https://forums.fast.ai/t/fastai-v0-7-install-issues-thread/24652
- following: https://course.fast.ai/start_gcp.html
- Port forward from sudo user to get jupter lab. port forward from jupyter user gives notebook
- add instructions on installing awscli and boto3 on GCP server
- install Make and gcloud commandline tools
- create Google cloud account and project
- Go to https://console.cloud.google.com/iam-admin/quotas?
- find "Compute Engine API: GPUs (all regions)"
- Check this quota and ask to edit edit quota
- set up gcloud credentials from command line: Auth
- get your kaggle.json file into this directory Instructions
- make sure you agree to any kaggle competitions first
- update makefile with your project, zone, and deployment name, then run
make deploy
- this should be succeed so you see a new machine running, but it will take a bit of time for it to setup and be ready for an SSH connection which is needed to run the next setup command.
make setup
make data
- now you should be able to
make lab
ABLE TO GET ENVIRONMENTS IN JUPYTER LAB AFTER THIS https://ipython.readthedocs.io/en/stable/install/kernel_install.html#kernels-for-different-environments