Building the Image

To build the image, please follow the steps below:

1. Run the script automation-scripts/build-image.sh with the following command:

./build-image.sh {model_path} {vocabs_path}

Make sure to provide two arguments:

• {model_path}: the path to the NMT model
• {vocabs_path}: the path to the SentencePiece vocabs model with .spm extension

Starting the Container

After building the image, you can start the container using the following command:

docker run \
-p {port_to_bind}:81 \
-e NUM_ENGINES={num_engines} \
--cpus {cpus}\
--memory {memory} \
cuni-multilingual-translation-service:2.1.0

Make sure to replace the placeholders with appropriate values:

• {port_to_bind}: the port number you want to bind to port 81 in the container
• {num_engines}: the number of translation engines you want to run
• {cpus}: the number of CPUs you want to allocate to the container
• {memory}: the amount of memory you want to allocate to the container

This will start the container with the specified configurations using the cuni-multilingual-translation-service image version 2.1.0. For more details about the API endpoints, please refer to the client-readme.md file.

Contributors

Hashem Sellat
Mateusz Krubiński

License

Unless otherwise stated, our code is released under the MIT License.