diff --git a/Example 1 - a pitch detection network with Dense layers.ipynb b/Example 1 - a pitch detection network with Dense layers.ipynb
index 155aa82..87f22f8 100644
--- a/Example 1 - a pitch detection network with Dense layers.ipynb	
+++ b/Example 1 - a pitch detection network with Dense layers.ipynb	
@@ -15,9 +15,7 @@
   {
    "cell_type": "code",
    "execution_count": 31,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [],
    "source": [
     "%matplotlib inline\n",
@@ -122,7 +120,6 @@
    "cell_type": "code",
    "execution_count": 35,
    "metadata": {
-    "collapsed": false,
     "scrolled": false
    },
    "outputs": [
@@ -205,7 +202,6 @@
    "cell_type": "code",
    "execution_count": 9,
    "metadata": {
-    "collapsed": false,
     "scrolled": false
    },
    "outputs": [],
@@ -231,9 +227,7 @@
   {
    "cell_type": "code",
    "execution_count": 10,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -269,7 +263,6 @@
    "cell_type": "code",
    "execution_count": 11,
    "metadata": {
-    "collapsed": false,
     "scrolled": true
    },
    "outputs": [
@@ -496,7 +489,6 @@
    "cell_type": "code",
    "execution_count": 34,
    "metadata": {
-    "collapsed": false,
     "scrolled": false
    },
    "outputs": [
@@ -549,9 +541,7 @@
   {
    "cell_type": "code",
    "execution_count": 13,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -576,9 +566,7 @@
   {
    "cell_type": "code",
    "execution_count": 14,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -597,9 +585,7 @@
   {
    "cell_type": "code",
    "execution_count": 33,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -663,7 +649,6 @@
    "cell_type": "code",
    "execution_count": 32,
    "metadata": {
-    "collapsed": false,
     "scrolled": false
    },
    "outputs": [
@@ -727,9 +712,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.11"
+   "version": "2.7.13"
   }
  },
  "nbformat": 4,
- "nbformat_minor": 0
+ "nbformat_minor": 1
 }
diff --git a/Example 2 - a chord recognition network with Convolutional layers.ipynb b/Example 2 - a chord recognition network with Convolutional layers.ipynb
index 6029878..de51154 100644
--- a/Example 2 - a chord recognition network with Convolutional layers.ipynb	
+++ b/Example 2 - a chord recognition network with Convolutional layers.ipynb	
@@ -4,7 +4,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# A very simple chord recognition convnet\n",
+    "# Example 2 - A very simple chord recognition convnet\n",
     "\n",
     "We're gonna use synthesize data and use CQT as representation. "
    ]
@@ -12,19 +12,13 @@
   {
    "cell_type": "code",
    "execution_count": 1,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stderr",
      "output_type": "stream",
      "text": [
-      "Using TensorFlow backend.\n",
-      "/Users/gnu/anaconda/lib/python2.7/site-packages/pandas/core/computation/__init__.py:18: UserWarning: The installed version of numexpr 2.4.4 is not supported in pandas and will be not be used\n",
-      "The minimum supported version is 2.4.6\n",
-      "\n",
-      "  ver=ver, min_ver=_MIN_NUMEXPR_VERSION), UserWarning)\n"
+      "Using TensorFlow backend.\n"
      ]
     }
    ],
@@ -166,9 +160,7 @@
   {
    "cell_type": "code",
    "execution_count": 5,
-   "metadata": {
-    "collapsed": false
-   },
+   "metadata": {},
    "outputs": [],
    "source": [
     "datagen = DataGen()\n",
@@ -184,17 +176,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 7,
    "metadata": {
-    "collapsed": false,
-    "scrolled": true
+    "scrolled": false
    },
    "outputs": [
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArcAAAD8CAYAAABtnYzxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAsUAAAGfCAYAAAC3Nq8UAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlAAAAABJRU5ErkJggg==\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x1199be6d0>"
+       "<matplotlib.figure.Figure at 0x11c6053d0>"
       ]
      },
      "metadata": {},
@@ -202,7 +193,7 @@
     }
    ],
    "source": [
-    "plt.figure(figsize=(14, 4))\n",
+    "plt.figure(figsize=(14, 7))\n",
     "if K.image_data_format == 'channels_first':\n",
     "    for i in range(4):\n",
     "        plt.subplot(1, 10, i+1)\n",
@@ -233,7 +224,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
+   "execution_count": 8,
    "metadata": {
     "collapsed": true
    },
@@ -252,7 +243,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
+   "execution_count": 9,
    "metadata": {
     "collapsed": true
    },
@@ -276,10 +267,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 10,
+   "metadata": {},
    "outputs": [],
    "source": [
     "np.random.seed(12345)\n",
@@ -299,10 +288,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 11,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -341,9 +328,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
+   "execution_count": 12,
    "metadata": {
-    "collapsed": false,
     "scrolled": true
    },
    "outputs": [
@@ -352,205 +338,211 @@
      "output_type": "stream",
      "text": [
       "Epoch 1/100\n",
-      "100/100 [==============================] - 0s - loss: 0.6720 - acc: 0.6387 - val_loss: 0.6337 - val_acc: 0.9004\n",
+      "100/100 [==============================] - 1s - loss: 0.6721 - acc: 0.6398 - val_loss: 0.6344 - val_acc: 0.8633\n",
       "Epoch 2/100\n",
-      "100/100 [==============================] - 0s - loss: 0.5632 - acc: 0.9853 - val_loss: 0.4632 - val_acc: 1.0000\n",
+      "100/100 [==============================] - ETA: 0s - loss: 0.5667 - acc: 0.980 - 0s - loss: 0.5637 - acc: 0.9814 - val_loss: 0.4638 - val_acc: 1.0000\n",
       "Epoch 3/100\n",
-      "100/100 [==============================] - 0s - loss: 0.3368 - acc: 1.0000 - val_loss: 0.2189 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.3380 - acc: 1.0000 - val_loss: 0.2163 - val_acc: 1.0000\n",
       "Epoch 4/100\n",
-      "100/100 [==============================] - 0s - loss: 0.1512 - acc: 1.0000 - val_loss: 0.0993 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.1511 - acc: 1.0000 - val_loss: 0.1013 - val_acc: 1.0000\n",
       "Epoch 5/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0759 - acc: 1.0000 - val_loss: 0.0561 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0759 - acc: 1.0000 - val_loss: 0.0587 - val_acc: 1.0000\n",
       "Epoch 6/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0462 - acc: 1.0000 - val_loss: 0.0372 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0462 - acc: 1.0000 - val_loss: 0.0374 - val_acc: 1.0000\n",
       "Epoch 7/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0315 - acc: 1.0000 - val_loss: 0.0250 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0319 - acc: 1.0000 - val_loss: 0.0278 - val_acc: 1.0000\n",
       "Epoch 8/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0247 - acc: 1.0000 - val_loss: 0.0227 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0242 - acc: 1.0000 - val_loss: 0.0184 - val_acc: 1.0000\n",
       "Epoch 9/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0185 - acc: 1.0000 - val_loss: 0.0158 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0185 - acc: 1.0000 - val_loss: 0.0154 - val_acc: 1.0000\n",
       "Epoch 10/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0156 - acc: 1.0000 - val_loss: 0.0144 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0159 - acc: 1.0000 - val_loss: 0.0133 - val_acc: 1.0000\n",
       "Epoch 11/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0130 - acc: 1.0000 - val_loss: 0.0103 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0128 - acc: 1.0000 - val_loss: 0.0110 - val_acc: 1.0000\n",
       "Epoch 12/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0109 - acc: 1.0000 - val_loss: 0.0091 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0108 - acc: 1.0000 - val_loss: 0.0119 - val_acc: 1.0000\n",
       "Epoch 13/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0099 - acc: 1.0000 - val_loss: 0.0107 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0099 - acc: 1.0000 - val_loss: 0.0091 - val_acc: 1.0000\n",
       "Epoch 14/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0089 - acc: 1.0000 - val_loss: 0.0085 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0088 - acc: 1.0000 - val_loss: 0.0110 - val_acc: 1.0000\n",
       "Epoch 15/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0080 - acc: 1.0000 - val_loss: 0.0073 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0079 - acc: 1.0000 - val_loss: 0.0074 - val_acc: 1.0000\n",
       "Epoch 16/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0073 - acc: 1.0000 - val_loss: 0.0059 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0075 - acc: 1.0000 - val_loss: 0.0052 - val_acc: 1.0000\n",
       "Epoch 17/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0063 - acc: 1.0000 - val_loss: 0.0057 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0064 - acc: 1.0000 - val_loss: 0.0058 - val_acc: 1.0000\n",
       "Epoch 18/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0062 - acc: 1.0000 - val_loss: 0.0067 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0060 - acc: 1.0000 - val_loss: 0.0056 - val_acc: 1.0000\n",
       "Epoch 19/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0055 - acc: 1.0000 - val_loss: 0.0048 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0053 - acc: 1.0000 - val_loss: 0.0048 - val_acc: 1.0000\n",
       "Epoch 20/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0049 - acc: 1.0000 - val_loss: 0.0063 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0051 - acc: 1.0000 - val_loss: 0.0041 - val_acc: 1.0000\n",
       "Epoch 21/100\n",
       "100/100 [==============================] - 0s - loss: 0.0052 - acc: 1.0000 - val_loss: 0.0040 - val_acc: 1.0000\n",
       "Epoch 22/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0044 - acc: 1.0000 - val_loss: 0.0033 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0044 - acc: 1.0000 - val_loss: 0.0041 - val_acc: 1.0000\n",
       "Epoch 23/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0040 - acc: 1.0000 - val_loss: 0.0032 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0041 - acc: 1.0000 - val_loss: 0.0035 - val_acc: 1.0000\n",
       "Epoch 24/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0040 - acc: 1.0000 - val_loss: 0.0043 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0042 - acc: 1.0000 - val_loss: 0.0036 - val_acc: 1.0000\n",
       "Epoch 25/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0038 - acc: 1.0000 - val_loss: 0.0038 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0037 - acc: 1.0000 - val_loss: 0.0030 - val_acc: 1.0000\n",
       "Epoch 26/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0034 - acc: 1.0000 - val_loss: 0.0025 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0036 - acc: 1.0000 - val_loss: 0.0039 - val_acc: 1.0000\n",
       "Epoch 27/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0036 - acc: 1.0000 - val_loss: 0.0031 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0034 - acc: 1.0000 - val_loss: 0.0035 - val_acc: 1.0000\n",
       "Epoch 28/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0032 - acc: 1.0000 - val_loss: 0.0027 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0030 - acc: 1.0000 - val_loss: 0.0026 - val_acc: 1.0000\n",
       "Epoch 29/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0030 - acc: 1.0000 - val_loss: 0.0024 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0032 - acc: 1.0000 - val_loss: 0.0045 - val_acc: 1.0000\n",
       "Epoch 30/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0031 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0031 - acc: 1.0000 - val_loss: 0.0024 - val_acc: 1.0000\n",
       "Epoch 31/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0030 - acc: 1.0000 - val_loss: 0.0043 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0028 - acc: 1.0000 - val_loss: 0.0028 - val_acc: 1.0000\n",
       "Epoch 32/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0026 - acc: 1.0000 - val_loss: 0.0032 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0026 - acc: 1.0000 - val_loss: 0.0023 - val_acc: 1.0000\n",
       "Epoch 33/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0024 - acc: 1.0000 - val_loss: 0.0037 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0027 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
       "Epoch 34/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0025 - acc: 1.0000 - val_loss: 0.0026 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0025 - acc: 1.0000 - val_loss: 0.0022 - val_acc: 1.0000\n",
       "Epoch 35/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0024 - acc: 1.0000 - val_loss: 0.0024 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0024 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
       "Epoch 36/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0021 - acc: 1.0000 - val_loss: 0.0018 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0024 - acc: 1.0000 - val_loss: 0.0019 - val_acc: 1.0000\n",
       "Epoch 37/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0021 - acc: 1.0000 - val_loss: 0.0019 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0022 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
       "Epoch 38/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0021 - acc: 1.0000 - val_loss: 0.0023 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0020 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 39/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0021 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
       "Epoch 40/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0020 - acc: 1.0000 - val_loss: 0.0018 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0022 - acc: 1.0000 - val_loss: 0.0014 - val_acc: 1.0000\n",
       "Epoch 41/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0022 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0020 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
       "Epoch 42/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0021 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0020 - acc: 1.0000 - val_loss: 0.0016 - val_acc: 1.0000\n",
       "Epoch 43/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0018 - acc: 1.0000 - val_loss: 0.0017 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0016 - val_acc: 1.0000\n",
       "Epoch 44/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0018 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0017 - val_acc: 1.0000\n",
       "Epoch 45/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0017 - val_acc: 1.0000\n",
       "Epoch 46/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0014 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0019 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
       "Epoch 47/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0018 - acc: 1.0000 - val_loss: 0.0014 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
       "Epoch 48/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0016 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
-      "Epoch 49/100\n",
       "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0021 - val_acc: 1.0000\n",
+      "Epoch 49/100\n",
+      "100/100 [==============================] - 0s - loss: 0.0016 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 50/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0016 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0010 - val_acc: 1.0000\n",
       "Epoch 51/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0025 - val_acc: 1.0000\n",
       "Epoch 52/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0017 - acc: 1.0000 - val_loss: 0.0019 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0016 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
       "Epoch 53/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0016 - acc: 1.0000 - val_loss: 0.0014 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0014 - acc: 1.0000 - val_loss: 0.0020 - val_acc: 1.0000\n",
       "Epoch 54/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0014 - acc: 1.0000 - val_loss: 0.0010 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0014 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 55/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0014 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0015 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
       "Epoch 56/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0015 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
       "Epoch 57/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0013 - acc: 1.0000 - val_loss: 0.0024 - val_acc: 1.0000\n",
       "Epoch 58/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0013 - acc: 1.0000 - val_loss: 8.4317e-04 - val_acc: 1.0000\n",
-      "Epoch 59/100\n",
       "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
+      "Epoch 59/100\n",
+      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0016 - val_acc: 1.0000\n",
       "Epoch 60/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0010 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 7.7586e-04 - val_acc: 1.0000\n",
       "Epoch 61/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 9.2067e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
       "Epoch 62/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0010 - acc: 1.0000 - val_loss: 8.1308e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0013 - acc: 1.0000 - val_loss: 0.0014 - val_acc: 1.0000\n",
       "Epoch 63/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 9.8218e-04 - val_acc: 1.0000\n",
       "Epoch 64/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0017 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.7383e-04 - acc: 1.0000 - val_loss: 9.6991e-04 - val_acc: 1.0000\n",
       "Epoch 65/100\n",
-      "100/100 [==============================] - 0s - loss: 9.9405e-04 - acc: 1.0000 - val_loss: 8.4340e-04 - val_acc: 1.0000\n",
-      "Epoch 66/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0010 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "Epoch 66/100\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 9.9847e-04 - val_acc: 1.0000\n",
       "Epoch 67/100\n",
-      "100/100 [==============================] - 0s - loss: 9.9135e-04 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.7296e-04 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 68/100\n",
-      "100/100 [==============================] - 0s - loss: 9.0876e-04 - acc: 1.0000 - val_loss: 7.5313e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.4549e-04 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 69/100\n",
-      "100/100 [==============================] - 0s - loss: 9.7704e-04 - acc: 1.0000 - val_loss: 8.0214e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.7626e-04 - acc: 1.0000 - val_loss: 9.9434e-04 - val_acc: 1.0000\n",
       "Epoch 70/100\n",
-      "100/100 [==============================] - 0s - loss: 9.3957e-04 - acc: 1.0000 - val_loss: 7.1984e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
       "Epoch 71/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0012 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.0783e-04 - acc: 1.0000 - val_loss: 9.0944e-04 - val_acc: 1.0000\n",
       "Epoch 72/100\n",
-      "100/100 [==============================] - 0s - loss: 9.8385e-04 - acc: 1.0000 - val_loss: 8.3657e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0010 - acc: 1.0000 - val_loss: 6.6179e-04 - val_acc: 1.0000\n",
       "Epoch 73/100\n",
-      "100/100 [==============================] - 0s - loss: 9.8486e-04 - acc: 1.0000 - val_loss: 5.3685e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.3634e-04 - acc: 1.0000 - val_loss: 8.0140e-04 - val_acc: 1.0000\n",
       "Epoch 74/100\n",
-      "100/100 [==============================] - 0s - loss: 9.1324e-04 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0010 - acc: 1.0000 - val_loss: 7.8819e-04 - val_acc: 1.0000\n",
       "Epoch 75/100\n",
-      "100/100 [==============================] - 0s - loss: 8.7023e-04 - acc: 1.0000 - val_loss: 9.1055e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.2504e-04 - acc: 1.0000 - val_loss: 8.1632e-04 - val_acc: 1.0000\n",
       "Epoch 76/100\n",
-      "100/100 [==============================] - 0s - loss: 8.8312e-04 - acc: 1.0000 - val_loss: 9.5169e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.4269e-04 - acc: 1.0000 - val_loss: 6.8168e-04 - val_acc: 1.0000\n",
       "Epoch 77/100\n",
-      "100/100 [==============================] - 0s - loss: 9.1021e-04 - acc: 1.0000 - val_loss: 6.5630e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 8.5385e-04 - acc: 1.0000 - val_loss: 7.7865e-04 - val_acc: 1.0000\n",
       "Epoch 78/100\n",
-      "100/100 [==============================] - 0s - loss: 9.1778e-04 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.1713e-04 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
       "Epoch 79/100\n",
-      "100/100 [==============================] - 0s - loss: 8.3930e-04 - acc: 1.0000 - val_loss: 5.9322e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.2030e-04 - acc: 1.0000 - val_loss: 8.8455e-04 - val_acc: 1.0000\n",
       "Epoch 80/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.9223e-04 - acc: 1.0000 - val_loss: 9.3847e-04 - val_acc: 1.0000\n",
       "Epoch 81/100\n",
-      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 8.3612e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 8.4465e-04 - acc: 1.0000 - val_loss: 5.5140e-04 - val_acc: 1.0000\n",
       "Epoch 82/100\n",
-      "100/100 [==============================] - 0s - loss: 8.5075e-04 - acc: 1.0000 - val_loss: 0.0013 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 8.0997e-04 - acc: 1.0000 - val_loss: 6.3979e-04 - val_acc: 1.0000\n",
       "Epoch 83/100\n",
-      "100/100 [==============================] - 0s - loss: 7.1350e-04 - acc: 1.0000 - val_loss: 0.0015 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 0.0011 - acc: 1.0000 - val_loss: 0.0012 - val_acc: 1.0000\n",
       "Epoch 84/100\n",
-      "100/100 [==============================] - 0s - loss: 9.1387e-04 - acc: 1.0000 - val_loss: 7.6505e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.1042e-04 - acc: 1.0000 - val_loss: 7.4020e-04 - val_acc: 1.0000\n",
       "Epoch 85/100\n",
-      "100/100 [==============================] - 0s - loss: 9.6451e-04 - acc: 1.0000 - val_loss: 0.0025 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.5559e-04 - acc: 1.0000 - val_loss: 5.8007e-04 - val_acc: 1.0000\n",
       "Epoch 86/100\n",
-      "100/100 [==============================] - 0s - loss: 8.1832e-04 - acc: 1.0000 - val_loss: 5.6651e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 8.4894e-04 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n",
       "Epoch 87/100\n",
-      "100/100 [==============================] - 0s - loss: 7.2781e-04 - acc: 1.0000 - val_loss: 0.0010 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.3317e-04 - acc: 1.0000 - val_loss: 0.0010 - val_acc: 1.0000\n",
       "Epoch 88/100\n",
-      "100/100 [==============================] - 0s - loss: 8.0625e-04 - acc: 1.0000 - val_loss: 6.3334e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.6423e-04 - acc: 1.0000 - val_loss: 7.8742e-04 - val_acc: 1.0000\n",
       "Epoch 89/100\n",
-      "100/100 [==============================] - 0s - loss: 7.6087e-04 - acc: 1.0000 - val_loss: 6.4656e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.5540e-04 - acc: 1.0000 - val_loss: 8.3755e-04 - val_acc: 1.0000\n",
       "Epoch 90/100\n",
-      "100/100 [==============================] - 0s - loss: 7.5786e-04 - acc: 1.0000 - val_loss: 6.8113e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.3232e-04 - acc: 1.0000 - val_loss: 5.9540e-04 - val_acc: 1.0000\n",
       "Epoch 91/100\n",
-      "100/100 [==============================] - 0s - loss: 8.1643e-04 - acc: 1.0000 - val_loss: 4.7783e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.7998e-04 - acc: 1.0000 - val_loss: 5.8214e-04 - val_acc: 1.0000\n",
       "Epoch 92/100\n",
-      "100/100 [==============================] - 0s - loss: 7.5797e-04 - acc: 1.0000 - val_loss: 6.2071e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 6.2086e-04 - acc: 1.0000 - val_loss: 4.7887e-04 - val_acc: 1.0000\n",
       "Epoch 93/100\n",
-      "100/100 [==============================] - 0s - loss: 7.6714e-04 - acc: 1.0000 - val_loss: 4.3379e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 9.1907e-04 - acc: 1.0000 - val_loss: 7.2314e-04 - val_acc: 1.0000\n",
       "Epoch 94/100\n",
-      "100/100 [==============================] - 0s - loss: 9.7814e-04 - acc: 1.0000 - val_loss: 0.0024 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.4112e-04 - acc: 1.0000 - val_loss: 7.4215e-04 - val_acc: 1.0000\n",
       "Epoch 95/100\n",
-      "100/100 [==============================] - 0s - loss: 8.0752e-04 - acc: 1.0000 - val_loss: 5.3576e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 7.1894e-04 - acc: 1.0000 - val_loss: 4.7123e-04 - val_acc: 1.0000\n",
       "Epoch 96/100\n",
-      "100/100 [==============================] - 0s - loss: 7.6372e-04 - acc: 1.0000 - val_loss: 6.8184e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 6.7798e-04 - acc: 1.0000 - val_loss: 5.7319e-04 - val_acc: 1.0000\n",
       "Epoch 97/100\n",
-      "100/100 [==============================] - 0s - loss: 6.7852e-04 - acc: 1.0000 - val_loss: 3.9996e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 6.6198e-04 - acc: 1.0000 - val_loss: 7.4432e-04 - val_acc: 1.0000\n",
       "Epoch 98/100\n",
-      "100/100 [==============================] - 0s - loss: 6.5877e-04 - acc: 1.0000 - val_loss: 6.1557e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 6.3420e-04 - acc: 1.0000 - val_loss: 4.6074e-04 - val_acc: 1.0000\n",
       "Epoch 99/100\n",
-      "100/100 [==============================] - 0s - loss: 6.7418e-04 - acc: 1.0000 - val_loss: 4.2036e-04 - val_acc: 1.0000\n",
+      "100/100 [==============================] - 0s - loss: 6.7688e-04 - acc: 1.0000 - val_loss: 4.2424e-04 - val_acc: 1.0000\n",
       "Epoch 100/100\n",
-      "100/100 [==============================] - 0s - loss: 6.0691e-04 - acc: 1.0000 - val_loss: 4.3382e-04 - val_acc: 1.0000\n"
+      "100/100 [==============================] - 0s - loss: 7.5914e-04 - acc: 1.0000 - val_loss: 0.0011 - val_acc: 1.0000\n"
      ]
     }
    ],
@@ -561,26 +553,24 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 13,
+   "metadata": {},
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "<matplotlib.legend.Legend at 0x119bcacd0>"
+       "<matplotlib.legend.Legend at 0x11c087110>"
       ]
      },
-     "execution_count": 12,
+     "execution_count": 13,
      "metadata": {},
      "output_type": "execute_result"
     },
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAEWCAYAAABynMHOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAswAAAEWCAYAAABynMHOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKeoDQR0AAAAASUVORK5CYII=\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x119ad9850>"
+       "<matplotlib.figure.Figure at 0x11c589210>"
       ]
      },
      "metadata": {},
@@ -605,16 +595,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 14,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "loss: 0.000604408510844, accuracy: 1.0\n"
+      "loss: 0.000917075452162, accuracy: 1.0\n"
      ]
     }
    ],
@@ -625,10 +613,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 15,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -666,16 +652,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 48,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 16,
+   "metadata": {},
    "outputs": [
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDUAAAEICAYAAABPrM81AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xu8HXV97//3e4cEkDsEuSQRUOMFbas2J9jDT+UonEeg\nSnoeRQ0eFTxY2j7kVFtsi+JBSttTPT1equIlVQp4ARH1GDUWFUHEAk2KiCYpklI1OwRCwi1cQ/Tz\n+2Mm6WK7916z9p4181l7vZ4+1sM1a2bPfNYm896zPus7M44IAQAAAAAADJqRtgsAAAAAAACYCpoa\nAAAAAABgINHUAAAAAAAAA4mmBgAAAAAAGEg0NQAAAAAAwECiqQEAAAAAAAZSuqaG7afZfsj2rJbr\nON/2Z/q4/tNtX9+v9U+wzTW2j5tk/jdsn9ZgSUDryJxcbL/E9m1t1wFg+mwfZ3u07ToATB/HS/3D\nZ7TpS9fUiIifR8TeEfGLtmuZaSLieRFxrTR+IETEiRFxSSvFAS0hc9plO2w/c+d0RHwvIp5dw3pP\nsf0D2w/avsX24umuE8Dkxu7PFZY/3/Ydth+wvdL2U8vXj7P9/fL1222f1L+qAVTB8VL/8Blt+tI1\nNfrF9m4tbrvVjiaA5pE5rTtM0u9JOkDS5yRd3Go1AMazm6TjJB0i6XFJ55evP03SuyQdVP7/ZW1m\nKoD+4XgJdWikqWH7p7b/1Patth+2/Snbh5RDabbZ/rbtA8pljyw7/buV09fa/suyY7/N9jdtz+1Y\n98nlkJ37y2WfO2a7f277VkkPj7fT2H6e7W/Zvtf23bbf2TF7ju1Ly+2usb2o4+eeW27v/nLeyR3z\nLrb9sfJbh4cl/RfbB9leUX5r+M+SnjHJ72vn7+BM23fa3mT77R3zd7f9wXLeneXz3ct5c21/razr\nXtvfsz3S8fs43vYSSe+U9NpyGNkPO37Xby7Xf7/t53ds82Dbj3Z8i/LK8tvP+23/k+1f7/bvAGgK\nmTPlzDnN9s9tb7F9bsf8Edvn2P4321ttX2H7wI75b7T9s3Le/9qZNeW8xbZvKOveZPsjtueU864r\nV/HDMote647h6uXv8soxtf6d7Q+Vz/cr/9tusr3R9l+5PECJiA9HxOryG6XvqfjQBGAKbL/J9lc7\npm+3/YWO6Q22Hywnd+3PHfPPtr253FfftPP1iHhX+e3vY5L+WeV+GhGXRsQ1EbFD0nWS9pH0lP6+\nS2D4cLzEZ7QZIyL6/pD0U0k3qvhjNU/SZkk3S3qhpD0kfUfSu8tlj5QUknYrp6+V9G+SniVpz3L6\nPeW8Z0l6WNIJkmZL+jNJ6yXN6djuLZIWSNpznLr2kbRJ0tllHftIOqacd76kxySdJGmWpL+RdGM5\nb3a5nXdKmiPp5ZK2SXp2Of9iSQ9IOlZF42gPSZdLukLSXpKeL2mjpOsn+H3t/B1cVi7/a5LukXR8\nOf+C8vf5VEkHS/onSX9ZzvsbSR8va5wt6SWS3PH7OL7j/X1mzHavlfTm8vlFkv66Y95bJP1j+fyF\n5X/DY8rfzWnlundv4t8TDx7dHmTOlDPn78v3/BsqvjV9bjn/reXvc76k3SV9QtJl5byjJT0k6f8r\na/u/kp7oyJrflPRiFd/IHilpnaS3dWw7JD2zY/o4SaPl8yMkPSJpn3J6Vvn7e3E5/eWylr1U5OE/\nS/r9Me9tL0k3SHpf2/8uefAY1Iekp0u6v8yXwyX9rGM/fbqk+8p54+3PO1Qct8wu8+0RSQeMWf+z\nVBznnDzm9Vnlfv7Ftn8HPHjMxIc4XuIz2gx5NHn6yYcj4u6I2KjiW7ObIuIHUXTnv1z+R5jIP0TE\nTyLiURX/6F5Qvv5aSV+PiG9FxBMqDqb3lPSfO372QxGxofzZsV4p6a6IeF9EPBYR2yLipo7510fE\nyii+6fu0igN9qThA31vFjrs9Ir4j6WuSTu342a9ExPcj4pcqDvB/V9J5EfFwRPxYUpXzov6iXP5H\nkv6hY/3/XdIFEbE5Iu6R9BeS3lDOe0LFsOsjIuKJKM5PjwrbGutzkpZ1TL+ufE2SzpT0iYi4KSJ+\nEcU5Xo+r+L0AWZA5U8ucRyPih5J+2LH9P5B0bkSMRsTOIeKnlN+snCLpqxFxfURsl3Seij/4kqSI\n+JeIuDEidkTET1U0IV5WoRZFxM9UHFz9t/Kll0t6JCJutH2IigOat5XvcbOkD+jJuSUV//3ulPTn\nVbYJ4FdFxB0qPhi8QNJLJV0l6U7bz1GxP3+vzJ7xPKHimOWJiFipogm667o5tg+S9G1J/zsiVoz5\n2Q+r+HDABfKA/uF4ic9oA6/JpsbdHc8fHWd670l+9q6O5490LLvz2wJJUvmPc4OKTuNOGyZZ7wIV\nHcaq292jPIg/XNKGMX/AfzbJdg9W8S3lhjHLdzN2+cPL509632Pm/a2KDuU3XVx865wK2xnPNZKe\nYvsY20eqCKkvl/OOkHR2Oazpftv3q/hdHj7umoB2kDm9Z85E7/sISV/u2N/XSfqFim92Du/cTkQ8\nImnrzmnbzyqHW95VDk//35J2DU+t4HP6j4OFzj/cR6j4pmNTR12fUPEBaNe2VXzgekMUw9gBTN13\nVYy8eGn5/FoV+9fLyumJbB2z/3Vmi1R8oFgfER/o/CHbe0r6fUmnRsRD0y0ewIQ4XuIz2sAb9AuF\n3qniP54kybZV/Ifb2LHMZB2wDSqGTU5luwt2ngdVetok271HxfDLBWOW72bs8nd2bP+I8eaVncyz\nI+Lpkk6W9Ce2XzHOuiftDJadzytUfJg4VdLXImJbOXuDimFP+3c8nhIRl1V4T8Agm+mZM5ENkk4c\ns8/vUX6rs0nFaSmSdn0QOajjZz8m6V8lLYyIfVUMCXUP2/6CpONsz1cxYmNnU2ODim8f5nbUtG9E\nPK/jZw+TdH/ZaAEwPTubGi8pn39X1Zoa3Rym/zi+6fRUFcep480DkNtMP17iM1oyg97UuELSb9t+\nhe3ZKs67elzF+UtVfE3SYbbfVl54ZR/bx1T4uZtUdAX/zPZsF/cVfpWKc7J+RfmP70uSzrf9FNtH\nq9pQyv9VLv88SW+S9Pny9cskvau8MMxcFcO9PyPtujjMM8vweEDFt6njDQm9W9KRY3b6sT6nYvjY\nf9d/fJCQivPu/6DsENr2XrZ/2/Y+Fd4TMMhmeuZM5OOS/tr2EdKui1ItLeddKelVtv+ziwuAnq8n\nNy32kfSgpIfKoep/OGbdd2uSA5dy+Oa1KoZ3/ntErCtf3yTpm5LeZ3tfFxczfYbtzlNbbtDkw2YB\nVPddSf9FxfnvoyqGqS9R0cT8QbnMpPvzBN6n4pzwsUYlHcYoK2AgzfTjJT6jJTPQTY2IuE3S61Wc\nc7lFxT/aV5XndVf5+W0qLmDzKhXDmG5X8Qe7289tL3/mxHK7H5X0xoj410l+7CwVQ7LuUnGRmn+o\nUOJ3VQxTulrS/42Ib5av/5Wk1ZJulfQjFeec/1U5b6GKc1MfUnFA/9GIuGacde+8avlW2zePt/Hy\n3LWHVQxZ+kbH66tV3CrxIyouDrZe0ukV3g8w0IYgcybyd5JWqBgyuU3FRbCOKWtbI+l/qjhg2KQi\nezarOHiRpLerOG1km4o/tp/Xk50v6ZJymORrJtj+5yQdryf/4ZakN6q4ENhaFVl0pYpvfXc6RtL1\nPbxPABOIiJ+o2L+/V04/KOkOSd8vPxhI1fbnsf5IRWNjrHmS1ptbLgIDZwiOl/iMlszOK64ikfL8\nqH+XNJtvKAAMEtt7q7hLwsKI+Pe26wEAAKgDn9HyGuiRGgCA9tl+VTkMcy8VVzj/kYpbiAEAAAB9\nRVMDADBdS1VcCOtOFcMrl03xNmUAAABATzj9BAAAAAAADCRGagAAAAAAgIG0Wz9W6t32DM8ZoDvH\nzNmj7Qp64O6LJBGPbFVsf6hywbP2PSJix6PV1v3oPVdFxJIpF4ehcMCBB8Xh84/ovmASs0YGZ//+\n5QCN8tu44ee6794tZBFas8/+B8bcw+e3XUZlP/v5PW2XUNk+B+7fdgmVPXrvJm1/6H6yCK3x7nuH\n9zyo7TKq217t3z96E9u3KXY82tNBZ9U8aiuL+tPUmLOPdn921Tt5JTDvuW1XUN2swbmz2ePf/Zue\nlo8dj1b+d/PYLRfOnUpNGC6Hzz9Cl339u22XUdn+T5nddgmVPbL9F90XSuKUJS/paXmyCHWbe/h8\nXXDp19suo7Iz3/aJtkuo7MWnvrLtEiq78b1v6ml5sgh1854Haffjzm27jOp+/qO2K6hugL7sefy2\nK3r+map51FYW9aWpAUyNJXNGFIC2kUUAMiCLAGSRO49oaiAPSxoZnJEoAGYosghABmQRgCyS5xFN\nDeTiwbmmAIAZjCwCkAFZBCCLxHmUdwwJhlA5rKnKAwD6hiwCkAFZBCCLinnUbS32AtvX2F5re43t\nt46zjG1/yPZ627faflG39ZKCyMWu9gCAfiKLAGRQYxbZXmL7tvKDwjnjzP+T8oPGrbavtj04tw8D\n0H/1ZNEOSWdHxNGSXizpLbaPHrPMiZIWlo8zJX2s20ppaiAPi28kALSPLAKQQY1ZZHuWpAtVfFg4\nWtKp43yQ+IGkRRHx65KulPR/6n1DAAZW1TzqIiI2RcTN5fNtktZJmjdmsaWSLo3CjZL2t33YZOvl\niAyJVPw2gm9HAfQVWQQgg1qzaLGk9RFxR0Rsl3S5ig8Ou0TENRHxSDl5o6T5tb4dAAOschbNtb26\n43HmhGu0j5T0Qkk3jZk1T9KGjulR/Wrj40m4UChySXxVXQBDhCwCkEH1LJpre3XH9PKIWN4xPd6H\nhGMmWd8Zkr5RdeMAhkC1PNoSEYu6LWR7b0lflPS2iHhwuqXR1EAiue9/DGBYkEUAMugpiyp9kKi0\nVfv1khZJelkd6wMwE9R3bGR7toqGxmcj4kvjLLJR0oKO6fnlaxPiqA15WAz5BtA+sghABvVmUaUP\nCbaPl3SupJMj4vE63gaAGaBqHnVbjW1Jn5K0LiLeP8FiKyS9sbwLyoslPRARmyZbLyM1kAvfjgLI\ngCwCkEF9WbRK0kLbR6loZiyT9Lonbcp+oaRPSFoSEZvr2jCAGaKePDpW0hsk/cj2LeVr75T0NEmK\niI9LWinpJEnrJT0i6U3dVkpTA4kw5BtABmQRgAzqy6KI2GH7LElXSZol6aKIWGP7AkmrI2KFpL+V\ntLekLxRfpurnEXFyLQUAGHD15FFEXF+sbNJlQtJbelkvTQ3kYUmzuDgfgJaRRQAyqDmLImKlim9A\nO187r+P58bVtDMDMkvzYiKYGcuEcdQAZkEUAMiCLAGSROI9oaiARhnwDyIAsApABWQQgi9x5RFMD\nuSTuAAIYImQRgAzIIgBZJM4jmhrIJXEHEMAQIYsAZEAWAcgicR5Vqsz2Etu32V5v+5x+F4UhVfVe\n7Im7hOgvsgiNIIvQBVmERpBF6IIsQmOSZ1HXkRq2Z0m6UNIJkkYlrbK9IiLW9rs4DKGRvFfVRbvI\nIjSKLMIEyCI0iizCBMgiNC5xHlUZqbFY0vqIuCMitku6XNLS/paF4VRegKbKA8OILEJD6ssi2wts\nX2N7re01tt86zjK2/aHym7Zbbb+oL28LdSGL0BCOizApsggNqphHLamy5XmSNnRMj5avPYntM22v\ntr06djxaV30YNsmHNqFVPWfRffduaaw4zDD1ZdEOSWdHxNGSXizpLbaPHrPMiZIWlo8zJX2szreC\n2vWcRdvuu7ex4jDDcFyEifX+GW37Q40VhxkocRbV1k6JiOURsSgiFnm3PetaLYaJxTcSmLbOLDrg\nwLltl4NBVGMWRcSmiLi5fL5N0jr96kHnUkmXRuFGSfvbPqzmd4WGdWbRPgcc2HY5GEQcF6EGT/qM\nNmfvtsvBoKqaRy2pcveTjZIWdEzPL18Dapb7/sdoHVmEhvQni2wfKemFkm4aM2uib9s21V4E6kAW\noSEcF2FSZBEalDuPqjQ1VklaaPsoFTvKMkmv62tVGF6JL0CD1pFFaE71LJpre3XH9PKIWD52Idt7\nS/qipLdFxIM1VIj2kEVoDsdFmBhZhGYlzqOuTY2I2GH7LElXSZol6aKIWNP3yjCcajwXy/YSSX+n\n4t/tJyPiPWPmP03SJZL2L5c5JyJW1lYAakUWoVHVs2hLRCyafFWeraKh8dmI+NI4i/Bt2wAhi9Ao\nrpeBCZBFaFziPKoyUkPlBz0+7KG/XN+wpoq3uXqXpCsi4mPlhftWSjqylgLQF2QRGlFvFlnSpySt\ni4j3T7DYCkln2b5c0jGSHogITj1JjCxCI2rMIsxMZBEakzyPKjU1gMbU1wHcdZurYrXeeZurzqZG\nSNq3fL6fpDvr2jiAAVdfFh0r6Q2SfmT7lvK1d0p6miRFxMdVHJCeJGm9pEckvamujQMYcIm/GQUw\nZBLnEU0NpOLqO0u389jHu/DeMWPWcb6kb9r+n5L2knR8b9UCmKl6yKJJRcT1Kq4ZPtkyIekttWwQ\nwIxSVxYBwHRlziOaGkjD6mln6XoeewWnSro4It5n+7ckfdr28yPil9NcL4AB1mMWAUBfkEUAssie\nRzQ1kIctj9S2s1S58N4ZkpZIUkTcYHsPSXMlba6rCAADqN4sAoCpIYsAZJE8j/Je7QNDyXalRwW7\nbnNle46K21ytGLPMzyW9otzucyXtIemeGt8OgAFVYxYBwJSRRQCyyJxFjNRAKjWexz7uba5sXyBp\ndUSskHS2pL+3/ccqLhp6enluO4Ahx4cEABmQRQCyyJxHNDWQSp07y3i3uYqI8zqer1VxZwIAeJLM\nf7gBDA+yCEAWmfOIpgbysLrcIwAAGkAWAciALAKQRfI8oqmBNCzOCwXQPrIIQAZkEYAssucRTQ2k\nMjLCtWsBtI8sApABWQQgi8x5RFMDqWTuAAIYHmQRgAzIIgBZZM4jmhrII/m5WgCGBFkEIAOyCEAW\nyfOIpgZSydwBBDA8yCIAGZBFALLInEc0NZBG9gvQABgOZBGADMgiAFlkzyOaGkjFI3l3FgDDgywC\nkAFZBCCLzHlEUwN5OPewJgBDgiwCkAFZBCCL5HnUl6bGUw48UL/xutf0Y9V98ZKjD2m7hMq2Pf6L\ntkuo7AtrPtLzz2TeWTB41v10s45500fbLqO6+ze1XUF1v9jRdgWVPX77aM8/QxahThs2P6Q/vvD7\nbZdR2fFvXNp2CZWd/1+f3XYJlb3hE3v0/DNkEeo0/9B9dfbZJ7RdRmXr7j627RIqe/yJwfmM9tVz\nr5nSz2XOI0ZqIJXMOwuA4UEWAciALAKQReY8oqmBNLJfgAbAcCCLAGRAFgHIInsejbRdAPAkrvgA\ngH4iiwBkQBYByKKmLLJ9ke3Ntn88wfzjbD9g+5bycV63dTJSA3lYGhmhzwagZWQRgAzIIgBZ1JtH\nF0v6iKRLJ1nmexHxyqorpKmBVDIPawIwPMgiABmQRQCyqCuPIuI620fWsrIS7V/kwjBLABmQRQAy\nqDGLbC+xfZvt9bbPGWf+S23fbHuH7VNqegcAZopmj4t+y/YPbX/D9vO6LcxIDaTCNxIAMiCLAGRQ\nVxbZniXpQkknSBqVtMr2iohY27HYzyWdLunttWwUwIxSMY/m2l7dMb08Ipb3uKmbJR0REQ/ZPknS\n/5O0cLIfoKmBNOzcV9UFMBzIIgAZ1JxFiyWtj4g7ynVfLmmppF1NjYj4aTnvl3VtFMDM0EMebYmI\nRdPZVkQ82PF8pe2P2p4bEVsm+hmaGkiFDxIAMiCLAGTQQxZ1+3Z0nqQNHdOjko6ZZnkAhkhTx0a2\nD5V0d0SE7cUqLpmxdbKfoamBVDzCBwkA7SOLAGTQQxZN+9tRAJhMXcdGti+TdJyKZuyopHdLmi1J\nEfFxSadI+kPbOyQ9KmlZRMRk66SpgVT4dhRABmQRgAxqzKKNkhZ0TM8vXwOASmq8+8mpXeZ/RMUt\nXyujqYE8zAcJAAmQRQAyqDeLVklaaPsoFc2MZZJeV9fKAcxwyY+NuKUr0rAku9oDAPqFLAKQQZ1Z\nFBE7JJ0l6SpJ6yRdERFrbF9g+2RJsv2fyqHgr5b0Cdtr+vbmAAyUqnnUFkZqIBHuOAAgA7IIQAb1\nZlFErJS0csxr53U8X6XitBQAGCP3sREjNZDKyIgrPaqwvcT2bbbX2z5ngmVeY3ut7TW2P1frmwEw\nsOrMIgCYKrIIQBaZs4iRGsijxmFLtmdJulDSCSpuW7bK9oqIWNuxzEJJ75B0bETcZ/up9WwdwEDj\n1BIAGZBFALJInkc0NZCGpTo7fIslrY+IOyTJ9uWSlkpa27HM70m6MCLuk6SI2FzXxgEMrpqzCACm\nhCwCkEX2POp6+onti2xvtv3jJgrCcOvhglhzba/ueJw5ZlXzJG3omB4tX+v0LEnPsv192zfaXtK3\nN4ZpI4vQJC4UismQR2gKWYTJkEVoUuYsqnJNjYsl8WEPjbBd6SFpS0Qs6ngsn8LmdpO0UNJxkk6V\n9Pe296/v3aBmF4ssQkN6yCIMp4tFHqEBZBG6uFhkERqSOYu6NjUi4jpJ9zZQC4ZdxW8jKu4vGyUt\n6JieX77WaVTSioh4IiL+XdJPVDQ5kBBZhMbUm0WYgcgjNIIsQhdkERqTPIu4+wnSsKyRkZFKjwpW\nSVpo+yjbcyQtk7RizDL/T8UoDdmeq+J0lDvqe0cABlGdWdRtaLDt42w/YPuW8nHeeMsBGD41HxcB\nwJRVzaO21Hah0PKaBmdK0pz9D6lrtRgydXX4ImKH7bMkXSVplqSLImKN7QskrY6IFeW8/2p7raRf\nSPrTiNhaTwVoS2cWaQ/OJsLU1Phtw8WSPiLp0kmW+V5EvLK2LSKFziwa2Wtuy9VgUDEKA9PVmUUH\nHHJ4y9VgkGXOo9qaGuU1DZZL0t7znxN1rRfDpc5zsSJipaSVY147r+N5SPqT8oEZojOLRvadTxZh\nSurKooi4zvaRtawMA6Uzi3ab+3SyCFPC9TIwXZ1Z9LTn/BpZhCnLnEfc0hV5cF4ogAyaz6Lfsv1D\nSXdKentErGl06wBy4rgIQBbJ86jKLV0vk3SDpGfbHrV9Rv/LwjCyuMo3JkYWoSk9ZlG320t3c7Ok\nIyLiNyR9WMW1fpAceYQmcFyEbsgiNKVqHrWl60iNiDi1iUIAKXcHEO0ii9CkHrJoS0Qsmup2IuLB\njucrbX/U9tyI2DLVdaL/yCM0heMiTIYsQpMy5xGnnyCVkZHEewuAodFUFtk+VNLdERG2F6sYQckF\niwFI4rgIQB6Z84imBvJw7gvQABgSNWZROTT4OBWnqYxKerek2ZIUER+XdIqkP7S9Q9KjkpaVFzEG\nMOw4LgKQRfI8oqmBNIpztdquAsCwqzOLug0NjoiPqLjlKwA8CcdFALLInkc0NZAIF7sCkAFZBCAD\nsghAFrnziKYGUkm8rwAYImQRgAzIIgBZZM4jmhrIw7kvQANgSJBFADIgiwBkkTyPaGogjZ33PwaA\nNpFFADIgiwBkkT2PaGoglcw7C4DhQRYByIAsApBF5jyiqYFUEu8rAIYIWQQgA7IIQBaZ84imBlLJ\n3AEEMDzIIgAZkEUAssicRzQ1kIdzdwABDAmyCEAGZBGALJLnEU0NpGE59VV1AQwHsghABmQRgCyy\n5xFNDaQykrkFCGBokEUAMiCLAGSROY9oaiCVxPsKgCFCFgHIgCwCkEXmPKKpgTTs3BegATAcyCIA\nGZBFALLInkc0NZBK4lO1AAwRsghABmQRgCwy51FfmhrPOHgvfeH3XtyPVffFbXdta7uEyr4/el/b\nJVQ2ewr/8jNfgAaDZ8999tJzXra47TIq27bt8bZLqGzffXdvu4TK1n3sWz3/DFmEOv3ykYf08C3X\nt11GZRd/+JS2S6jsrgcea7uEyqbyLSdZhDodsOccvfrX5rVdRmX3PX172yVUtuWhwan1n54ye0o/\nlzmPGKmBNKziyroA0CayCEAGZBGALLLnEU0NpJK4AQhgiJBFADIgiwBkkTmPRtouANjFlis+AKBv\nyCIAGZBFALKoMYtsX2R7s+0fTzDftj9ke73tW22/qNs6aWogleLKut0f1dblJbZvK3eIcyZZ7ndt\nh+1Fdb0PAIOtziwCgKkiiwBkUWMWXSxpySTzT5S0sHycKelj3VbI6SdIw5JGavrLbHuWpAslnSBp\nVNIq2ysiYu2Y5faR9FZJN9WyYQADr84sAoCpIosAZFFnHkXEdbaPnGSRpZIujYiQdKPt/W0fFhGb\nJvoBmhpIpcar6i6WtD4i7pAk25er2EHWjlnuLyW9V9Kf1rVhAIMv8xW+AQwPsghAFhXzaK7t1R3T\nyyNieY+bmidpQ8f0aPnahE0NTj9BGlWHWFZsEk60M3Rszy+StCAivl7bmwAw8GrOIgCYkrqzqNtp\nubZ3t/35cv5NXb5JBTBEesiiLRGxqOPRa0NjShipgVR6GNY0rS6g7RFJ75d0evXqAAwLhnwDyKDh\n03LPkHRfRDzT9jIVI1lfW0sBAAZeg8dGGyUt6JieX742IZoaSKWHXWVLREx2Yc9uO8M+kp4v6dry\nSr2HSlph++SI6GyWABhCtDQAZFBjFlU5LXeppPPL51dK+ohtl+e1AxhyDR4brZB0VplTx0h6YLLr\naUg0NZBMjbclWyVpoe2jVDQzlkl63c6ZEfGApLkd271W0ttpaACQas0iAJiyHrKo2wjW8U7LPWbM\nOnYtExE7bD8g6SBJW3oqGsCMVNexke3LJB2nIrdGJb1b0mxJioiPS1op6SRJ6yU9IulN3dZJUwNp\nFFfVrWdd5R/jsyRdJWmWpIsiYo3tCyStjogV9WwJwExTZxYBwFT1mEXdRrACwJTV/Dnt1C7zQ9Jb\nelknTQ3kYdd6le+IWKmi09f52nkTLHtcbRsGMNhqziIAmJJ6s6jKOeo7lxm1vZuk/SRtrasAAAMs\n+bERTQ2kwpBvABmQRQAyaOq03NIKSadJukHSKZK+w/U0AOyU+diIpgbSYMg3gAzIIgAZtHBa7qck\nfdr2ekn3qmh8AED6YyOaGkglcwcQwPAgiwBkUGcWdTstNyIek/Tq2jYIYEbJfGxEUwOp5N1VAAwT\nsghABmRYkG/sAAAZVElEQVQRgCwy5xFNDaRhS7Myj2sCMBTIIgAZkEUAssieRzQ1kErmYU0AhgdZ\nBCADsghAFpnzaKTbArYX2L7G9lrba2y/tYnCMJzsag8MH7IITSKLMBGyCE0iizARsghNy5xFVUZq\n7JB0dkTcbHsfSf9i+1sRsbbPtWHIWNYIf5kxMbIIjSCL0AVZhEaQReiCLEJjsudR16ZGRGyStKl8\nvs32OknzJLHDoF5824BJkEVoDFmESZBFaAxZhEmQRWhU8jzqevpJJ9tHSnqhpJvGmXem7dW2V2/d\nsqWe6jB0bFd6YLhVzaIdDz/QdGmYIerKItsX2d5s+8cTzLftD9leb/tW2y+q/c2gb6pmUex4tOnS\nMENwXIQq+IyGJmTOospNDdt7S/qipLdFxINj50fE8ohYFBGLDpo7t84aMSQsaZZd6YHh1UsW7bbX\nfs0XiIFXcxZdLGnJJPNPlLSwfJwp6WPTrR/N6CWLvNuezReIgcdxEargMxqaUDWP2lLp7ie2Z6vY\nWT4bEV/qb0kYZonvFIQEyCI0pa4siojrym/QJrJU0qUREZJutL2/7cPKYcVIiixCUzguwmTIIjQp\ncx51bWq4GEfyKUnrIuL9/S8JwyzzzoJ2kUVoUg9ZNNf26o7p5RGxvIdNzZO0oWN6tHyNpkZSZBGa\nxHERJkIWoWmZ86jKSI1jJb1B0o9s31K+9s6IWNm/sjCMilsBJd5b0DayCI3oMYu2RMSiftaDdMgi\nNILjInRBFqEx2fOoyt1PrldxGg3Qd5k7gGgXWYQmNZhFGyUt6JieX76GpMgiNInjIkyELELTMudR\nT3c/AfrNrvYAgH5qMItWSHpjeReUF0t6gOtpANiJ4yIAWWTOokoXCgWaYEm78ZcZQMvqzCLbl0k6\nTsW1N0YlvVvSbEmKiI9LWinpJEnrJT0i6U21bBjAwOO4CEAW2fOIpgZSSbyvABgidWVRRJzaZX5I\neks9WwMw03BcBCCLzHlEUwNp2NZI5r0FwFAgiwBkQBYByCJ7HtHUQCqJ9xUAQ4QsApABWQQgi8x5\nRFMDqWS+qi6A4UEWAciALAKQReY8oqmBNCxpVua9BcBQIIsAZEAWAcgiex7R1EAezt0BBDAkyCIA\nGZBFALJInkc0NZCKlXhvATA0yCIAGZBFALLInEcjbRcA7GQVHcAqj0rrs5fYvs32etvnjDP/T2yv\ntX2r7attH1HzWwIwgOrOIgCYCrIIQBZV86gtjNRAKnXtDLZnSbpQ0gmSRiWtsr0iItZ2LPYDSYsi\n4hHbfyjp/0h6bT0VABhkfEgAkAFZBCCLzHlEUwOpuL57BS2WtD4i7ijXe7mkpZJ2NTUi4pqO5W+U\n9Pq6Ng5gsNWYRQAwZWQRgCwy5xFNDaRhS7OqnxA11/bqjunlEbG8Y3qepA0d06OSjplkfWdI+kbl\nrQOYsXrMIgDoC7IIQBbZ84imBlIZqd4B3BIRi+rYpu3XS1ok6WV1rA/A4OshiwCgb8giAFlkziOa\nGkhj5wVoarJR0oKO6fnla0/epn28pHMlvSwiHq9t6wAGVs1ZBABTQhYByCJ7HtHUQCo1NgBXSVpo\n+ygVzYxlkl735G35hZI+IWlJRGyubcsABl7iLyMADBGyCEAWmfOoL02NH/10q57xPz7dj1X3xfFL\nfr3tEiq7+ZbRtkuo7J77H+3xJ6yRmu5/HBE7bJ8l6SpJsyRdFBFrbF8gaXVErJD0t5L2lvSF8sI3\nP4+Ik2spACkcut8eesernt12GZX9IqLtEiq777En2i6hsr/57O49/kR9WQRIkvfcS3OeO9llnXJ5\n97dub7uEyp7z1D3aLqGybY/3mptkEeq1buMDOuZdg3MJuacesm/bJVS25Z6H2i6hsrvueXgKP5U7\njxipgTSsejuAEbFS0soxr53X8fz4+rYGYKaoO4sAYCrIIgBZZM+jxNcwxdCxtNuIKz0AoG/IIgAZ\nkEUAsqiYR5VWZS+xfZvt9bbPGWf+6bbvsX1L+Xhzt3UyUgNpZO8AAhgOZBGADMgiAFnUlUe2Z0m6\nUNIJkkYlrbK9IiLWjln08xFxVtX10tRAKplvFQRgeJBFADIgiwBkUVMeLZa0PiLukCTbl0taKmls\nU6MnnH6CVOxqDwDoJ7IIQAZkEYAsKmbRXNurOx5njlnNPEkbOqZHy9fG+l3bt9q+0vaCbrUxUgNp\nWHTZALSPLAKQQVNZZPtASZ+XdKSkn0p6TUTcN85y/yjpxZKuj4hXNlAagCR6yKMtEbFompv7qqTL\nIuJx278v6RJJL5/sBzhuQx4uhjVVeQBA35BFADJoLovOkXR1RCyUdHU5PZ6/lfSG6W4MwACqmEcV\nbJTUOfJifvnaLhGxNSIeLyc/Kek3u62UpgbSsPggAaB9ZBGADBrMoqUqvglV+f+/M95CEXG1pG3T\n3RiAwVM1jypYJWmh7aNsz5G0TNKKJ23LPqxj8mRJ67qtlNNPkAofEQBkQBYByKCHLJpre3XH9PKI\nWF7xZw+JiE3l87skHVJ9swCGRR3HRhGxw/ZZkq6SNEvSRRGxxvYFklZHxApJf2T7ZEk7JN0r6fRu\n66WpgVT44hNABmQRgAx6yKJJz2O3/W1Jh44z69zOiYgI21F5qwCGRl3HRhGxUtLKMa+d1/H8HZLe\n0cs6aWogEct8kgDQOrIIQAb1ZVFEHD/hVuy7bR8WEZvKYd+ba9kogBkk97ER19RAGjuvqlvlAQD9\nQhYByKDBLFoh6bTy+WmSvjL9VQKYSarmUVsYqYFUuPAegAzIIgAZNJRF75F0he0zJP1M0mskyfYi\nSX8QEW8up78n6TmS9rY9KumMiLiqiQIBtC/zsRFNDeRhpR7WBGBIkEUAMmgoiyJiq6RXjPP6aklv\n7ph+Sd+LAZBT8mMjmhpIY+ewJgBoE1kEIAOyCEAW2fOIpgZSydwBBDA8yCIAGZBFALLInEc0NZBK\n3l0FwDAhiwBkQBYByCJzHnVtatjeQ9J1knYvl78yIt7d78IwfCxpVuIOINpFFqEpZBEmQxahKWQR\nJkMWoUnZ86jKSI3HJb08Ih6yPVvS9ba/ERE39rk2DKHE+wraRxahMWQRJkEWoTFkESZBFqFRmfOo\na1MjIkLSQ+Xk7PIR/SwKw8py6oFNaBNZhOaQRZgYWYTmkEWYGFmEZuXOo0oXMbU9y/YtkjZL+lZE\n3NTfsjCs7GoPDCeyCE0hizAZsghNIYswGbIITcqcRZWaGhHxi4h4gaT5khbbfv7YZWyfaXu17dXx\n+La668QQKG4V5EoPDKdes+jB+7c2XyQGXt1ZZHuJ7dtsr7d9zjjzT7d9j+1byseb635PqFfPx0WP\ncVyE3nFchG56zaJfPvpg80ViRqiaR23p6XazEXG/pGskLRln3vKIWBQRi7z7PnXVh2FS8dsIvpFA\n1Szad/+Dmi8Og6/GLLI9S9KFkk6UdLSkU20fPc6in4+IF5SPT9b6ftA3lY+L9uC4CFPAcREqqppF\nI3vu23xxmBmSZ1HXpobtg23vXz7fU9IJkv6134VhOI3YlR4YPmQRmlRjFi2WtD4i7oiI7ZIul7S0\nr8Wjr8giNInjIkyELELTMmdRlbufHCbpkvLbphFJV0TE1/pbFoaRJY3wdxkTI4vQiB6zaK7t1R3T\nyyNiecf0PEkbOqZHJR0zznp+1/ZLJf1E0h9HxIZxlkEOZBEawXERuiCL0JjseVTl7ie3SnphA7UA\ntV5V1/YSSX8naZakT0bEe8bM313SpZJ+U9JWSa+NiJ/WVgBqRRahST1k0ZaIWDTNzX1V0mUR8bjt\n35d0iaSXT3Od6BOyCE3KfLcBtIssQtMy51FP19QA+q3h89jPkHRfRDxT0gckvbfedwNgUNV4HvtG\nSQs6pueXr+0SEVsj4vFy8pMqGq0AwDU1AKSROYtoaiAVV/xfBVXOY1+q4htRSbpS0itsDg0A1JpF\nqyQttH2U7TmSlkla8aRt2Yd1TJ4saV1tbwTAQKsxiwBgWjJnUZVragCNaOE89l3LRMQO2w9IOkjS\nlh7KBjDD1HneaJktZ0m6SsWpcBdFxBrbF0haHRErJP2R7ZMl7ZB0r6TT69k6gEGW/Rx2AMMjex7R\n1EAevV01t47z2AHgV9V8Be+IWClp5ZjXzut4/g5J76htgwBmBu5sAiCL5HlEUwOp1LirdD2PvWOZ\nUdu7SdpPxQVDAQy5vH+2AQwTsghAFpnziKYG0iiGNdW2u+w6j11F82KZpNeNWWaFpNMk3SDpFEnf\niYioqwAAg6nmLAKAKSGLAGSRPY9oaiCVunaViuexf0rSp22vV3Ee+7KaNg9gwOX9sw1gmJBFALLI\nnEc0NZBLjXtLhfPYH5P06vq2CGDGyPyXG8DwIIsAZJE4j2hqIJXMw5oADA+yCEAGZBGALDLnEU0N\npJJ3VwEwTMgiABmQRQCyyJxHNDWQS+a9BcDwIIsAZEAWAcgicR7R1EAaluTMewuAoUAWAciALAKQ\nRfY8oqmBPCwlPlULwLAgiwBkQBYByCJ5HtHUQCqJ9xUAQ4QsApABWQQgi8x5RFMDiVjO3AIEMCTI\nIgAZkEUAssidRzQ1kErifQXAECGLAGRAFgHIInMe9aWp8YzD99MHL3hlP1bdF887dL+2S6js0P/x\nn9ouobJjb/rrnpa3cg9rwuDZf8/ZeuXzD2+7jMo23vto2yVUdsh+u7ddQmXLnzKnp+XJItQttm/X\n9jt/2nYZlf3Vkte1XUJle8ye1XYJlf3DXr3lJlmEuj1/wf76/gdObruMyu59aHvbJVR20D6Dc1x0\n7Pf/ouefyZ5HjNRALpn3FgDDgywCkAFZBCCLxHk00nYBQCdX/B8A9BNZBCADsghAFnVlke0ltm+z\nvd72OePM393258v5N9k+sts6aWogFbvaAwD6iSwCkAFZBCCLOrLI9ixJF0o6UdLRkk61ffSYxc6Q\ndF9EPFPSByS9t9t6aWogj4p/uPnjDaCvyCIAGZBFALKoL4sWS1ofEXdExHZJl0taOmaZpZIuKZ9f\nKekV7nLrFa6pgVQYQgkgA7IIQAZkEYAsKubRXNurO6aXR8Tyjul5kjZ0TI9KOmbMOnYtExE7bD8g\n6SBJWybaKE0NpGHxbQOA9pFFADIgiwBk0UMebYmIRf2t5ldx+glSccUHAPQTWQQggyayyPaBtr9l\n+/by/w8YZ5kX2L7B9hrbt9p+7TQ3C2DA1JRFGyUt6JieX7427jK2d5O0n6Stk62UpgZy4ZMEgAzI\nIgAZNJNF50i6OiIWSrq6nB7rEUlvjIjnSVoi6YO295/2lgEMjnqyaJWkhbaPsj1H0jJJK8Yss0LS\naeXzUyR9JyJispVy+glSGWGcJYAEyCIAGTSURUslHVc+v0TStZL+vHOBiPhJx/M7bW+WdLCk+5so\nEED76sij8hoZZ0m6StIsSRdFxBrbF0haHRErJH1K0qdtr5d0r4rGx6RoaiAVPkYAyIAsApBBD1nU\n7eJ8kzkkIjaVz++SdMikNdmLJc2R9G/VywMw6Oo6NoqIlZJWjnntvI7nj0l6dS/rpKmBXPgkASAD\nsghABtWzaNKL89n+tqRDx5l1budERITtCYd52z5M0qclnRYRv6xcHYDBl/jYiKYG0ihOxUq8twAY\nCmQRgAzqzKKIOH7C7dh32z4sIjaVTYvNEyy3r6SvSzo3Im6spTAAAyH7sREXCkUeLm4VVOUBAH1D\nFgHIoLks6rwo32mSvvIrpRQX9PuypEsj4sppbxHAYEl+XERTA6lwwwEAGZBFADJoKIveI+kE27dL\nOr6clu1Ftj9ZLvMaSS+VdLrtW8rHC6a/aQCDIvNxEaefIBHLfPUJoHVkEYAMmsmiiNgq6RXjvL5a\n0pvL55+R9Jm+FwMgqdzHRozUQCpNDLO0faDtb9m+vfz/A8ZZ5gW2b7C9xvattl87va0CGCScfgIg\nA7IIQBaZs4imBtKoOsSyhv3lHElXR8RCSVeX02M9IumNEfE8SUskfdD2/tPfNIDsGswiAJgQWQQg\ni+xZRFMDuTSzxyyVdEn5/BJJvzN2gYj4SUTcXj6/U8WVwA+e9pYBDIbsf70BDAeyCEAWibOoclPD\n9izbP7D9tX4WhOHmiv+TNNf26o7HmT1s5pCI2FQ+v0vSIZPWZC+WNEfSv03pTaFWZBGa0EMWYUiR\nRWgCWYRuyCI0JXMW9XKh0LdKWidp3z7VAvRyLtaWiFg08Xr8bUmHjjPr3M6JiAjbMcl6DpP0aUmn\nRcQvK1eHfiKL0Heco44KyCL0HVmECsgiNCJzHlUaqWF7vqTflvTJbssCU2ZppOKjm4g4PiKeP87j\nK5LuLpsVO5sWm8ctx95X0tclnRsRN9b3RjFVZBEaUWMWYWYii9AIsghdkEVoTPIsqnr6yQcl/Zmk\nCb+ptn3mzlMBHrjv3lqKwzBq5OTRFZJOK5+fJukrv1KFPUfSlyVdGhFXTneDqE1PWXTPlnuaqwwz\nDCeyY1I9ZVE88XBzlWGGIYswqZ6yaAvHRZiWvFnUtalh+5WSNkfEv0y2XEQsj4hFEbFovwMOrK1A\nDA9LTd267D2STrB9u6Tjy2nZXmR7Z6f7NZJeKul027eUjxdMe8uYsqlk0cFzubYreld3FtleYvs2\n2+tt/8rdlmzvbvvz5fybbB9Z6xtCraaSRZ69V0PVYSZp8LgIA2gqWTSX4yJMUdU8akuVa2ocK+lk\n2ydJ2kPSvrY/ExGv729pGEZN7AsRsVXSK8Z5fbWkN5fPPyPpMw2Ug+rIIjSmriyyPUvShZJOkDQq\naZXtFRGxtmOxMyTdFxHPtL1M0nslvbamElA/sgiNoV+BSZBFaFTmPOo6UiMi3hER8yPiSEnLJH2H\nnQX9wjcSmAhZhCbVmEWLJa2PiDsiYruky1XcVrpT522mr5T0Cpuky4osQpM4LsJEyCI0LXMW9XL3\nE6DvOI4HkEEPWTTX9uqO6eURsbxjep6kDR3To5KOGbOOXctExA7bD0g6SNKWnooGMONwXAQgi8x5\n1FNTIyKulXRtXyoBlHtYE/Igi9BvPWTRpLeXxsxGFqHfOC5CFWQRmpA5jxipgTTaHrYEAFLtWbRR\n0oKO6fnla+MtM2p7N0n7SdpaWwUABhLHRQCyyJ5HVW/pCjTCFf8HAP1UYxatkrTQ9lHlraKXqbit\ndKfO20yfouK86KjtzQAYWBwXAcgicxYxUgO58HcZQAY1ZVF5jYyzJF0laZakiyJije0LJK2OiBWS\nPiXp07bXS7pXReMDADguApBH4jyiqYFUEu8rAIZInVkUESslrRzz2nkdzx+T9OoaNwlghuC4CEAW\nmfOIpgYSsUYyn6wFYEiQRQAyIIsAZJE7j2hqIA0r9wVoAAwHsghABmQRgCyy5xEXCgUAAAAAAAOJ\nkRpIJXMHEMDwIIsAZEAWAcgicx7R1EAq3JYMQAZkEYAMyCIAWWTOI5oayMO5O4AAhgRZBCADsghA\nFsnziKYG0sh+ARoAw4EsApABWQQgi+x5RFMDqWQe1gRgeJBFADIgiwBkkTmPaGoglcwdQADDgywC\nkAFZBCCLzHlEUwOpJN5XAAwRsghABmQRgCwy5xFNDeSSeW8BMDzIIgAZkEUAskicRzQ1kIYljWQe\n1wRgKJBFADIgiwBkkT2PHBH1r9S+R9LPal7tXElbal5nPw1Svf2q9YiIOLjqwrb/saylii0RsWRq\nZWFYkEWSBqtesggzUp+ySGL/7qd+1EsWoVVkkaTBqlVKkEVST3nUShb1panRD7ZXR8SituuoapDq\nHaRagbYN2v4ySPUOUq1ABoO0zwxSrdLg1Qu0aZD2l0GqVRq8etsy0nYBAAAAAAAAU0FTAwAAAAAA\nDKRBamosb7uAHg1SvYNUK9C2QdtfBqneQaoVyGCQ9plBqlUavHqBNg3S/jJItUqDV28rBuaaGgAA\nAAAAAJ0GaaQGAAAAAADALjQ1AAAAAADAQBqIpobtJbZvs73e9jlt1zMZ2xfZ3mz7x23X0o3tBbav\nsb3W9hrbb227JiAzsqg/yCKgN2RRf5BFQG/Iov4gi3qX/poatmdJ+omkEySNSlol6dSIWNtqYROw\n/VJJD0m6NCKe33Y9k7F9mKTDIuJm2/tI+hdJv5P1dwu0iSzqH7IIqI4s6h+yCKiOLOofsqh3gzBS\nY7Gk9RFxR0Rsl3S5pKUt1zShiLhO0r1t11FFRGyKiJvL59skrZM0r92qgLTIoj4hi4CekEV9QhYB\nPSGL+oQs6t0gNDXmSdrQMT0q/qPWzvaRkl4o6aZ2KwHSIosaQBYBXZFFDSCLgK7IogaQRdUMQlMD\nfWZ7b0lflPS2iHiw7XoADCeyCEAGZBGADMii6gahqbFR0oKO6fnla6iB7dkqdpbPRsSX2q4HSIws\n6iOyCKiMLOojsgiojCzqI7KoN4PQ1FglaaHto2zPkbRM0oqWa5oRbFvSpySti4j3t10PkBxZ1Cdk\nEdATsqhPyCKgJ2RRn5BFvUvf1IiIHZLOknSVioukXBERa9qtamK2L5N0g6Rn2x61fUbbNU3iWElv\nkPRy27eUj5PaLgrIiCzqK7IIqIgs6iuyCKiILOorsqhH6W/pCgAAAAAAMJ70IzUAAAAAAADGQ1MD\nAAAAAAAMJJoaAAAAAABgINHUAAAAAAAAA4mmBgAAAAAAGEg0NQAAAAAAwECiqQEAAAAAAAbS/w/o\noyK1xBtpUQAAAABJRU5ErkJggg==\n",
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAABDUAAAEICAYAAABPrM81AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nABhIFDUAAAAAAMBA+v+oqA0+tSPt8wAAAABJRU5ErkJggg==\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x11bc18890>"
+       "<matplotlib.figure.Figure at 0x11cad9990>"
       ]
      },
      "metadata": {},
@@ -705,7 +689,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 58,
+   "execution_count": 18,
    "metadata": {
     "collapsed": true
    },
@@ -717,7 +701,7 @@
     "def plot_convolution(x, kernels):\n",
     "    \"\"\"kernels: 3d, (height, width, channel) and four channels.\"\"\"\n",
     "    n_kernels = kernels.shape[2]\n",
-    "    plt.figure(figsize=(6, 3))\n",
+    "    plt.figure(figsize=(9, 5))\n",
     "    plt.subplot(1, n_kernels+1, 1)\n",
     "    plt.imshow(x)\n",
     "    plt.title('Input x')\n",
@@ -730,10 +714,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 59,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 19,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -744,9 +726,9 @@
     },
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAVwAAADSCAYAAAAG23gwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n1qcxhlf5AAAAAElFTkSuQmCC\n",
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAesAAAE/CAYAAABmXOuYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABE7v8DRWDSuQ12dQUAAAAASUVORK5CYII=\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x11d2da1d0>"
+       "<matplotlib.figure.Figure at 0x11c5a9490>"
       ]
      },
      "metadata": {},
@@ -760,10 +742,8 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 60,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 20,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
@@ -774,9 +754,9 @@
     },
     {
      "data": {
-      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAADSCAYAAAB0FBqGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
+      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe8AAAE/CAYAAABvt0viAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRU5ErkJggg==\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x11d6cae50>"
+       "<matplotlib.figure.Figure at 0x11e81a290>"
       ]
      },
      "metadata": {},
@@ -799,19 +779,17 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 61,
-   "metadata": {
-    "collapsed": false
-   },
+   "execution_count": 21,
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[[-0.49479741  2.32843733]\n",
-      " [ 2.9154563  -3.87876177]\n",
-      " [-0.61376089 -0.32271883]\n",
-      " [-3.05479431  3.24747348]]\n"
+      "[[-0.50099331  2.33464241]\n",
+      " [ 2.9126389  -3.8759408 ]\n",
+      " [-0.6134612  -0.3230224 ]\n",
+      " [-3.05488276  3.24755979]]\n"
      ]
     }
    ],
@@ -864,9 +842,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.11"
+   "version": "2.7.13"
   }
  },
  "nbformat": 4,
- "nbformat_minor": 0
+ "nbformat_minor": 1
 }
diff --git a/Example_5-1.py b/Example_5-1.py
index 657aeeb..c958fd8 100644
--- a/Example_5-1.py
+++ b/Example_5-1.py
@@ -16,7 +16,6 @@
 import pandas as pd
 import numpy as np
 import models_time_varying
-import models_MLP
 
 from sklearn.preprocessing import LabelEncoder