CTGAN
CTGAN copied to clipboard
sdv-dev
Conditional GAN for generating synthetic tabular data.
When doing the TVAE or CTGAN tutorial, one gets a lot of `SettingWithCopyWarning` due to the following [line](https://github.com/sdv-dev/CTGAN/blob/357177806a0523ce94a0b199c0914fd8849e51b3/ctgan/data_transformer.py#L111). I found a bug by deeply looking at the code and trying...
The method `transform` of the `BayesGMMTransformer` is the [following](https://github.com/sdv-dev/RDT/blob/74ffcb72aaddf434b31f09bcbdfcda66457539b0/rdt/transformers/base.py#L232): `def transform(self, data, drop=True):` where `drop` is a bool according to the doc. However, [here](https://github.com/sdv-dev/CTGAN/blob/357177806a0523ce94a0b199c0914fd8849e51b3/ctgan/data_transformer.py#L113) during the transformation we use a...
Normalization of protected-like attributes for TVAE and CTGAN. See #218 for more details.
### Problem Description At the moment it is not possible to see the training progress when training a CTGAN or TVAE model. For CTGAN logging when an epoch has finished...
### Environment Details Please indicate the following details about the environment in which you found the bug: * CTGAN version:0.5.0 * Python version:3.9.7 * Operating System:windows 10 ### Error Description...
The following test is currently an integration test and needs to be changed to a proper unit test: https://github.com/sdv-dev/CTGAN/blob/89c6650dee814c448a45e95393e6ca1d12eb8e51/tests/unit/synthesizer/test_ctgan.py#L265-L296
Not too familiar with pytorch so let me know if this makes sense... It seems like we're instantiating the model each time `fit()` is called https://github.com/DAI-Lab/CTGAN/blob/7aa29685045ffdba84bd87432354c133e05699e6/ctgan/ctgan_model.py#L458-L465 Would it make sense...
Any tips on optimizing performance/training time for the Bayesian Gaussian mixture training phase? Could we consider exposing the parameters and perhaps include sampling the training set? This piece doesn't seem...
See #215 for details. I did not add the test since it is supposed to be captured by one of the tests already. However, we could use the tutorial to...