notes.txt

train a model:
  sketch_rnn_train --log_root=training_logs --data_dir=datasets/<datasetName> --hparams="data_set=[dataset_filename.npz]"
pip install from requirements text file: 
  pip install -r /path/to/requirements.txt
output current pip dependencies to requirements file:
  pip freeze > requirements.txt
create a virtual environment file:
  python3 -m venv <folder_name>

Start the server:
python3 server.py or
flask run --port <port_to_run_on>

CuDNNLSTM will likely raise some issues if you try to train the sketch_rnn locally with accelerate_LSTM set to True.
It should be fine on CSUA though.

DOCKER:
docker build -t image_name directory <- run this in the same directory as the Dockerfile
docker run -p host_port:forwards_to_docker_port docker-container 

mounting: these do not change the image. Still need to rebuild before deploying
docker run -p host_port:forwards_to_docker_port -v /full_path_to_folder_volume:container_directory_to_mount docker-container