Use softmax

Author: GraemeMalcolm

05b - Convolutional Neural Networks (Tensorflow).ipynb

@@ -171,7 +171,7 @@
     "\n",
     "# Now we'll flatten the feature maps and generate an output layer with a predicted probability for each class\n",
     "model.add(Flatten())\n",
-    "model.add(Dense(train_generator.num_classes, activation='sigmoid'))\n",
+    "model.add(Dense(train_generator.num_classes, activation='softmax'))\n",
     "\n",
     "# With the layers defined, we can now compile the model for categorical (multi-class) classification\n",
     "model.compile(loss='categorical_crossentropy',\n",