Usage instructions unclear about rootpainter3d HOT 4 OPEN

We agreed a video is a good way to solve this. With an installation guide per OS. Perhaps a video per OS also.

But first we will get the auto-build of the installers working as discussed here: #19

from rootpainter3d.

The client and server are designed to be ran on different machines.

The server is typically ran from source, while the client installers should work OK. At least they used to, and I am sure it wont be hard to get this working again.

I think there are benefits to keeping them separate as in many situations the client machine does not have a suitable GPU and the only way to run the application is via connection to a remote GPU.

So the client build process is not a problem for me using my old scripts, but I'm not sure I totally understand what you have done recently so we might need to work through that together.

Bundling the server into the client installer might be really tricky, as it has a dependency on pytorch and CUDA. I'm also not sure it fits the intended use case.

from rootpainter3d.

I will work on creating some instructions. They will include terminal though. The idea is that teams are composed of people with different abilities. Someone (linux sys admin) can setup the server and get it started and then other users (no command line / terminal knowledge) can just install the client that connects with the server. This is based on what was developed / proposed for https://nph.onlinelibrary.wiley.com/doi/full/10.1111/nph.18387 or preprint: https://www.biorxiv.org/content/10.1101/2020.04.16.044461v1.full

Moving away from this model might not be entirely straight forward, but perhaps there is also scope for a more streamlined installer for users who have high powered deep learning workstations. That would include an application that bundled both client and server in one installer (I get the impression that's what you are looking for?).

from rootpainter3d.

I downloaded the hepatic vessels data from http://medicaldecathlon.com/

I will work on trying to get a model that segments something in this dataset and then try to detail the install process.

from rootpainter3d.