The deep tensor neural network (DTNN) enables spatially and chemically resolved insights into quantum-mechanical observables of molecular systems.
Requirements:
- python 3.4
- ASE
- numpy
- tensorflow (>=1.0)
See the examples folder for scripts for training and evaluation of a DTNN
model for predicting the total energy (U0) for the GDB-9 data set.
The data set will be downloaded and converted automatically.
Basic usage:
python train_dtnn_gdb9.py -h
If you use deep tensor neural networks in your research, please cite:
K.T. Schütt. F. Arbabzadah. S. Chmiela, K.-R. Müller, A. Tkatchenko.
Quantum-chemical insights from deep tensor neural networks.
Nature Communications 8. 13890 (2017)
doi: 10.1038/ncomms13890