Unable to create_cnn due to insufficient rights

create_cnn requires write access on the data source directory, which on Floydhub is the input directory. Unfortunately this is read only, has anyone got a work around?

learn = create_cnn(data, resnet34, metrics=error_rate)

Downloading: "https://download.pytorch.org/models/resnet34-333f7ec4.pth" to /root/.torch/models/resnet34-333f7ec4.pth
100%|██████████| 87306240/87306240 [00:01<00:00, 62466155.89it/s]

OSError: [Errno 30] Read-only file system: '/floyd/input/birds_vs_not_birds/models'

OK got a workaround (not optimal) on the Floydhub forum:

You can copy the dataset into a data directory in the persistent folder ( /floyd/home ), then use this new path ( /floyd/home/data ) in your code.

Note : This is suboptimal (because it breaks one of the well established best practice that we are enforcing), but it will unblock you.