Character baseed BiLSTM-CTF model
We setup the development environment in a Docker container with the following command.
make init
This command gets the resources for training and testing, and then prepares the Docker image for the experiments. After creating the Docker image, you run the following command.
make create-container
The above command creates a Docker container from the Docker image which we create with make init, and then
login to the Docker container. Now we made the development environment. For create and evaluate the model,
you run the following command.
This section shows how we develop with the created Docker container.
Most of the source codes of this project, lstm-crf are stored in the lstm_crf directory.
Generated Docker container mounts the project directory to /work of the container and therefore
when you can edit the files in the host environment with your favorite editor
such as Vim, Emacs, Atom or PyCharm. The changes in host environment are reflected in the Docker container environment.
When we need to add libraries in Dockerfile or requirements.txt
which are added to working environment in the Docker container, we need to drop the current Docker container and
image, and then create them again with the latest setting. To remove the Docker the container and image, run make clean-docker
and then make init-docker command to create the Docker container with the latest setting.
Only the first time you need to create a Docker container, from the image created in make init command.
make create-container creates and launch the lstm_crf container.
After creating the container, you just need run make start-container.
When you logout from shell in Docker container, please run exit in the console.
When you check the code quality, please run make lint
When you run test in tests directory, please run make test
When you want to download data in remote data sources such as Amazon S3 or NFS, sync-from-remote target downloads them.
When you modify the data in local environment, sync-to-remote target uploads the local files stored in data to specified data sources such as S3 or NFS directories.
When you see the status of Docker container, please run make profile in host machine.
To launch Jupyter Notebook, please run make jupyter in the Docker container. After launch the Jupyter Notebook, you can
access the Jupyter Notebook service in https://localhost:8888.
This package was created with Cookiecutter and the cookiecutter-docker-science project template.