Hi all, I see recently a proliferation of papers that not only release the code and some supplementary results in a github repo, but have complete sites with interactive demos.

Examples: http://gannotsee.net/, http://gandissect.res.ibm.com/ganpaint.html

To me, these seem like a difficult exercise that not all students / labs, have access to. Unless there is a quick tool/pipeline that people use.

Are all these websites dynamic and have server side stuff to run the model?
or is there a simpler way to do this browser (client-side)?

I guess the answer depends on the type of demo (e.g. here everything is pre-done: https://dmitryulyanov.github.io/deep_image_prior), but I am just wondering if there is a tool etc to make the whole process easier.

To clarify: I am aware of tensorflow-js, and in browser running of models, but I am referring to a bigger picture.