with tf.Graph().as_default() as graph:
with tf.compat.v1.Session() as sess:
tf.constant([imageHeight,imageWidth], dtype="float32",name = "imageSize")
tf.constant([outputTensorName], name = "OutputTensorName")
X = tf.compat.v1.placeholder(tf.float32, [None, imageHeight,imageWidth,1], name='Input')
Y = tf.compat.v1.placeholder(tf.float32, [None, imageHeight,imageWidth,1])
Z = tf.compat.v1.placeholder(tf.float32, [None, 300], name = 'InputLatentVector')
isTrain = tf.compat.v1.placeholder(tf.bool, name="isTrain");
model = Model(X,Y,Z,isTrain,learningRate)
prediction = model.prediction()
error = model.error()
data.add_variable_summary(error, "Loss")
optimizer = model.optimize()
#train_dataset, val_dataset, test_dataset = data.createTensorflowDatasets(0.8,0.1,0.1)
train_dataset, val_dataset = data.createTensorflowDatasets()
merged_summary_operation = tf.compat.v1.summary.merge_all()
modelname = "model_" + "learningRate_" + str(model.learningRate) + "epochs_" + str(epochs) + "_channelsize_" + str(channelSize) + loss
print (modelname)
train_summary_writer = tf.compat.v1.summary.FileWriter(savePath + modelname + '/tmp/' + modelname + "train")
validation_summary_writer = tf.compat.v1.summary.FileWriter(savePath + modelname + '/tmp/' + modelname+ "validation")
init = tf.compat.v1.global_variables_initializer()
init_l = tf.compat.v1.local_variables_initializer()
sess.run(init)
sess.run(init_l)
print("Initialisation completed")
flops = tf.compat.v1.profiler.profile(graph,options=tf.compat.v1.profiler.ProfileOptionBuilder.float_operation())
print('FLOP = ', flops.total_float_ops)
print("Initialisation completed")
trainingClass = train(sess,data,optimizer,error,model,merged_summary_operation)
for epoch in range(epochs):
return int(Datasize/batchSize)
with tf.Graph().as_default() as graph:
with tf.Session() as sess:
tf.constant([imageHeight,imageWidth], dtype="float32",name = "imageSize")
tf.constant([outputTensorName], name = "OutputTensorName")
X = tf.placeholder(tf.float32, [None, imageHeight,imageWidth,1], name='Input')
Y = tf.placeholder(tf.float32, [None, imageHeight,imageWidth,5])
model = Model(X,Y,learningRate)
prediction = model.prediction()
error = model.error()
data.add_variable_summary(error, "Loss")
optimizer = model.optimize()
train_dataset, val_dataset, test_dataset = data.createTensorflowDatasets(0.8,0.1,0.1)
merged_summary_operation = tf.summary.merge_all()
modelname = "model_" + "learningRate_" + str(model.learningRate) + "epochs_" + str(epochs)
print (modelname)
train_summary_writer = tf.summary.FileWriter(savePath + modelname + '/tmp/' + modelname + "train")
validation_summary_writer = tf.summary.FileWriter(savePath + modelname + '/tmp/' + modelname+ "validation")
init = tf.global_variables_initializer()
init_l = tf.local_variables_initializer()
sess.run(init)
sess.run(init_l)
print("Initialisation completed")
trainingClass = train(sess,data,optimizer,error,model,merged_summary_operation)
for epoch in range(epochs):
print ("Current epoch is",epoch)
batches = getnumberofBatches(data.train_size, trainbatchSize)
print ("Number of batches", batches)