Repository contanings the notebooks and codes to reproduce the results presented in the paper:
"Sampling with flows, diffusion and autoregressive neural networks: A spin-glass perspective"
To be able to reproduce the simulations, it is required to have Python 3 installed with the following packages:
- Numpy
- Scipy
- Pandas
The codes for the SE equations, AMP and Sampling are in the folder Codes
Inside the folder Analysis we put the Python code necessary to reproduce our simulations
The folder Data contains all the data frames, and files which are needed to run the notebooks in the folder Analysis
The folder Figures contains all the plots of the paper, obtained directly from the notebooks present in the folder Analysis.