PQAI service for transforming data into representations suitable as inputs for machine-learning pipelines. For example, transforming a piece of text into an embedding in a high-dimensional vector space.
The following representations are supported at the moment:
- Text -> Dense embedding
- Text -> Sequence of word vectors
- Text -> Bag of entities
For more detailed information, please refer to PQAI Wiki.
| Method | Route | Description |
|---|---|---|
POST |
/encode |
Returns a representation of given data |
The project is open-source under the MIT license.
We welcome contributions.
To make a contribution, please follow these steps:
- Fork this repository.
- Create a new branch with a descriptive name
- Make copy of env file as .env and docker-compose.dev.yml as docker-compose.yml
- Download and extract the file from
https://s3.amazonaws.com/pqai.s3/public/pqai-assets-latest.zipto assets/ - Bring encoder to life
docker-compose up - Make the changes you want and add new tests, if needed
- Make sure all tests are passing
docker exec -i dev_pqai_encoder_api python -m unittest discover ./tests/ - Commit your changes
- Submit a pull request
Please create an issue if you need help.