Question about the NN worflow

Good evening everyone,

I started using the NN workflow for fun but as I only have a basic level in coding and everything, I must say this is a bit vague for me.

Is a pre-trained model working like a macro ? At the moment, I must use 4 different workflows to analyze my pictures, would it be possible to gather all the steps into one single pre-trained model where I could load all my data and do everything at once ?

Hello @Antoine_Cure,

a pre-trained network could be indeed looked at as a macro that does a specific task (in our case it’s usually segmentation). The difference to, let’s say a ImageJ macro is, that all processing is learned during training) by looking at lots of examples with correct (hand-generated) results and trying to replicate those as closely as possible.

Currently, we allow only a single network on a GPU at a time. So in order to chain networks, you’d have to work with them one by one, exporting the data from each step as an input to the following one.

Did you find 4 networks, that you want to chain?

Thank you for the Answer Dominik, I didn’t have a look yet but I’ll let you know if something interests me :slight_smile: