I have developed a NER pipeline using Spacy, to use in an open source project I am working on. In order to deploy the model, I wrapped it in a Dokcer (while exposing an API) and noticed the Image size is very big (~ 6GB)
When installing the model-
pip install https://huggingface.co/royashcenazi/en_parsigs/resolve/main/en_parsigs-any-py3-none-any.whl
Though the model size is ~500MB there are additional installations of spacy׳s dependencies that result in ~5GB.
Any way to optimize this? adding a link to the model I have deployed in huggingface.