changes relative to the spliting of mldataset in data

skluc/convert_image_to_records.py

@@ -21,12 +21,9 @@ from __future__ import print_function
 import argparse
 import os
 import sys
-from collections import namedtuple
 
 import tensorflow as tf
 
-from skluc.mldatasets import Cifar10Dataset
-
 from tensorflow.contrib.learn.python.learn.datasets import mnist
 
 FLAGS = None

skluc/examples/fc_nn.py

@@ -7,7 +7,7 @@ where the input comes from memory.
 
 import tensorflow as tf
 import numpy as np
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.tensorflow_.utils import inference_mnist,  batch_generator
 
 tf.logging.set_verbosity(tf.logging.ERROR)

skluc/examples/fc_nn_cifar.py

@@ -7,7 +7,7 @@ where the input comes from memory.
 
 import tensorflow as tf
 import numpy as np
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.tensorflow_.utils import inference_cifar10, batch_generator
 import matplotlib.pyplot as plt
 tf.logging.set_verbosity(tf.logging.ERROR)

skluc/examples/fc_nn_timed.py

 import tensorflow as tf
 import numpy as np
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.tensorflow_.utils import fully_connected, get_next_batch, tf_op
 from skluc.utils import time_fct
 

skluc/examples/tfrecord_nn.py

@@ -9,7 +9,7 @@ the case of tfrecords data source.
 
 import tensorflow as tf
 import numpy as np
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.convert_image_to_records import convert_to
 from skluc.tensorflow_.utils import inference_mnist
 

skluc/examples/write_read_tfrecords.py

@@ -5,7 +5,7 @@ Example (with mnist) on how to read and write tf records.
 import tensorflow as tf
 import os
 import numpy as np
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.convert_image_to_records import convert_to
 import matplotlib.pyplot as plt
 

skluc/keras_/kernel_approximation/nystrom_layer.py

@@ -7,7 +7,7 @@ from skluc.tensorflow_.utils import classification_cifar, batch_generator
 from time import time
 from keras.applications.vgg19 import VGG19
 
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 import numpy as np
 import tensorflow as tf
 
@@ -20,7 +20,8 @@ if __name__ == '__main__':
     cifar.normalize()
     cifar.data_astype(np.float32)
     cifar.labels_astype(np.float32)
-    cifar.resize_zoom(2)
+
+    cifar.rescale(2)
 
     num_epoch = 100
     batch_size = 54

skluc/tensorflow_/kernel_approximation/nystrom_layer.py

@@ -7,7 +7,7 @@ import time as t
 import tensorflow as tf
 import numpy as np
 
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 from skluc.tensorflow_.utils import get_next_batch, classification_mnist, convolution_mnist
 from skluc.tensorflow_.kernel import tf_rbf_kernel
 
@@ -26,7 +26,19 @@ X_val, Y_val = mnist.validation
 X_test, Y_test = mnist.test
 
 
-def nystrom_layer(input_x, input_subsample, output_dim, kernel, **kernel_params):
+def nystrom_layer(input_x, input_subsample, kernel, W_matrix=None, output_dim=0, **kernel_params):
+    """
+    Create the deepstrom layer.
+
+    If output_dim is 0, then there is no learnable W created.
+
+    :param input_x:
+    :param input_subsample:
+    :param output_dim:
+    :param kernel:
+    :param kernel_params:
+    :return:
+    """
     nystrom_sample_size = input_subsample.shape[0]
     with tf.name_scope("nystrom"):
         init_dim = np.prod([s.value for s in input_x.shape[1:] if s.value is not None])
@@ -35,10 +47,16 @@ def nystrom_layer(input_x, input_subsample, output_dim, kernel, **kernel_params)
         with tf.name_scope("kernel_vec"):
             kernel_vector = kernel(h_conv_flat, h_conv_nystrom_subsample_flat, **kernel_params)
 
+        if output_dim != 0:
+            if W_matrix is None:
                 W = tf.get_variable("W_nystrom", [nystrom_sample_size, output_dim],
                                     initializer=tf.random_normal_initializer(stddev=0.1))
+            else:
+                W = tf.Variable(initial_value=W_matrix, trainable=False, name="W_nystrom")
             tf.summary.histogram("W_nystrom", W)
             out_fc = tf.matmul(kernel_vector, W)
+        else:
+            out_fc = kernel_vector
         tf.summary.histogram("output_nystrom", out_fc)
     return out_fc
 

skluc/test/test_kernel.py

@@ -3,7 +3,7 @@ import tensorflow as tf
 import numpy as np
 from sklearn.metrics.pairwise import rbf_kernel, chi2_kernel, additive_chi2_kernel, sigmoid_kernel
 
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 
 from skluc.tensorflow_.kernel import tf_rbf_kernel, tf_chi_square_CPD_exp, tf_chi_square_CPD, tf_sigmoid_kernel
 

skluc/test/test_mldatasets.py

 import unittest
-import skluc.mldatasets as dataset
+import skluc.data.mldatasets as dataset
 import os