def chooseModel(opts):
print('selecting model to test...')
if 'cnn' == opts.model:
if 'onlyIncoming' == opts.mode:
params = cnn.generate_default_onlyIncoming_params(opts.dataType)
else:
params = cnn.generate_default_whole_params(opts.dataType)
modelObj = cnn.CNN(opts, params)
elif 'cudnnLstm' == opts.model:
if 'onlyIncoming' == opts.mode:
params = cudnnLstm.generate_default_onlyIncoming_params(opts.dataType)
else:
params = cudnnLstm.generate_default_whole_params(opts.dataType)
modelObj = cudnnLstm.LSTM(opts, params)
elif 'sae' == opts.model:
if 'onlyIncoming' == opts.mode:
params = sae.generate_default_onlyIncoming_params(opts.dataType)
else:
params = sae.generate_default_whole_params(opts.dataType)
modelObj = sae.SAE(opts, params)
return params, modelObj
def chooseModel(opts, ifTest=True, NUM_CLASS=100, modelPath=''):
if 'cnn' == opts.model:
if 'onlyIncoming' == opts.mode:
params = cnn.generate_default_onlyIncoming_params(opts.dataType)
else:
params = cnn.generate_default_whole_params(opts.dataType)
modelObj = cnn.CNN(opts, params)
elif 'cudnnLstm' == opts.model:
if 'onlyIncoming' == opts.mode:
params = cudnnLstm.generate_default_onlyIncoming_params(opts.dataType)
else:
params = cudnnLstm.generate_default_whole_params(opts.dataType)
modelObj = cudnnLstm.LSTM(opts, params)
elif 'sae' == opts.model:
if 'onlyIncoming' == opts.mode:
params = sae.generate_default_onlyIncoming_params(opts.dataType)
else:
params = sae.generate_default_whole_params(opts.dataType)
modelObj = sae.SAE(opts, params)
if ifTest:
modelObj = modelObj.create_model(NUM_CLASS)
modelObj.load_weights(modelPath)
return modelObj, params
def __init__(self, opts, params, networkName='SAE'):
self.networkName = networkName
self.verbose = opts.verbose
self.sae_model = sae.SAE(opts, params)
mnist = datamodel.nus_wide_class(batch_size)
dims = [500, 128, 32]
graph = tf.Graph()
with graph.as_default():
w_init_func = tf.random_normal
b_init_func = tf.random_normal
def loss_func(x, y):
return 0.5 * tf.reduce_mean(tf.pow(x - y, 2))
reg_fac = 0.1
act_func = tf.sigmoid
sae = sae.SAE(dims, w_init_func, b_init_func, loss_func, act_func, input,
reg_fac)
input = tf.placeholder('float32', [None, dims[0]])
loss = sae.loss_sae(input)
v_error = sae.loss_sae(mnist.valid.images)
optimizer = tf.train.AdamOptimizer(learning_rate=0.01).minimize(loss)
init = tf.global_variables_initializer()
print('model constructed')
# Launch the graph
with tf.Session(graph=graph) as sess:
last_loss = 0
sess.run(init)