verbose = 0
bSize = 850
fntrain = "/home/philgun/Documents/PhD/Modelica/receiver-data/training_data_cascade_constant_size_AR_clean.csv"
fntest = "/home/philgun/Documents/PhD/Modelica/receiver-data/validation_data_cascade_constant_size_AR_const.csv"
wd = "./multi_aperture_6"
#MinMax or StandardScaler
scaler = "MinMax"
if not os.path.exists(wd):
os.makedirs(wd)
print(len(pd.read_csv(fntrain)))
arr = lib.preprocessing(wd, fntrain, fntest, 13, 1, scaling_method=scaler)
#Print X and y train raw - to check whether training and test data have been parsed correctly
print("X train raw :\n", arr[-3])
print("y train :\n", arr[-2])
#Print X and y test raw - to check whether training and test data have been parsed correctly
print("X test raw :\n", arr[-1])
print("y test raw :\n", arr[5])
#******************************* Test build model
model = lib.generate_model(arr, dropout=0.1, nPercent=0.25, nShrink=0.9)
model.summary()
#******************************* Partially initialise the eval_net func
objfunc = functools.partial(lib.eval_net, wd, verbose, scaler, arr, bSize)
import time
import functools
from matplotlib import pyplot as plt
import bayesian as lib
if __name__ == "__main__":
verbose = 0
bSize = 850
fntrain = "/home/philgun/Documents/PhD/Modelica/receiver-data/training_data_constant_AR_H_drop_T_out.csv"
fntest = "/home/philgun/Documents/PhD/Modelica/receiver-data/validation_data_constant_AR_H_drop_T_out.csv"
wd = "./single_aperture_constant_T_out_2"
arr = lib.preprocessing(wd,fntrain,fntest,7,1)
print(arr[-2])
print(arr[-1])
print(arr[5])
#******************************* Test build model
model = lib.generate_model(arr,dropout=0.1,nPercent=0.25,nShrink=0.9)
model.summary()
objfunc = functools.partial(
lib.eval_net,
wd,
verbose,
arr,
#************************************* Create dir w.r.t. mode
if not os.path.exists(wd):
os.makedirs(wd)
#************************************* Prep data
data = utils.Data("./data/LC_model_temp_control.mat")
df = data.get_data()[3:].drop(columns="time")
split_df(df, fraction=0.3)
scaler = "MinMax"
inputsize = df.shape[1] - 1
outputsize = 1
bSize = 64
#*************************************** Generate training test data set and also scalers etc.
arr = lib.preprocessing(wd, fntrain, fntest, inputsize, outputsize, scaler)
#Print X and y train raw - to check whether training and test data have been parsed correctly
print("X train raw :\n", arr[-3])
print("y train raw:\n", arr[-2])
#Print X and y test raw - to check whether training and test data have been parsed correctly
print("X test raw :\n", arr[-1])
print("y test raw :\n", arr[5])
#************************************** Function test
model = lib.generate_model(arr, dropout=0.1, nPercent=0.25, nShrink=0.9)
model.summary()
#************************************** Partially initialise the objective function
objfunc = functools.partial(lib.eval_net,