tomaat/services repository overview

Here you can find docker images containing deep learning models accessible through TOMAAT.

If you want to simply use these models, you don't need to get these docker images. These models are already available in the cloud (running on a remote host). Just download a compatible client (for example the 3D slicer client, that you can get from the slicer extension manager) and access the services.

If you want to host these services locally, or extend them or even use them as a reference for your own development, use the images in this registry.

TOMAAT allows you to make deep learning models public by providing them with an interface to communicate with the “outside” world. Find more info on

• https://tomaat.cloud⁠
• https://tomaat.readthedocs.io⁠
• https://github.com/faustomilletari/TOMAAT⁠

In this registry you can find services that are ready to be deployed and can be accessed via TOMAAT.

You can run these services by executing this command: nvidia-docker run -it -p 9001:9001 --rm tomaat/services:tag python serve.py start_service

NOTE: Some services might require a different run command. For example services that are derived from NiftyNet or other model zoo need a different command to run. Refer to TOMAAT documentations to know more.

where the tag can be obtained by visiting https://hub.docker.com/r/tomaat/services/tags/⁠