Synthetic data generation with CTGAN | GANs for Tabular Synthetic Data Generation with Code in Hindi

[@mukeshrajpurohit5593]

Can we use CTGAN for class balancing technique? Does it have any advantage over SMOTE or SMOTEENN?

[@DataScienceWallah]

Yes, CTGAN can address class imbalance by generating synthetic data, offering advantages such as preserving original data distributions and generating more diverse samples compared to SMOTE.

[@vanshikakejriwal6980]

How can we see the losses of generator and discriminator epoch wise?

[@DataScienceWallah]

Set verbos =True , you will get loss in every epoch

[@vanshikakejriwal6980]

@DataScienceWallah i have done that but it is getting overwritten on the same line. If i want a list of all epochs then what to do?

[@DataScienceWallah]

@@vanshikakejriwal6980 i am getting result for every epoch using same code , try using it on Jupyter or VS code

[@vanshikakejriwal6980]

​@DataScienceWallah when i am running this code in vscode then all the feature graphs are coming in 1 fugure. Each subplot is very small and not fully visible. In google colab mu model is not running fully as my dataset size is large snd it is giving resources constraints.
How can i get each of the graphs clearly in vscode?

[@DataScienceWallah]

@@vanshikakejriwal6980 I recommend utilizing Jupyter Notebook or Jupyter within VSCode for your coding tasks. Additionally, please make sure to include the magic command %matplotlib inline at the beginning of your notebook to display the plots inline within the notebook interface.