The tutorials has been very clear on how to use google cloud to run the code that is rooted in localhost:8080/tree. If now I have a model that I want to train with GPU, how should I use google cloud on the code? Say I have mode written in either Jupyter notebook or .py file.
I think you can either create a github repo and clone it to the instance or by using jupyter gui where you can find button UPLOAD. Then once you have a file you can run it either through jupyter(.ipynb files) or terminal(.py files).