Error on trying to run experiments about cfpq-on-gpgpu

[ciromdrs]

Ps.: I had ran the tests before downloading the files with git lfs and the python interpreter threw an error about not finding triples in the file.
That was true, since I had not downloaded the real content of the files with git lfs yet. No problem about this, so.
But what matters for me is that it indicates that the python interpreter is actually reading the files, and this may serve as debugging information.

[NikitaMishin]

Hello,
About the second question--- practically all implementations use CUDA, moreover, CUDA is only available on Nvidia cards.
I think it should be mentioned in the documentation.
As for the first question, maybe it refers to this problem docker/for-linux#503 ?

[SpirinEgor]

Hi,
First of all, you should generate test description. So, you should check, that in data folder you have real data and run python test_utils/build_testset.py data. After that, you will see tests.csv with
all information about tests.
If you have completed all preparation, run tests via docker run -v /<path to project>/data:/data/ cfpq, notice, that docker needs the full path to your data folder.

[ciromdrs]

Hi

I have managed to get the tests running.
One mistake was to pass the . (dot) instead of the full path.
The other one was in the command. The command that worked for me was
docker run -v /<path to project>:/work/ cfpq
(path to project root, not data; and after the : I put the work directory)
That worked for me.

As expected, my GPU does not support cuda, so the tests that use GPU present empty results.

Thanks for the help.

[SpirinEgor]

I'm glad, that you could run it if you have other questions, feel free to ask.