Stanford MRNet Competition trouble downloading dataset to gcp

Hi, I see that Stanford just issued a dataset for a competition:

Just finished the week 3 lesson and thought I would try to use the learnings on this. The zipped dataset file is large 5.7 Gig. So, I was hoping that I could load the dataset directly to my-fastai-instance on GCP. I had to sign a submission (not for commercial use etc) and get an email, but when I wget (wget http address for download in the email) the download link in GCP shell I get:

Resolving (…
Connecting to (||:80… connected.
HTTP request sent, awaiting response… 200 OK
Length: 6087523606 (5.7G) [application/zip] Permission denied
Cannot write to ‘’ (Permission denied).

Am I doing something wrong? or can I not use that link outside of my email (since that is where the permission is). I contacted the competition as well. I would prefer not to have to download to my desktop (with my current connect says it will take a few hours) and then upload from my desktop into my-fastai-instance. Thanks, Mark

I hit a similar issue on Windows. I emailed the team but they replied they didn’t know why it would happen.

There is this Google Chrome extension that shows you the curl command to use. After loggin in, use that extension and you would be able to resolve permission denied. The problem is happening because you need to be logged in to download, and when using wget address the permissions are not showed.

Link to extension

1 Like

Thanks, I have not heard back from them yet, will holler with the response

Thanks, not sure what you mean by logging in though. There is no login, only a form that is filled out and then they send an email with the download link in the email.

When downloading from that email use the curl widget, to get the complete curl command to download data with all your login credentials in the curl command.

Thanks for the help, but tried that and when downloading from the link in the email (or pasting into a browser) that widget is grayed out, so cannot get the command you describe.