NLP-Microservice

An example for a Deep Learning NLP model hosted as a RESTful API.

Python Environment

• Conda environment: conda create -n nlp-ms python=3.8
• Dependencies:
  • Run pip install -r ./requirements.txt to install all dependencies (except PyTorch)
  • API Stuff:
    • Flask
    • Flask-RESTX
  • 🤗Transformers is used as NLP library
  • TensorFlow 2.0 or PyTorch
    • I used Pytorch (see here for installation instructions)

Usage

To run the Text-Generation-API simply execute python src/service.py in the root directory.

The script sample_request.py can be used to issue requests to a running Text-Generation-API: python sample_request.py <api-host-or-ip-adress>.

Tutorial for Creating such a Microservice

Please refer to the folder tutorial.