- Pull image
docker pull <my-image-summited>, - Setup some environment variables
export <variable-environment>=<value>, following by:- BERT_NAME: name of model pretrain (ex: bert-base-multilingual-cased)
- MODEL_PATH: path of model from "location" (ex: /fastApiProject/bert_regression.pt)
- HOST: target host running (ex: 0.0.0.0 or 127.0.0)
- PORT: target port running (ex: 8000)
- Run docker container:
docker run -it -p $PORT:$PORT -v $MODEL_PATH:/trmanh29_hackathon/output -e MODEL_PATH=/trmanh29_hackathon/output/bert_regression.pt -e HOST=$HOST -e PORT=$PORT <my-image-summited> - Expose api GET:
<my-host>:<my-port>/review-solver/solve?review_sentence=<my-sentence-test>