compatibility to latest sklearn, tf, np versions

DEC.py

@@ -12,13 +12,12 @@
 
 from time import time
 import numpy as np
-import keras.backend as K
-from keras.engine.topology import Layer, InputSpec
-from keras.layers import Dense, Input
-from keras.models import Model
-from keras.optimizers import SGD
-from keras import callbacks
-from keras.initializers import VarianceScaling
+import tensorflow.keras.backend as K
+from tensorflow.keras.layers import Layer, InputSpec, Dense, Input
+from tensorflow.keras.models import Model
+from tensorflow.keras.optimizers import SGD
+from tensorflow.keras import callbacks
+from tensorflow.keras.initializers import VarianceScaling
 from sklearn.cluster import KMeans
 import metrics
 
@@ -158,7 +157,7 @@ def on_epoch_end(self, epoch, logs=None):
                                           self.model.get_layer(
                                               'encoder_%d' % (int(len(self.model.layers) / 2) - 1)).output)
                     features = feature_model.predict(self.x)
-                    km = KMeans(n_clusters=len(np.unique(self.y)), n_init=20, n_jobs=4)
+                    km = KMeans(n_clusters=len(np.unique(self.y)), n_init=20)
                     y_pred = km.fit_predict(features)
                     # print()
                     print(' '*8 + '|==>  acc: %.4f,  nmi: %.4f  <==|'
@@ -298,13 +297,13 @@ def fit(self, x, y=None, maxiter=2e4, batch_size=256, tol=1e-3,
         pretrain_epochs = 300
         init = VarianceScaling(scale=1. / 3., mode='fan_in',
                                distribution='uniform')  # [-limit, limit], limit=sqrt(1./fan_in)
-        pretrain_optimizer = SGD(lr=1, momentum=0.9)
+        pretrain_optimizer = SGD(learning_rate=1, momentum=0.9)
     elif args.dataset == 'reuters10k':
         update_interval = 30
         pretrain_epochs = 50
         init = VarianceScaling(scale=1. / 3., mode='fan_in',
                                distribution='uniform')  # [-limit, limit], limit=sqrt(1./fan_in)
-        pretrain_optimizer = SGD(lr=1, momentum=0.9)
+        pretrain_optimizer = SGD(learning_rate=1, momentum=0.9)
     elif args.dataset == 'usps':
         update_interval = 30
         pretrain_epochs = 50

datasets.py

@@ -2,9 +2,9 @@
 
 
 def extract_vgg16_features(x):
-    from keras.preprocessing.image import img_to_array, array_to_img
-    from keras.applications.vgg16 import preprocess_input, VGG16
-    from keras.models import Model
+    from tensorflow.keras.preprocessing.image import img_to_array, array_to_img
+    from tensorflow.keras.applications.vgg16 import preprocess_input, VGG16
+    from tensorflow.keras.models import Model
 
     # im_h = x.shape[1]
     im_h = 224
@@ -93,7 +93,7 @@ def make_reuters_data(data_dir):
 
 def load_mnist():
     # the data, shuffled and split between train and test sets
-    from keras.datasets import mnist
+    from tensorflow.keras.datasets import mnist
     (x_train, y_train), (x_test, y_test) = mnist.load_data()
     x = np.concatenate((x_train, x_test))
     y = np.concatenate((y_train, y_test))
@@ -104,7 +104,7 @@ def load_mnist():
 
 
 def load_fashion_mnist():
-    from keras.datasets import fashion_mnist  # this requires keras>=2.0.9
+    from tensorflow.keras.datasets import fashion_mnist  # this requires keras>=2.0.9
     (x_train, y_train), (x_test, y_test) = fashion_mnist.load_data()
     x = np.concatenate((x_train, x_test))
     y = np.concatenate((y_train, y_test))
@@ -179,7 +179,7 @@ def load_reuters(data_path='./data/reuters'):
         print('making reuters idf features')
         make_reuters_data(data_path)
         print(('reutersidf saved to ' + data_path))
-    data = np.load(os.path.join(data_path, 'reutersidf10k.npy')).item()
+    data = np.load(os.path.join(data_path, 'reutersidf10k.npy'), allow_pickle=True).item()
     # has been shuffled
     x = data['data']
     y = data['label']
@@ -190,8 +190,8 @@ def load_reuters(data_path='./data/reuters'):
 
 
 def load_retures_keras():
-    from keras.preprocessing.text import Tokenizer
-    from keras.datasets import reuters
+    from tensorflow.keras.preprocessing.text import Tokenizer
+    from tensorflow.keras.datasets import reuters
     max_words = 1000
 
     print('Loading data...')
@@ -210,8 +210,8 @@ def load_retures_keras():
 
 
 def load_imdb():
-    from keras.preprocessing.text import Tokenizer
-    from keras.datasets import imdb
+    from tensorflow.keras.preprocessing.text import Tokenizer
+    from tensorflow.keras.datasets import imdb
     max_words = 1000
 
     print('Loading data...')
@@ -246,7 +246,7 @@ def load_newsgroups():
 
 
 def load_cifar10(data_path='./data/cifar10'):
-    from keras.datasets import cifar10
+    from tensorflow.keras.datasets import cifar10
     (train_x, train_y), (test_x, test_y) = cifar10.load_data()
     x = np.concatenate((train_x, test_x))
     y = np.concatenate((train_y, test_y)).reshape((60000,))

metrics.py

@@ -22,6 +22,6 @@ def acc(y_true, y_pred):
     w = np.zeros((D, D), dtype=np.int64)
     for i in range(y_pred.size):
         w[y_pred[i], y_true[i]] += 1
-    from sklearn.utils.linear_assignment_ import linear_assignment
-    ind = linear_assignment(w.max() - w)
-    return sum([w[i, j] for i, j in ind]) * 1.0 / y_pred.size
+    from scipy.optimize import linear_sum_assignment
+    ind = linear_sum_assignment(np.amax(w) - w)
+    return sum([w[i, j] for i, j in zip(*ind)]) * 1.0 / y_pred.size

run_exp.py

@@ -1,8 +1,7 @@
 from DEC import DEC
 import os, csv
-import datasets
-from keras.optimizers import SGD
-from keras.initializers import VarianceScaling
+from tensorflow.keras.optimizers import SGD
+from tensorflow.keras.initializers import VarianceScaling
 import numpy as np
 
 expdir='./results/exp1'
@@ -34,13 +33,13 @@
         pretrain_epochs = 300
         init = VarianceScaling(scale=1. / 3., mode='fan_in',
                                distribution='uniform')  # [-limit, limit], limit=sqrt(1./fan_in)
-        pretrain_optimizer = SGD(lr=1, momentum=0.9)
+        pretrain_optimizer = SGD(learning_rate=1, momentum=0.9)
     elif db == 'reuters10k':
         update_interval = 30
         pretrain_epochs = 50
         init = VarianceScaling(scale=1. / 3., mode='fan_in',
                                distribution='uniform')  # [-limit, limit], limit=sqrt(1./fan_in)
-        pretrain_optimizer = SGD(lr=1, momentum=0.9)
+        pretrain_optimizer = SGD(learning_rate=1, momentum=0.9)
     elif db == 'usps':
         update_interval = 30
         pretrain_epochs = 50