This repository is essentially the same as this repository.
The main contents of this repository are the following below:
- This code has been rewritten using pytorch instead of theano.
- This code has been rewritten so that you can read almost all variables.
The thesne_using_theano has exactly the same contents as original thesne just by rewriting the variables.
The plotted figures in result/gaussian_figure are the experimental results obtained by performing the same experiment as the original paper using pytorch, and you can see that the code has been rewritten normally.
When you run examples/gaussians.py, you hove to change some variables, dir_path and save_path, to suit your environment.
You can select your gpu device as os.environ["CUDA_VISIBLE_DEVICES"].
This code automatically run on cpu without you need to rewrite if you can't use gpu, but it is recommended to use gpu in terms of calculation time.
Nothing has been rewritten this time, so if you want to run this code, you need to rewrite it yourself.