Home

Welcome to the speech-to-text wiki!

Check EC2

This will become deprecated at some point. Info as of Dec 16 2024

1. Find IP address for EC2 instance. It is pinned in slack in #dlss-text-extraction channel.
2. SSH into instance. You will need your public SSH key placed on the server for authentication. Ask for help if needed.
3. speech-to-text repo has been cloned on the instance. It is in the speech-to-text folder.
4. The .env files are placed on the server with specific configuration.
5. The code is updated from Github via a git pull.
6. If you need to build an updated container, look in history for docker build command in order to build the updated container. Command is docker build --tag sul-speech-to-text . Note that you must be in the speech-to-text folder to build the image.
7. Building the container may take a while.
8. docker images shows the built images
9. docker ps shows the running images with container IDs
10. docker logs <CONTAINER ID> shows the logs from the running container. Add a --follow to follow the logs.
11. docker stop <CONTAINER ID> to force stop it from running.
12. docker rmi <IMAGE ID> to remove old images to save space. You may need to remove the container first with docker rm <CONTAINER ID>. docker ps -a to see containers.
13. docker run --rm --detach --env-file .env.qa --gpus all sul-speech-to-text to start it up again.