UNIQORN -- The Universal Neural-network Interface for Quantum Observable Readout from N-body wavefunctions
|
Variables | |
training_generator | |
validation_generator | |
y_val = np.array(y_val) | |
X_train | |
y_train | |
X_val | |
test_predictions | |
Tweak input parameters to optimize performance of custom NNs. More... | |
UNIQORN.test_predictions |
Tweak input parameters to optimize performance of custom NNs.
MAIN ##########
train (or load already trained) model and validate it
inp.layers=[1024,512,64] inp.regularizations=[0.4,0.2,0.6] ....
UNIQORN.training_generator |
LOADING FILENAMES AND INSTANTIATING DATA GENERATORS #######
Instantiation of the data generators for batch processing with the paths to the files used as data. This construction is needed to load the data and labels on the fly, to not overload the memory for very large data sets.
UNIQORN.validation_generator |
UNIQORN.X_train |
UNIQORN.X_val |
UNIQORN.y_train |
UNIQORN.y_val = np.array(y_val) |