You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
SDV version: 0.11.0
Python version: 3.6.8
Operating System: Windows 10
Error Description
When applying a UniqueCombinations constraint with handling_strategy = 'transform', the CTGAN model produces some synthetic observations that exhibit nan values for those columns specified within the constraint (despite training the CTGAN model on nan-free data).
This screen capture illustrates the bug:
Steps to reproduce
The full Jupyter Notebook that reproduces this bug is attached (the notebook pulls it's own data from the web): generate_adversarial_data-ctgan.zip
The text was updated successfully, but these errors were encountered:
Environment Details
SDV version: 0.11.0
Python version: 3.6.8
Operating System: Windows 10
Error Description
When applying a
UniqueCombinations
constraint withhandling_strategy = 'transform'
, the CTGAN model produces some synthetic observations that exhibitnan
values for those columns specified within the constraint (despite training theCTGAN
model on nan-free data).This screen capture illustrates the bug:
Steps to reproduce
The full Jupyter Notebook that reproduces this bug is attached (the notebook pulls it's own data from the web):
generate_adversarial_data-ctgan.zip
The text was updated successfully, but these errors were encountered: