This is the implementation for the paper "Alleviating ``Posterior Collapse'' in Deep Topic Models via Policy Gradient".
Since we take the SawETM as an example to demonstrate the building of a dc-ETM, some of the code is based on the "https://github.com/BoChenGroup/SawETM".

For running, just run the main files like "python main_xxx.py".

The code would be public after the paper is released.