Hi folks, at the end I was able to "integrate" Deep Java Library (djl) with Nextflow and create a nf-plugin (although still in an early stage) !!! In case you want to play with it:

export NXF_PLUGINS_TEST_REPOSITORY="<https://github.com/jagedn/nf-djl/releases/download/0.0.1/nf-djl-0.0.1-meta.json>"
wget <https://raw.githubusercontent.com/jagedn/nf-djl/main/example/main.nf>
(you can edit the <http://main.nf|main.nf> and add/remove images)
./nextflow run <http://main.nf|main.nf> -plugins nf-djl

and you will obtain a list of urls + detected objects

Jorge I have no idea about

djl

.....sounds cool...what does it do ?

Nor me. Just started to play with it. It is a java deep learning library and I have create a plugin to use it in nf By the moment I have create a nf function to detect objects in an image flllowing the tutorial

And this would mean that the Nextflow head job would be able to run/create DL models and then perhaps use them in different processes downstream