Fix a couple of bugs: replaced the fizzbuzz call whose signature no longer...

Fix a couple of bugs: replaced the fizzbuzz call whose signature no longer matched, and remove the Flatter layer which seems to be buggy. Also include minor edits to match the style guide.

samples/core/guide/autograph.ipynb

@@ -593,19 +593,19 @@
       "cell_type": "code",
       "source": [
         "@autograph.convert()\n",
-        "def fizzbuzz_each(nums):\n",
+        "def squares(nums):\n",
         "\n",
         "  result = []\n",
-        "  autograph.set_element_type(result, tf.string)\n",
+        "  autograph.set_element_type(result, tf.int64)\n",
         "\n",
         "  for num in nums: \n",
-        "    result.append(fizzbuzz(num))\n",
+        "    result.append(num * num)\n",
         "    \n",
         "  return autograph.stack(result)\n",
         "    \n",
         "with tf.Graph().as_default():  \n",
         "  with tf.Session() as sess:\n",
-        "    print(sess.run(fizzbuzz_each(tf.constant(np.arange(10)))))"
+        "    print(sess.run(squares(tf.constant(np.arange(10)))))"
       ],
       "execution_count": 0,
       "outputs": []
@@ -679,24 +679,24 @@
         "\n",
         "@autograph.convert()\n",
         "def collatz(x):\n",
-        "  x=tf.reshape(x,())\n",
-        "  assert x>0\n",
+        "  x = tf.reshape(x,())\n",
+        "  assert x > 0\n",
         "  n = tf.convert_to_tensor((0,)) \n",
-        "  while not tf.equal(x,1):\n",
-        "    n+=1\n",
+        "  while not tf.equal(x, 1):\n",
+        "    n += 1\n",
         "    if tf.equal(x%2, 0):\n",
-        "      x = x//2\n",
+        "      x = x // 2\n",
         "    else:\n",
-        "      x = 3*x+1\n",
+        "      x = 3 * x + 1\n",
         "      \n",
         "  return n\n",
         "\n",
         "with tf.Graph().as_default():\n",
         "  model = tf.keras.Sequential([\n",
-        "    tf.keras.layers.Lambda(collatz, input_shape=(1,), output_shape=(), )\n",
+        "    tf.keras.layers.Lambda(collatz, input_shape=(1,), output_shape=())\n",
         "  ])\n",
         "  \n",
-        "result = model.predict(np.array([6171])) #261\n",
+        "result = model.predict(np.array([6171]))\n",
         "result"
       ],
       "execution_count": 0,
@@ -738,7 +738,7 @@
         "  def build(self,input_shape):\n",
         "    super().build(input_shape.as_list())\n",
         "    self.depth = len(self.layers)\n",
-        "    self.plims = np.linspace(self.pfirst, self.plast, self.depth+1)[:-1]\n",
+        "    self.plims = np.linspace(self.pfirst, self.plast, self.depth + 1)[:-1]\n",
         "    \n",
         "  @autograph.convert()\n",
         "  def call(self, inputs):\n",
@@ -749,7 +749,7 @@
         "    \n",
         "    p = tf.random_uniform((self.depth,))\n",
         "    \n",
-        "    keeps = p<=self.plims\n",
+        "    keeps = (p <= self.plims)\n",
         "    x = inputs\n",
         "    \n",
         "    count = tf.reduce_sum(tf.cast(keeps, tf.int32))\n",
@@ -781,7 +781,7 @@
       },
       "cell_type": "code",
       "source": [
-        "train_batch = np.random.randn(64, 28,28,1).astype(np.float32)"
+        "train_batch = np.random.randn(64, 28, 28, 1).astype(np.float32)"
       ],
       "execution_count": 0,
       "outputs": []
@@ -811,9 +811,9 @@
         "  for n in range(20):\n",
         "    model.add(\n",
         "          layers.Conv2D(filters=16, activation=tf.nn.relu,\n",
-        "                        kernel_size=(3,3), padding='same'))\n",
+        "                        kernel_size=(3, 3), padding='same'))\n",
         "\n",
-        "  model.build(tf.TensorShape((None, None, None,1)))\n",
+        "  model.build(tf.TensorShape((None, None, None, 1)))\n",
         "  \n",
         "  init = tf.global_variables_initializer()"
       ],
@@ -918,7 +918,6 @@
       "source": [
         "def mlp_model(input_shape):\n",
         "  model = tf.keras.Sequential((\n",
-        "      tf.keras.layers.Flatten(),\n",
         "      tf.keras.layers.Dense(100, activation='relu', input_shape=input_shape),\n",
         "      tf.keras.layers.Dense(100, activation='relu'),\n",
         "      tf.keras.layers.Dense(10, activation='softmax')))\n",
@@ -927,7 +926,7 @@
         "\n",
         "\n",
         "def predict(m, x, y):\n",
-        "  y_p = m(x)\n",
+        "  y_p = m(tf.reshape(x, (-1, 28 * 28)))\n",
         "  losses = tf.keras.losses.categorical_crossentropy(y, y_p)\n",
         "  l = tf.reduce_mean(losses)\n",
         "  accuracies = tf.keras.metrics.categorical_accuracy(y, y_p)\n",
@@ -959,7 +958,7 @@
         "def get_next_batch(ds):\n",
         "  itr = ds.make_one_shot_iterator()\n",
         "  image, label = itr.get_next()\n",
-        "  x = tf.to_float(image)/255.0\n",
+        "  x = tf.to_float(image) / 255.0\n",
         "  y = tf.one_hot(tf.squeeze(label), 10)\n",
         "  return x, y "
       ],