Upload New File

import numpy as np

import h5py

from tensorflow.keras.models import Sequential

from tensorflow.keras.layers import Dense

from tensorflow.keras.optimizers import Adam

from tensorflow.keras import metrics

from tensorflow.keras.layers import Dropout

from tensorflow.keras.layers import BatchNormalization

from data_loader import *

nepochs = 300

n_in = 5

n_out = 1

train_fraction = 0.8

val_fraction = 0.0

#infilename = "/m100_work/IscrC_CD-DLS/simulations/full_dtb_newBH12b_csf3dR_z0.0"

infilename = "/m100_work/IscrC_CD-DLS/simulations/"

rho, T, vx, vy, vz, Bx, By, Bz, B2 = load_data(infilename, [0, 0, 0], [200, 200, 200])

n_cells = rho.size

n_train = n_cells

Xtrain = np.zeros((n_train, n_in))

Ytrain = np.zeros((n_train, n_out))

xxx = rho.flatten()

Xtrain[0:n_train,0] = xxx[0:n_train]

xxx = T.flatten()

Xtrain[0:n_train,1] = xxx[0:n_train]

xxx = vx.flatten()

Xtrain[0:n_train,2] = xxx[0:n_train]

xxx = vy.flatten()

Xtrain[0:n_train,3] = xxx[0:n_train]

xxx = vz.flatten()

Xtrain[0:n_train,4] = xxx[0:n_train]

if n_out == 3:

    xxx = Bx.flatten()

    Ytrain[0:n_train,0] = xxx[0:n_train]

    xxx = By.flatten()

    Ytrain[0:n_train,1] = xxx[0:n_train]

    xxx = Bz.flatten()

    Ytrain[0:n_train,2] = xxx[0:n_train]

if n_out == 1:

    xxx = B2.flatten()

    Ytrain[0:n_train,0] = xxx[0:n_train]

# define the keras model

n_nodes = 400

model = Sequential()

model.add(Dense(n_nodes, input_shape=(n_in,), activation='relu'))

#model.add(Dropout(0.2))

model.add(BatchNormalization())

model.add(Dense(n_nodes, activation='relu'))

#model.add(Dropout(0.2))

model.add(BatchNormalization())

model.add(Dense(n_nodes, activation='relu'))

#model.add(Dropout(0.2))

model.add(BatchNormalization())

model.add(Dense(n_nodes, activation='relu'))

#model.add(Dropout(0.2))

model.add(BatchNormalization())

model.add(Dense(n_nodes, activation='relu'))

#model.add(Dropout(0.2))

model.add(BatchNormalization())

model.add(Dense(n_out, activation='linear'))

model.summary()

print("Model training for number of outputs = ",n_out)

# compile the keras model

opt = Adam(lr=0.0001)

model.compile(loss='mse', optimizer=opt, metrics=[metrics.mae])

# train the network

model.fit(Xtrain, Ytrain, validation_split=0.2, epochs=nepochs, batch_size=500)

# save the trained network

ckptfile = 'models/trained_networ.ckpt'

model.save(ckptfile)