Bolzmann

This is my own implementation to the network described in https://github.com/EverettYou/EFL and the corresponding article (https://arxiv.org/abs/1709.01223). The calculation parts of the Code are taken from the original repository and simply 'translated' to pytorch.

Instead of tensorflow, which is used in the original implementation, in this one pytorch is used.

Limitations

I have not implemented the full network functionality, but the parts I needed and found the most interesting. Therefore, this implementation misses:

• All sample methods besides weighted since the article states 'We use the interval-weighted scheme to sample the entanglement regions and prepare the training data for this study.' (See 'B. Choosing the Central Charge' at https://arxiv.org/abs/1709.01223)