This repository implements the neural-network quantum states [1] with Python 3 and Tensorflow 2 library to speed-up the process with graphics processing units (GPU). It also has some applications for finding extreme eigenvalues of a Hermitian matrix.
The implementation is greatly inspired by the old version of NetKet library [2]. For similar library with Tensorflow 1 library with Python 2 (which is not maintained anymore), please see the following repository.
This code is used in [3,4] where we propose several transfer learning protocols to improve the scalability, efficiency, and effectiveness of neural-network quantum states.
This project is based on Python. The code is tested on Python 3.8.5. These are the main library requirements for the project:
tensorflow==2.3.0to run on CPUs ortensorflow-gpu==2.3.0to run on GPUsscipysklearnjupyterto run the notebooks.matplotlibfor plotting purposes.networkxif one wants to define their own graph.
It is also available as requirements.txt in the project and do
pip install -r requirements.txt
to install the necessary libraries.
The different examples to run the code is available in the notebooks directory.
[1] G. Carleo and M. Troyer, Science 355, 602 (2017)
[2] G. Carleo, K. Choo, D. Hofmann, J. E. T. Smith,T. Westerhout, F. Alet, E. J. Davis, S. Efthymiou,I. Glasser, S.-H. Lin, M. Mauri, G. Mazzola, C. B. Mendl,E. van Nieuwenburg, O. O’Reilly, H. Th ́eveniaut, G. Tor-lai, F. Vicentini, and A. Wietek, SoftwareX , 100311(2019).
[3] Zen, R., My, L., Tan, R., Hébert, F., Gattobigio, M., Miniatura, C., Poletti, D., Bressan, S.: Finding quantum critical points with neural-network quantum states. In: ECAI 2020 - 24th European Conference on Artificial Intelligence. Frontiers in Artificial Intelligence and Applications, vol. 325, pp. 1962–1969. IOS Press (2020)
[4] Zen, R., My, L., Tan, R., Hébert, F., Gattobigio, M., Miniatura, C., Poletti, D., Bressan, S.: Transfer learning for scalability of neural-network quantum states. Physical Review E 101(5), 053301 (2020)