Giandomenico Cornacchia

It should be ok now, @gaugup . Maybe I have an older version because I have never changed the imports.

line 51 of file feasible_base_vae.py and line 50 of file feasible_model_approx.py the base_model_dir get error: ``` self.base_model_dir = '../../../dice_ml/utils/sample_trained_models/' ``` i've resolved with: ``` import os self.base_model_dir = ('/').join(os.path.dirname(__file__).split('/')[:-1]+['utils','sample_trained_models','']) ```...

okay, another issue is that the classifier and the adversary network are trained before the nested gradient training in the paper. I'll open a PR and implement a pretraining cycle...

@pronics2004 @nrkarthikeyan i'm trying to pass the continuous integration test for the Pull Request but did not pass the test Continuous Integration / build-py (3.6/3.7/3.8) since the adversary test is...

try to use the Sklearn version instead of the TF version and you will have no problem with pickle dump after the fit. if you want to use the TF...

try to install earlier versions like tensorflow=="2.5.0" or check if the import tensorflow.compat.v1 is imported correctly. if you are interested in a tf2 version of adv.deb. check the forked AIF360...

Can you share the code? i think that some Hyper-parameter blow-up the rating!

Hi @amirhk , I would like to use one of the CF explanation methodologies in the Repo. I'm using MACE, but I see that the CF generation is model-dependent. Exist...