neelkamath/spacy-server repository overview

⁠spaCy Server

This project provides industrial-strength NLP via spaCy⁠ and sense2vec⁠ over a containerized HTTP API. No knowledge of spaCy or sense2vec is required. sense2vec provides phrases similar to the entity (e.g., machine learning and computer vision are related to deep learning). This project is open source on GitHub⁠. You can get help, submit bug reports, and request features on the issue tracker⁠.

⁠Usage

The container EXPOSEs port 8000. You can view the HTTP API docs here⁠.

⁠Newer Images

Images are named according to the format neelkamath/spacy-server:<VERSION>-<MODEL><SENSE2VEC> (e.g., neelkamath/spacy-server:2-en_core_web_sm-sense2vec is an English model for version 2 of the HTTP API with sense2vec enabled, neelkamath/spacy-server:2-fr_core_news_sm is a French model for version 2 of the HTTP API without a sense2vec model). <VERSION> is the HTTP API version (e.g., 2). <MODEL> is the name of the pretrained model⁠ (e.g., en_core_web_sm). SENSE2VEC is -sense2vec if a sense2vec model was bundled, and an empty string otherwise.

To serve at http://localhost:8000, run docker run --rm -p 8000:8000 neelkamath/spacy-server:<VERSION>-<MODEL><SENSE2VEC>.

⁠Version One Images

Images are named according to the format neelkamath/spacy-server:v1-<MODEL> (e.g., neelkamath/spacy-server:v1-en_core_web_sm is an English model for version 1 of the HTTP API). <MODEL> is the name of the pretrained model⁠ (e.g., en_core_web_sm).

To serve at http://localhost:8000, run docker run --rm -e SPACY_MODEL=<MODEL> -p 8000:8000 neelkamath/spacy-server:v1-<MODEL>.

⁠License

This project is under the MIT License⁠.