tf part1 student version

Author: avaamini

lab1/TF_Part1_Intro.ipynb

@@ -148,8 +148,7 @@
         "### Defining higher-order Tensors ###\n",
         "\n",
         "'''TODO: Define a 2-d Tensor'''\n",
-        "matrix = tf.constant([[1.0, 2.0, 3.0, 4.0], [5.0, 6.0, 7.0, 8.0]]) # TODO\n",
-        "# matrix = # TODO\n",
+        "matrix = # TODO\n",
         "\n",
         "assert isinstance(matrix, tf.Tensor), \"matrix must be a tf Tensor object\"\n",
         "assert tf.rank(matrix).numpy() == 2"
@@ -166,8 +165,7 @@
         "'''TODO: Define a 4-d Tensor.'''\n",
         "# Use tf.zeros to initialize a 4-d Tensor of zeros with size 10 x 256 x 256 x 3.\n",
         "#   You can think of this as 10 images where each image is RGB 256 x 256.\n",
-        "images = tf.zeros([10, 256, 256, 3]) # TODO\n",
-        "# images = # TODO\n",
+        "images = # TODO\n",
         "\n",
         "assert isinstance(images, tf.Tensor), \"matrix must be a tf Tensor object\"\n",
         "assert tf.rank(images).numpy() == 4, \"matrix must be of rank 4\"\n",
@@ -263,12 +261,9 @@
         "# Construct a simple computation function\n",
         "def func(a,b):\n",
         "  '''TODO: Define the operation for c, d, e (use tf.add, tf.subtract, tf.multiply).'''\n",
-        "  c = tf.add(a, b)\n",
-        "  # c = # TODO\n",
-        "  d = tf.subtract(b, 1)\n",
-        "  # d = # TODO\n",
-        "  e = tf.multiply(c, d)\n",
-        "  # e = # TODO\n",
+        "  c = # TODO\n",
+        "  d = # TODO\n",
+        "  e = # TODO\n",
         "  return e"
       ]
     },
@@ -349,12 +344,10 @@
         "\n",
         "  def call(self, x):\n",
         "    '''TODO: define the operation for z (hint: use tf.matmul)'''\n",
-        "    z = tf.matmul(x, self.W) + self.b # TODO\n",
-        "    # z = # TODO\n",
+        "    z = # TODO\n",
         "\n",
         "    '''TODO: define the operation for out (hint: use tf.sigmoid)'''\n",
-        "    y = tf.sigmoid(z) # TODO\n",
-        "    # y = # TODO\n",
+        "    y = # TODO\n",
         "    return y\n",
         "\n",
         "# Since layer parameters are initialized randomly, we will set a random seed for reproducibility\n",
@@ -402,8 +395,7 @@
         "# Remember: dense layers are defined by the parameters W and b!\n",
         "# You can read more about the initialization of W and b in the TF documentation :)\n",
         "# https://www.tensorflow.org/api_docs/python/tf/keras/layers/Dense?version=stable\n",
-        "dense_layer = Dense(n_output_nodes, activation='sigmoid') # TODO\n",
-        "# dense_layer = # TODO\n",
+        "dense_layer = # TODO\n",
         "\n",
         "# Add the dense layer to the model\n",
         "model.add(dense_layer)\n"
@@ -430,8 +422,7 @@
         "x_input = tf.constant([[1,2.]], shape=(1,2))\n",
         "\n",
         "'''TODO: feed input into the model and predict the output!'''\n",
-        "model_output = model(x_input).numpy()\n",
-        "# model_output = # TODO\n",
+        "model_output = # TODO\n",
         "print(model_output)"
       ]
     },
@@ -463,8 +454,7 @@
         "  def __init__(self, n_output_nodes):\n",
         "    super(SubclassModel, self).__init__()\n",
         "    '''TODO: Our model consists of a single Dense layer. Define this layer.'''\n",
-        "    self.dense_layer = Dense(n_output_nodes, activation='sigmoid') # TODO\n",
-        "    # self.dense_layer = '''TODO: Dense Layer'''\n",
+        "    self.dense_layer = '''TODO: Dense Layer'''\n",
         "\n",
         "  # In the call function, we define the Model's forward pass.\n",
         "  def call(self, inputs):\n",
@@ -529,13 +519,7 @@
         "\n",
         "  '''TODO: Implement the behavior where the network outputs the input, unchanged, under control of the isidentity argument.'''\n",
         "  def call(self, inputs, isidentity=False):\n",
-        "    x = self.dense_layer(inputs)\n",
-        "    if isidentity: # TODO\n",
-        "      return inputs # TODO\n",
-        "    return x\n",
-        "\n",
-        "  # def call(self, inputs, isidentity=False):\n",
-        "    # TODO"
+        "    ### TODO"
       ]
     },
     {
@@ -560,10 +544,8 @@
         "\n",
         "x_input = tf.constant([[1,2.]], shape=(1,2))\n",
         "'''TODO: pass the input into the model and call with and without the input identity option.'''\n",
-        "out_activate = model.call(x_input) # TODO\n",
-        "# out_activate = # TODO\n",
-        "out_identity = model.call(x_input, isidentity=True) # TODO\n",
-        "# out_identity = # TODO\n",
+        "out_activate = # TODO\n",
+        "out_identity = # TODO\n",
         "\n",
         "print(\"Network output with activation: {}; network identity output: {}\".format(out_activate.numpy(), out_identity.numpy()))"
       ]
@@ -658,8 +640,7 @@
         "for i in range(500):\n",
         "  with tf.GradientTape() as tape:\n",
         "    '''TODO: define the loss as described above'''\n",
-        "    loss = (x - x_f)**2 # \"forward pass\": record the current loss on the tape\n",
-        "    # loss = # TODO\n",
+        "    loss = # TODO\n",
         "\n",
         "  # loss minimization using gradient tape\n",
         "  grad = tape.gradient(loss, x) # compute the derivative of the loss with respect to x\n",
@@ -691,7 +672,7 @@
       "collapsed_sections": [
         "WBk0ZDWY-ff8"
       ],
-      "name": "TF_Part1_Intro.ipynb",
+      "name": "TF_Part1_Intro_Solution.ipynb",
       "provenance": []
     },
     "kernelspec": {
@@ -711,4 +692,4 @@
   },
   "nbformat": 4,
   "nbformat_minor": 0
-}
+}