From 9818b0b8283951526b9e06ba6ccb6de5ee080d37 Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Fri, 8 Mar 2019 01:07:16 +0000
Subject: Use softmax back again in training lenet

---
 lenet.py | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 4950fe9..6bea8dd 100644
--- a/lenet.py
+++ b/lenet.py
@@ -64,8 +64,7 @@ def get_lenet(shape):
   
   model.add(Dense(units=120, activation='relu'))
   model.add(Dense(units=84, activation='relu'))
-  #model.add(Dense(units=10, activation = 'softmax'))
-  model.add(Dense(units=10, activation = 'relu'))
+  model.add(Dense(units=10, activation = 'softmax'))
   return model
 
 def get_lenet_icp(shape):
-- 
cgit v1.2.3-54-g00ecf


From e58605e30e90bbfcbfd37dcac57e9d97d4c17a85 Mon Sep 17 00:00:00 2001
From: nunzip <np.scarh@gmail.com>
Date: Fri, 8 Mar 2019 02:03:51 +0000
Subject: Add confusion matrix

---
 lenet.py | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 6bea8dd..3440af7 100644
--- a/lenet.py
+++ b/lenet.py
@@ -11,6 +11,7 @@ from tensorflow.keras.metrics import categorical_accuracy
 import numpy as np
 import random
 from sklearn.metrics import accuracy_score
+from sklearn.metrics import confusion_matrix
 from sklearn.model_selection import train_test_split
 
 from classifier_metrics_impl import classifier_score_from_logits
@@ -139,7 +140,7 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   model.save_weights('./weights.h5')
   return model 
 
-def test_classifier(model, x_test, y_true):
+def test_classifier(model, x_test, y_true, conf_mat=False):
   x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
   y_pred = model.predict(x_test)
   logits = tf.convert_to_tensor(y_pred, dtype=tf.float32)
@@ -147,6 +148,13 @@ def test_classifier(model, x_test, y_true):
   y_pred = np.argmax(y_pred, axis=1)
   y_true = np.argmax(y_true, axis=1)
   plot_example_errors(y_pred, y_true, x_test)
+  cm = confusion_matrix(y_true, y_pred)
+  if conf_mat:
+    plt.matshow(cm, cmap='Blues')
+    plt.colorbar()
+    plt.ylabel('Actual')
+    plt.xlabel('Predicted')
+    plt.show()
   return accuracy_score(y_true, y_pred), inception_score
 
 def mix_data(X_train, y_train, X_validation, y_validation, train_gen, tr_labels_gen, val_gen, val_labels_gen, split=0):
-- 
cgit v1.2.3-54-g00ecf


From 035a11b98dc4b78ccab8ddc35a7fceaea9bb00c6 Mon Sep 17 00:00:00 2001
From: nunzip <np.scarh@gmail.com>
Date: Fri, 8 Mar 2019 11:11:41 +0000
Subject: Add PCA representation

---
 lenet.py | 27 +++++++++++++++++++++++++--
 1 file changed, 25 insertions(+), 2 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 3440af7..2ec7196 100644
--- a/lenet.py
+++ b/lenet.py
@@ -13,8 +13,9 @@ import random
 from sklearn.metrics import accuracy_score
 from sklearn.metrics import confusion_matrix
 from sklearn.model_selection import train_test_split
-
+from sklearn.decomposition import PCA
 from classifier_metrics_impl import classifier_score_from_logits
+from sklearn.utils import shuffle
 
 def import_mnist():
   from tensorflow.examples.tutorials.mnist import input_data
@@ -82,6 +83,19 @@ def get_lenet_icp(shape):
     model.add(Dense(units=10, activation = 'relu'))
     return model
 
+def get_lenet_pen(shape):
+    model = keras.Sequential()
+    model.add(Conv2D(filters=6, kernel_size=(3, 3), activation='relu', input_shape=(32,32,1)))
+    model.add(AveragePooling2D())
+      
+    model.add(Conv2D(filters=16, kernel_size=(3, 3), activation='relu'))
+    model.add(AveragePooling2D())
+    model.add(Flatten())
+
+    model.add(Dense(units=120, activation='relu'))
+    model.add(Dense(units=84, activation='relu'))
+    return model
+
 def plot_history(history, metric = None):
   # Plots the loss history of training and validation (if existing)
   # and a given metric
@@ -140,7 +154,7 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   model.save_weights('./weights.h5')
   return model 
 
-def test_classifier(model, x_test, y_true, conf_mat=False):
+def test_classifier(model, x_test, y_true, conf_mat=False, pca=False):
   x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
   y_pred = model.predict(x_test)
   logits = tf.convert_to_tensor(y_pred, dtype=tf.float32)
@@ -155,6 +169,15 @@ def test_classifier(model, x_test, y_true, conf_mat=False):
     plt.ylabel('Actual')
     plt.xlabel('Predicted')
     plt.show()
+  if pca:
+    set_pca = PCA(n_components=2)
+    pca_rep = set_pca.fit_transform(x_test)
+    pca_rep, y_tmp = shuffle(pca_rep, y_tmp, random_state=0)
+    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true, edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('spectral', 10))
+    plt.xlabel('component 1')
+    plt.ylabel('component 2')
+    plt.colorbar();
+  
   return accuracy_score(y_true, y_pred), inception_score
 
 def mix_data(X_train, y_train, X_validation, y_validation, train_gen, tr_labels_gen, val_gen, val_labels_gen, split=0):
-- 
cgit v1.2.3-54-g00ecf


From 54e248d0d28bb6faeaefa8b25195af236ec70150 Mon Sep 17 00:00:00 2001
From: nunzip <np.scarh@gmail.com>
Date: Fri, 8 Mar 2019 11:34:59 +0000
Subject: Fix PCA Bug

---
 lenet.py | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 2ec7196..7c4c48c 100644
--- a/lenet.py
+++ b/lenet.py
@@ -171,12 +171,16 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False):
     plt.show()
   if pca:
     set_pca = PCA(n_components=2)
-    pca_rep = set_pca.fit_transform(x_test)
-    pca_rep, y_tmp = shuffle(pca_rep, y_tmp, random_state=0)
-    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true, edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('spectral', 10))
+    pca_rep = np.reshape(x_test, (x_test.shape[0], x_test.shape[1]*x_test.shape[2]))
+    print(pca_rep.shape)
+    pca_rep = set_pca.fit_transform(pca_rep)
+    print(pca_rep.shape)
+    pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
+    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true[:100], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Spectral', 10))
     plt.xlabel('component 1')
     plt.ylabel('component 2')
     plt.colorbar();
+    plt.show()
   
   return accuracy_score(y_true, y_pred), inception_score
 
@@ -211,4 +215,4 @@ if __name__ == '__main__':
   x_train, y_train, x_val, y_val, x_t, y_t = import_mnist()
   print(y_t.shape)
   model = train_classifier(x_train[:100], y_train[:100], x_val, y_val, epochs=3)
-  print(test_classifier(model, x_t, y_t))
+  print(test_classifier(model, x_t, y_t, pca=True))
-- 
cgit v1.2.3-54-g00ecf


From 4a0b1bd9b6a622007e9e5b5e99f49c3a5249d16b Mon Sep 17 00:00:00 2001
From: nunzip <np.scarh@gmail.com>
Date: Fri, 8 Mar 2019 11:43:11 +0000
Subject: Change PCA color

---
 lenet.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 7c4c48c..39a5cd7 100644
--- a/lenet.py
+++ b/lenet.py
@@ -176,7 +176,7 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False):
     pca_rep = set_pca.fit_transform(pca_rep)
     print(pca_rep.shape)
     pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
-    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true[:100], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Spectral', 10))
+    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true[:100], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
     plt.xlabel('component 1')
     plt.ylabel('component 2')
     plt.colorbar();
-- 
cgit v1.2.3-54-g00ecf


From 97cc07ebc57a813f1fd4b32f314f455c033ab55a Mon Sep 17 00:00:00 2001
From: nunzip <np.scarh@gmail.com>
Date: Fri, 8 Mar 2019 18:30:14 +0000
Subject: Remove get_lenet_pen

---
 lenet.py | 13 -------------
 1 file changed, 13 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 39a5cd7..3d388de 100644
--- a/lenet.py
+++ b/lenet.py
@@ -83,19 +83,6 @@ def get_lenet_icp(shape):
     model.add(Dense(units=10, activation = 'relu'))
     return model
 
-def get_lenet_pen(shape):
-    model = keras.Sequential()
-    model.add(Conv2D(filters=6, kernel_size=(3, 3), activation='relu', input_shape=(32,32,1)))
-    model.add(AveragePooling2D())
-      
-    model.add(Conv2D(filters=16, kernel_size=(3, 3), activation='relu'))
-    model.add(AveragePooling2D())
-    model.add(Flatten())
-
-    model.add(Dense(units=120, activation='relu'))
-    model.add(Dense(units=84, activation='relu'))
-    return model
-
 def plot_history(history, metric = None):
   # Plots the loss history of training and validation (if existing)
   # and a given metric
-- 
cgit v1.2.3-54-g00ecf


From da913f9a4dabab31698669b09b69a215d7947c4e Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 17:01:42 +0000
Subject: Add TSNE and fix PCA

---
 lenet.py        | 31 ++++++++++++++++++++-----------
 report/paper.md |  8 ++++++--
 2 files changed, 26 insertions(+), 13 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 3d388de..3d9ed20 100644
--- a/lenet.py
+++ b/lenet.py
@@ -16,6 +16,7 @@ from sklearn.model_selection import train_test_split
 from sklearn.decomposition import PCA
 from classifier_metrics_impl import classifier_score_from_logits
 from sklearn.utils import shuffle
+from sklearn.manifold import TSNE
 
 def import_mnist():
   from tensorflow.examples.tutorials.mnist import input_data
@@ -141,12 +142,12 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   model.save_weights('./weights.h5')
   return model 
 
-def test_classifier(model, x_test, y_true, conf_mat=False, pca=False):
+def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False):
   x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
-  y_pred = model.predict(x_test)
-  logits = tf.convert_to_tensor(y_pred, dtype=tf.float32)
-  inception_score = tf.keras.backend.eval(classifier_score_from_logits(logits))
-  y_pred = np.argmax(y_pred, axis=1)
+  logits = model.predict(x_test)
+  tf_logits = tf.convert_to_tensor(logits, dtype=tf.float32)
+  inception_score = tf.keras.backend.eval(classifier_score_from_logits(tf_logits))
+  y_pred = np.argmax(logits, axis=1)
   y_true = np.argmax(y_true, axis=1)
   plot_example_errors(y_pred, y_true, x_test)
   cm = confusion_matrix(y_true, y_pred)
@@ -158,16 +159,24 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False):
     plt.show()
   if pca:
     set_pca = PCA(n_components=2)
-    pca_rep = np.reshape(x_test, (x_test.shape[0], x_test.shape[1]*x_test.shape[2]))
-    print(pca_rep.shape)
-    pca_rep = set_pca.fit_transform(pca_rep)
-    print(pca_rep.shape)
+    pca_rep = set_pca.fit_transform(logits)
     pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
-    plt.scatter(pca_rep[:100, 0], pca_rep[:100, 1], c=y_true[:100], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.scatter(pca_rep[:1000, 0], pca_rep[:1000, 1], c=y_true[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
     plt.xlabel('component 1')
     plt.ylabel('component 2')
     plt.colorbar();
     plt.show()
+  if tsne:
+    tsne = TSNE(n_components=2, random_state=0)
+    components = tsne.fit_transform(logits)
+    print(components.shape)
+    components, y_tmp = shuffle(components, y_true, random_state=0)
+    plt.scatter(components[:1000, 0], components[:1000, 1], c=y_true[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.xlabel('component 1')
+    plt.ylabel('component 2')
+    plt.colorbar();
+    plt.show()
+
   
   return accuracy_score(y_true, y_pred), inception_score
 
@@ -202,4 +211,4 @@ if __name__ == '__main__':
   x_train, y_train, x_val, y_val, x_t, y_t = import_mnist()
   print(y_t.shape)
   model = train_classifier(x_train[:100], y_train[:100], x_val, y_val, epochs=3)
-  print(test_classifier(model, x_t, y_t, pca=True))
+  print(test_classifier(model, x_t, y_t, pca=False, tsne=True))
diff --git a/report/paper.md b/report/paper.md
index 53cdb3f..e053353 100644
--- a/report/paper.md
+++ b/report/paper.md
@@ -151,7 +151,7 @@ architectures in Q2.**
 We measure the performance of the considered GAN's using the Inecption score [-inception], as calculated
 with L2-Net logits.
 
-$$ \textrm{IS}(x) = \exp(\mathcal{E}_x \left( \textrm{KL} ( p(y\|x) \|\| p(y) ) \right) ) $$
+$$ \textrm{IS}(x) = \exp(\mathbb{E}_x \left( \textrm{KL} ( p(y\mid x) \| p(y) ) \right) ) $$
 
 ```
 \begin{table}[]
@@ -252,7 +252,11 @@ as most of the testing images that got misclassified (mainly nines and fours) sh
 
 # Bonus
 
-This is an open question. Do you have any other ideas to improve GANs or
+## Relation to PCA
+
+Similarly to GAN's, PCA can be used to formulate **generative** models of a system. While GAN's are trained neural networks, PCA is a definite statistical procedure which perform orthogonal transformations of the data. While both attempt to identify the most important or *variant* features of the data (which we may then use to generate new data), PCA by itself is only able to extract linearly related features. In a purely linear system, a GAN would be converging to PCA. In a more complicated system, we would ndeed to identify relevant kernels in order to extract relevant features with PCA, while a GAN is able to leverage dense and convolutional neural network layers which may be trained to perform relevant transformations.
+
+* This is an open question. Do you have any other ideas to improve GANs or
 have more insightful and comparative evaluations of GANs? Ideas are not limited. For instance,
 
 \begin{itemize}
-- 
cgit v1.2.3-54-g00ecf


From 8a3c9ecfec05ce73f0fe69426bfc97110b6f189f Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 17:39:06 +0000
Subject: Use correct label for colour

---
 lenet.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 3d9ed20..273053d 100644
--- a/lenet.py
+++ b/lenet.py
@@ -161,7 +161,7 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False
     set_pca = PCA(n_components=2)
     pca_rep = set_pca.fit_transform(logits)
     pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
-    plt.scatter(pca_rep[:1000, 0], pca_rep[:1000, 1], c=y_true[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.scatter(pca_rep[:1000, 0], pca_rep[:1000, 1], c=y_tmp[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
     plt.xlabel('component 1')
     plt.ylabel('component 2')
     plt.colorbar();
@@ -171,7 +171,7 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False
     components = tsne.fit_transform(logits)
     print(components.shape)
     components, y_tmp = shuffle(components, y_true, random_state=0)
-    plt.scatter(components[:1000, 0], components[:1000, 1], c=y_true[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.scatter(components[:1000, 0], components[:1000, 1], c=y_tmp[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
     plt.xlabel('component 1')
     plt.ylabel('component 2')
     plt.colorbar();
-- 
cgit v1.2.3-54-g00ecf


From f60cf0650f8a1a4872308edff8ab2deb8503bb76 Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 18:21:29 +0000
Subject: Slightly improve graphs

---
 lenet.py | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 273053d..9b4bb2e 100644
--- a/lenet.py
+++ b/lenet.py
@@ -161,9 +161,9 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False
     set_pca = PCA(n_components=2)
     pca_rep = set_pca.fit_transform(logits)
     pca_rep, y_tmp = shuffle(pca_rep, y_true, random_state=0)
-    plt.scatter(pca_rep[:1000, 0], pca_rep[:1000, 1], c=y_tmp[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
-    plt.xlabel('component 1')
-    plt.ylabel('component 2')
+    plt.scatter(pca_rep[:5000, 0], pca_rep[:5000, 1], c=y_tmp[:5000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.xlabel('Feature 1')
+    plt.ylabel('Feature 2')
     plt.colorbar();
     plt.show()
   if tsne:
@@ -171,9 +171,9 @@ def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False
     components = tsne.fit_transform(logits)
     print(components.shape)
     components, y_tmp = shuffle(components, y_true, random_state=0)
-    plt.scatter(components[:1000, 0], components[:1000, 1], c=y_tmp[:1000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
-    plt.xlabel('component 1')
-    plt.ylabel('component 2')
+    plt.scatter(components[:5000, 0], components[:5000, 1], c=y_tmp[:5000], edgecolor='none', alpha=0.5, cmap=plt.cm.get_cmap('Paired', 10))
+    plt.xlabel('Feature 1')
+    plt.ylabel('Feature 2')
     plt.colorbar();
     plt.show()
 
-- 
cgit v1.2.3-54-g00ecf


From 717da7a7e173cd971baddd9f8edb6de668b1e815 Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 18:30:33 +0000
Subject: Add plot probas function

---
 lenet.py | 8 ++++++++
 1 file changed, 8 insertions(+)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 9b4bb2e..881cfd7 100644
--- a/lenet.py
+++ b/lenet.py
@@ -17,6 +17,7 @@ from sklearn.decomposition import PCA
 from classifier_metrics_impl import classifier_score_from_logits
 from sklearn.utils import shuffle
 from sklearn.manifold import TSNE
+import scikitplot as skplt
 
 def import_mnist():
   from tensorflow.examples.tutorials.mnist import input_data
@@ -142,6 +143,13 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   model.save_weights('./weights.h5')
   return model 
 
+def plot_probas(model, x_test, y_true):
+    probas = model.predict(x_test)
+    skplt.metrics.plot_roc(y_true, probas)
+    plt.show()
+    skplt.metrics.plot_precision_recall_curve(y_true, probas)
+    plt.show()
+
 def test_classifier(model, x_test, y_true, conf_mat=False, pca=False, tsne=False):
   x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
   logits = model.predict(x_test)
-- 
cgit v1.2.3-54-g00ecf


From 07ae3a62f460e8b1ae700194cd85f05b2d2f011d Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 18:46:24 +0000
Subject: Add padding to plot_probas

---
 lenet.py | 1 +
 1 file changed, 1 insertion(+)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 881cfd7..0fe8277 100644
--- a/lenet.py
+++ b/lenet.py
@@ -144,6 +144,7 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   return model 
 
 def plot_probas(model, x_test, y_true):
+    x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
     probas = model.predict(x_test)
     skplt.metrics.plot_roc(y_true, probas)
     plt.show()
-- 
cgit v1.2.3-54-g00ecf


From f9c0139f63438a2574d9931f732cb3aecd172485 Mon Sep 17 00:00:00 2001
From: Vasil Zlatanov <v@skozl.com>
Date: Sun, 10 Mar 2019 19:00:54 +0000
Subject: Fix y_true in plot_probas

---
 lenet.py | 1 +
 1 file changed, 1 insertion(+)

(limited to 'lenet.py')

diff --git a/lenet.py b/lenet.py
index 0fe8277..a94259b 100644
--- a/lenet.py
+++ b/lenet.py
@@ -144,6 +144,7 @@ def train_classifier(x_train, y_train, x_val, y_val, batch_size=128, epochs=100,
   return model 
 
 def plot_probas(model, x_test, y_true):
+    y_true = np.argmax(y_true, axis=1)
     x_test = np.pad(x_test, ((0,0),(2,2),(2,2),(0,0)), 'constant')
     probas = model.predict(x_test)
     skplt.metrics.plot_roc(y_true, probas)
-- 
cgit v1.2.3-54-g00ecf