Hugging Face API, it is recommended that you run the download_models.py script to locally download the prediction models as to prevent them from being downloaded from the transformers library every time the API is reloaded.
Run the following command in the root repository path to run the api
uvicorn main:app
You can test your token by running the file huggingface_api_test.py
The file summarization.py contains the logic of the api
The file main.py contains the request structure and the route declaration
The file parallel_test.py contains a file to test the performance of the summarization script while running sequentially in comparison to a parallel approach.