This is a NLP-NER docker based on huggingface-transformers and Flask.

Load Dataset	Train	Prediction	Evaluate
❌	✔️	✔️	❌

Architecture

docker usage

see https://hub.docker.com/r/genghonghu/docker-ner for more information.

How to use?

Step 1: clone this repository

$ git clone https://github.com/GodHu777777/docker-ner.git
$ cd docker-ner

Step 2: run NLP-NER image

$ docker pull genghonghu/docker-ner:v0.3
$ docker run -it --rm -p 3111:3111 genghonghu/docker-ner:v0.3 # 3111 since flask's port in pod is set to 3111

Step 4: start the server and get using it

$ python3 server.py

After running this command above, there will shows several information about working status, and you can access the given link(default is http://127.0.0.1:5000).

ver Mar 4:

It can be trained by conll2003 and do prediction locally. I am not sure if it is stable in docker container.

ver Mar 5:

Added simple Flask to present NER result through webpage.

ver Mar 10:

Added usage document.

TODO: hyper parameters list

config file

ver Mar 13:

All model(LLM) thing has been transferred into docker container, server.py only do message-transferring work and training(TODO).