This is the code for my paper: Adaptive Variational Memory Encoding for Recurrent Data Synthesis. The pdf is available here if you'd like to take a look.
Unfortunately, some parts of the original implementation appear to be broken in recent Pytorch updates and I haven't found time to fix them. However, the sampling script remains intact and works with the available pre-trained models. I'll briefly describe how one might set that up:
- Make sure you have Pytorch, Jupyter, Numpy, and Matplotlib installed wherever you plan to run the sampling script
- Clone the repo and start up a Jupyter notebook server
- Navigate to the
sample.ipynb
Jupyter notebook - Execute the single cell in the notebook and outputs should be displayed
If you have any questions regarding the paper or code feel free to let me know.