HTTP-Model

An HTTP server for deep learning models and natural language processing.

HTTP-Model is a microservice for your natural language processing models that can be updated dynamically via REST and JSON.

This is currently a work in progress. Below are the list of supported algorithms.

Supported Algorithms:

Naive Bayes

Usage

Data sets are represented by JSON objects. Below is an example of JSON that would be accepted by the system.

{
  "Data" : 
  [
    [
      "Hello how are you doing today?",
      "salutation"
    ],
    [
      "See you later and have a great day.",
      "valediction"
    ]
  ]
}

To create a model run the bin/create_model.sh script.

$ bin/create_model.sh <json_file> <model_name>

To update a model run the bin/update_model.sh script.

$ bin/update_model.sh <json_file> <model_name>

To delete a model run the bin/delete_model.sh script.

$ bin/delete_model.sh <model_name>

Disclaimer

HTTP-Model is still a work in progress with limited functionality. It is not ready to run in production.

Contributing

If you would like to contribute to this project send a pull request.

License