cross validated results

lcmd-epfl / Q-stack

Stack of codes for dedicated pre- and post-processing tasks for Quantum Machine Learning (QML)

MIT License

13 stars 5 forks source link

cross validated results #53

Closed YAY-C closed 1 month ago

YAY-C commented 1 month ago

[x] added save-file for results; added sub-selection parameters
[x] added cosine kernel
[x] added score (r2) evaluation (maybe not needed since individual predictions are saved)
[x] add explicit random-state to regression
[x] What about the --debug parameter ?
[x] add KWARG to save predictions for each seed
[x] add test for regression, hyperparameters

YAY-C commented 1 month ago

I realize now that we have no tests for the regression/prediction scripts! @briling how would you test them ? on a toy-example or a small dataset included as example ?

briling commented 1 month ago

how would you test them ? on a toy-example or a small dataset included as example ?

just give them any X and Y files. We have something in https://github.com/lcmd-epfl/spahm and https://github.com/lcmd-epfl/spahm-rho

briling commented 1 month ago

do we need the xyz files?

YAY-C commented 1 month ago

do we need the xyz files?

Not necessarily, but I thought we might need them for upcoming tests. Do you think it makes more sense to remove them ?

briling commented 1 month ago

do we need the xyz files?

Not necessarily, but I thought we might need them for upcoming tests. Do you think it makes more sense to remove them ?

Could you please add a test that generates the representation (used in this test) from these xyz files? so we remember they're related

YAY-C commented 1 month ago

do we need the xyz files?

Not necessarily, but I thought we might need them for upcoming tests. Do you think it makes more sense to remove them ?

Could you please add a test that generates the representation (used in this test) from these xyz files? so we remember they're related

[x] add test-data generation script