Is fastai the name of your instance? If so, then I think you should remove the “$”
1 Like
Thank you, Jeremy. I have just sent the first 37 referrals (everybody who filled in the form by now, both fast.ai students and @python_tip followers). I will wait until the decision about those 37 referrals is reached and continue if the most of them get through.
UPDATE (2018-10-28 20:23 PDT): So far, I have sent 110 referrals. All of them should have received an email from Google with the subject “Google Cloud Credit Program - You’ve Been Nominated!”. I still do not know how many people will get through the review phase. (It takes 2-3 days)
UPDATE (2018-10-29 14:39 PDT): First 25 referrals approved, so far nobody was rejected.
UPDATE (2018-10-30 14:40 PDT): Another 28 referrals approved (3 + 25 + 28 in total), so far nobody was rejected.
UPDATE (2018-10-31 23:30 PDT): I have added 54 referrals, thanks for your patience. Another 8 approved (seems to done in batches, results arrive every day at around 2:30pm). In total, 166 referrals sent and received “You’ve been nominated” email. From those applications, 68 have been finalized (“Here’s your credit” email). So far I have no indication that anybody would be rejected.
UPDATE (2018-11-02 19:00 PDT): In total 207 referrals sent (“You’ve been nominated” email) and 103 approved (“Here’s Your Credit!” email). I will keep the form open until Tuesday and then send all the referrals and close it.
UPDATE (2018-11-05 14:13 PDT): Eight referrals have been rejected today (“Application Update” email) including @jeremy. This is the first time that happens. I cannot identify a shared pattern, some of them have a weak justification but others seem solid. I apologize but there is nothing I can do. Another 32 referrals have been accepted today (“Here’s Your Credit!” email). In total, 135 credits given so far. As noted above, I will keep the form open until Tuesday (250 people applied so far).
I would like to thank to everybody reaching to me after receiving GCP credit. It is out to my capability to answer those messages but I appreciate them. Hope GCP will be useful both for fast.ai course and your future projects.
9 Likes
Filled the form. Thanks for helping us out peter
1 Like
Wow ! How very nice of you. Thanks for taking the time to help the community.
On a side note, are you still working on genomic data ? I’ve some rough plans on using bacterial genome data (NGS/illumina) to build some basic prediction models to begin with. Would be neat to team up with somebody with more domain knowledge.
1 Like
Is this a problem with permissions? Try
git config --global user.email "your_email@example.com"
git config --global user.name “your_name”
Hey,
Could you post this great offer in the main v3 thread or notify everyone in here?
This could be a really helpful especially for people like me who’ve created the GCP few years ago and have no credits.
Yes. That was the issue. It was a typo mistake.
Hi, what If I create an instance with a single P100 and a 200GB boot disk and then later want to create an instance with 2 P100s it possible to use the same disk that was allocated to the first instance?
1 Like
when I reach step 4 of the guide (access fastai materials and update packages), git command returns an error “fatal: Not a git repository (or any of the parent directories): .git”, could someone assist me?
Yes, possible. You need to tick the delete protection option to let GCP know not to delete the boot disk while deleting the GPU_instance. Next you can attach the disk while creating another 2 GPU_instance. I would suggest you to use an external disk for data storage purposes.
3 Likes
help, i’m not able to start my instance since yesterday
This error continues to pop up
3 Likes
As the message mentioned, try a different zone and change the GPU and machine type.
which means i have to loose the work that i have done on the previous instance
As I said in the message, you need to read the actual guide - each platform is slightly different. Have a look at my comments again, read the guide, then come back here and let us know if you get it working - and if so, what you did! 
1 Like
The options you described are grayed out, are they only accessible when the instance is active? Because, I too have trouble connecting to my instance atm
Click on the edit option above
hello
is gcp with gpu included in the $300 credits? i did the recommended gcp setup but getting this:
/opt/anaconda3/lib/python3.6/site-packages/torch/nn/functional.py in batch_norm(input, running_mean, running_var, weight, bias, training, momentum, eps)
1371 return torch.batch_norm(
1372 input, weight, bias, running_mean, running_var,
-> 1373 training, momentum, eps, torch.backends.cudnn.enabled
1374 )
1375
RuntimeError: CUDA error: out of memory
try running nvidia-smi from the terminal when connected to your virtual machine. that should show you whether there is a gpu there.
1 Like
oh my, that was silly of me, thanks a lot for your patience!
Has anyone had success transferring files between local machine and VM instance from the command line? I tried to use the gcloud compute scp command (ref: https://cloud.google.com/compute/docs/instances/transfer-files), but get the following error:
(gcloud.compute.scp) Could not fetch resource:
- Insufficient Permission
I am using Windows Subsytem for Linux on my local machine.