installation issue

[AffDk]

Hi,

I followed the installation procedure and it apparently went without a problem.
Then, I ran
pytest tests

and I got:

`platform linux -- Python 3.7.4, pytest-5.2.1, py-1.8.0, pluggy-0.13.0
rootdir: /home/afsamani/FilterNet
plugins: remotedata-0.3.2, arraydiff-0.3, openfiles-0.4.0, doctestplus-0.4.0
collected 25 items

tests/test_datasets.py . [ 4%]
tests/test_init.py . [ 8%]
tests/test_models.py .EEs [ 24%]
tests/test_train.py F [ 28%]
tests/datasets/test_har.py .... [ 44%]
tests/datasets/test_intention_recognition.py .Killed
`
but when I run the "the scripts in the scripts/ directory", I get an error like the following:

FileNotFoundError: [Errno 2] No such file or directory: '/home/afsamani/FilterNet/filternet/datasets/OpportunityUCIDataset/cache/df_train.df.pkl'

I also looked at the stated folder above and it is apparently empty.

Any suggestion on what could be wrong?

Afshin

[AffDk]

I managed to resolve the abovementioned issue by simply removing the folder "cache" which was empty and re-ran the code so it then generated the file df_train.df.pkl, however, I faced another issue when I continued running it. The code reached line #210 of the file train.py when it ran the function called "sliding_window_x_y", it uses up a lot of RAM so that the kernel resets.
in that function, the variable start_idxs will be assigned 31032 and each window of Xc will have a size of 512 by 113 so the variable X in line 15 will uses up about 53 GB of RAM which my machine does not have.

I could imagine that the dataloader could fetch data lazily and do not make such a big array but I was wondering whether I can simply download the saved model from somewhere without generating them by the code?