Hey! Great work!
I added a docker config for cuda configurations.
I just noticed you have a dev branch. Please let me know if you'd prefer PRs to dev.

I noticed there is an issue where this configuration will not convert multiple files.

This issue can be reproduced by setting the command: in docker-compose.yml to:

python convert.py /input /output --workers 10 --max 10 --min_length 10000

Then run docker compose run marker ; ls output/
Please note there are no output files.

Thank you for this, @gardner ! Is there any easy way to get this to work with or without a cuda device? I imagine it would need a different base image.

Also, any reason to use 11.8 over 12 for cuda?

Hi - I'm not an active contributor so please take what I say with a grain of salt. I would comment on an Issue if there was one. Please disregard if this isn't the place.

In my particular use case, a Docker container that exposes an API I can repeated call over HTTP would be significantly more helpful and a Docker container used as a sort of command line function call. This makes it useful in a cloud deployment situation. E.g. I could just push this to Google Cloud Run and have a functioning/scaling endpoint to convert PDFs to markdown.

Anyways, I'm going to be writing one for my own use but would be happy to post here when finished.

Anyways, I'm going to be writing one for my own use but would be happy to post here when finished.

@justinmclark, did you get around to doing this? Something like this would also be very useful for me right now!