There is a way to bypass this problem.

• create a databricks repo and include a folder called "Modules"
• download and unpack the whl/tar.gz to a folder (best practice, name the folder as the module itself) of the python module you'd like to install
• push that folder inside Modules folder in the databricks repo
• in your notebook, where you need to use that module, in the first cell add this -> sys.path.append("<full-path-to-Modules>").
• Import your module as usual..

In that way, you wouldn't have to install the wheel every time your cluster starts..