https://github.com/twilio-labs/ml-training-api

A service template for asynchronous machine learning model training
https://github.com/twilio-labs/ml-training-api

machine-learning python

Last synced: about 1 year ago
JSON representation

A service template for asynchronous machine learning model training

• Host: GitHub
• URL: https://github.com/twilio-labs/ml-training-api
• Owner: twilio-labs
• License: mit
• Created: 2021-02-05T18:42:32.000Z (over 5 years ago)
• Default Branch: main
• Last Pushed: 2024-11-24T05:01:08.000Z (over 1 year ago)
• Last Synced: 2025-05-07T14:58:37.315Z (about 1 year ago)
• Topics: machine-learning, python
• Language: Python
• Homepage:
• Size: 4.61 MB
• Stars: 16
• Watchers: 2
• Forks: 3
• Open Issues: 19
• Metadata Files:
  • Readme: README.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

README

          
# ml-training-api

[![A Twilio Labs Project](https://img.shields.io/static/v1?label=&message=Twilio-Labs&color=F22F46&labelColor=0D122B&logo=twilio&style=flat-square)](https://www.twilio.com/labs)

This is a simple python service template for training machine learning models asynchronously.

## Falcon + Celery

Example of how to handle training machine learning models asynchronously with Falcon, Celery, and message queues

![diagram](/diagram.png)

### Quick Start

Spin up the containers:

```sh

$ docker-compose up -d --build

```

Open your browser to http://localhost:8000/ping to view the app or to http://localhost:5555 to view the Flower dashboard.

Trigger a new task with hyperparameters (optional):

```sh

$ curl -X POST http://localhost:8000/create \

    -d '{"alpha":0.5}' \

    -H "Content-Type: application/json"

```

Check the status:

```sh

$ curl http://localhost:8000/status/

```

The [model](https://github.com/twilio-labs/ml-training-api/tree/main/src/model) used in this repo is a general text classficiation model built using td-idf features and naive bayes model. But you can plug in any model of your own choice.

## Let's work together

Everything in this toolkit is released under [Twilio Labs](https://www.twilio.com/docs/labs) and fully open-source. If you find any problems with this, [please file an issue](https://github.com/twilio-labs/ml-training-api/issues) or even create a pull request to work together with us on the toolkit. We would love to hear your ideas and feedback!

## License

MIT