from tensorflow.contrib.keras.python.keras import layers
import numpy as np
from tensorflow.contrib.keras.python.keras.models import Model
from tensorflow.contrib.keras.python.keras.optimizers import Adam
# load dataset: image arrays, labels arrays
from prep_data_utils_01_save_load_large_arrays_bcolz_np_pickle_torch import bz_load_array
val_img_array = bz_load_array("/Users/Natsume/Downloads/data_for_all/dogscats/results/val_img_array")
val_lab_array = bz_load_array("/Users/Natsume/Downloads/data_for_all/dogscats/results/val_lab_array")
# load dataset: image iterator and label iterator
from prep_data_02_vgg16_catsdogs_03_img_folder_2_iterators import val_batches
# create an constant tensor
tensor1 = K.constant(value=val_img_array, name='tensor1')
# create a input tensor (constant)
input_tensor = layers.Input(tensor=tensor1, name='input_tensor')
# create a input tensor (placeholder) without knowing num of samples
input_tensor = layers.Input(shape=val_img_array.shape[1:], name='input_tensor')
# create a relu tensor (placeholder) with input tensor
relu_tensor = K.relu(input_tensor)
"""
def relu(x, alpha=0., max_value=None):
Rectified linear unit
With default values, it returns element-wise `max(x, 0)`
Goals: How to use the following
1. K.constant()
2. K.placeholder()
3. Input()
"""
from tensorflow.contrib.keras.python.keras.layers import Input
from prep_data_utils_01_save_load_large_arrays_bcolz_np_pickle_torch import bz_load_array
test_img_array = bz_load_array(
"/Users/Natsume/Downloads/data_for_all/dogscats/results/test_img_array")
from tensorflow.contrib.keras.python.keras import backend as K
input = K.constant(test_img_array, name='test_img_tensor')
"""
Inputs:
- value: scalar or array is a must
- dtype, shape: optional, can be inferred from value or array
- name are optional, but good to have one
Return:
- tensor
def constant(value, dtype=None, shape=None, name=None):
('Creates a constant tensor.\n'
'\n'
'Arguments:\n'
' value: A constant value (or list)\n'
from tensorflow.contrib.keras.python.keras.layers import Input, Dense
from tensorflow.contrib.keras.python.keras import backend as K
import numpy as np
input_tensor = Input(shape=(100, ), name="input_tensor")
inter_tensor = Dense(30, name="my_layer")(input_tensor)
final_tensor = Dense(30, name="final_layer")(inter_tensor)
model = Model(input_tensor, final_tensor) # create the original model
layer_name = 'my_layer'
intermediate_layer_model = Model(inputs=model.input,
outputs=model.get_layer(layer_name).output)
# must compile before predict? No, but must compile before training
input_array1 = np.random.random((1000, 100)) * 9
input_tensor1 = K.constant(value=input_array1)
intermediate_output = intermediate_layer_model.predict(
input_array1) # return array
intermediate_output1 = intermediate_layer_model(
input_tensor1
) # return tensor not array; tensor is no use and a long way to go to reach array
"""
Alternatively, you can build a Keras function that will return the output of a certain layer given a certain input, for example:
"""
from tensorflow.contrib.keras.python.keras import backend as K
# with a Sequential model
get_3rd_layer_output = K.function([model.layers[0].input],
[model.layers[1].output])
layer_output = get_3rd_layer_output([input_array1
def __call__(self, shape, dtype=None): return K.constant(self.value, shape=shape, dtype=dtype)
def __call__(self, shape, dtype=None):
return K.constant(self.value, shape=shape, dtype=dtype)
