The accompanying Tensorflow models and corresponding normalization statistics are contained in models/
and data/
.
Our neural networks have been trained for local resolutions of 256, 266, and 276 equally-spaced nodes along
the Cartesian x- and y- directions.
For further information please contact the author/developer at lal@cs.ucsb.edu.
As of February 2, 2022, this repository has been archived. Please check our latest development on error-correcting neural networks.