I'm deploying an endpoint using a Pytorch model, in the input_fn I have the following:
os.mkdir('outputs')
os.mkdir('outputs/assets')
At the end of the outpu_fn I remove the output
folder
But when invoking the endpoint I get the following error:
OSError: [Errno 30] Read-only file system: 'outputs'
Am I not allowed to create a folder in /opt/ml/model
?