How can I save the data + learn objects on the p2.xlarge instance and re-access them on the free t2.micro instance?

I am using fastai AMI for this lesson. I am able to save the notebook and access the same notebook on both p2.xlarge and t2.micro(free) instances.

On the p2.xlarge instance I ran these lines and got the values for the data + learn objects.

data = ImageClassifierData.from_paths(PATH, tfms=tfms_from_model(resnet34, sz))
learn = ConvLearner.pretrained(resnet34, data, precompute=True)

But wondering how can I save this ‘data’ and ‘learn’ objects somehow,
and then re-load or access them again on the t2.micro instance.

Doing this will allow me to browse thru al the code on the t2.micro free instance.
So is there some way for me to access these objects in all notebooks? Instead of creating it each time?

Thanks

Arjun

1 Like

fastai lib requires CUDA. As t2.micro does not have a GPU, you cant run this code on this instance unless you modify it a bit. There was a thread on the forum where you can find lines of code to change.

2 Likes

Depends upon what you want to re-access , if it’s the predictions , then you can save & download to your local machine and upload - but i suppose this is a very tedious way ,
I’m new to AWS , so i suppose , there may be a role of S3 storage if we want to do everything on AWS itself.

However , if you’ve a fast Internet connection then, for now this is what i can think of , to download relevant information & use it back later on.

You can use FileZilla for ftp.

1 Like

This is true. @arjunrajkumar it’s not possible.

@jeremy I haven’t yet recieved confirmation for p2.xlarge instance limit increase… it’s been a couple of days. Is there anything you can do about this?

No there’s nothing I can do, sorry.

Hi Arjun

maybe this can help