typo fix and correction for state processor output shape

Author: ayberkydn

DQN/Deep Q Learning Solution.ipynb

@@ -79,7 +79,7 @@
     "            state: A [210, 160, 3] Atari RGB State\n",
     "\n",
     "        Returns:\n",
-    "            A processed [84, 84, 1] state representing grayscale values.\n",
+    "            A processed [84, 84] state representing grayscale values.\n",
     "        \"\"\"\n",
     "        return sess.run(self.output, { self.input_state: state })"
    ]
@@ -144,7 +144,7 @@
     "        gather_indices = tf.range(batch_size) * tf.shape(self.predictions)[1] + self.actions_pl\n",
     "        self.action_predictions = tf.gather(tf.reshape(self.predictions, [-1]), gather_indices)\n",
     "\n",
-    "        # Calcualte the loss\n",
+    "        # Calculate the loss\n",
     "        self.losses = tf.squared_difference(self.y_pl, self.action_predictions)\n",
     "        self.loss = tf.reduce_mean(self.losses)\n",
     "\n",

DQN/Deep Q Learning.ipynb

@@ -137,7 +137,7 @@
     "        gather_indices = tf.range(batch_size) * tf.shape(self.predictions)[1] + self.actions_pl\n",
     "        self.action_predictions = tf.gather(tf.reshape(self.predictions, [-1]), gather_indices)\n",
     "\n",
-    "        # Calcualte the loss\n",
+    "        # Calculate the loss\n",
     "        self.losses = tf.squared_difference(self.y_pl, self.action_predictions)\n",
     "        self.loss = tf.reduce_mean(self.losses)\n",
     "\n",

DQN/Double DQN Solution.ipynb

@@ -73,7 +73,7 @@
     "            state: A [210, 160, 3] Atari RGB State\n",
     "\n",
     "        Returns:\n",
-    "            A processed [84, 84, 1] state representing grayscale values.\n",
+    "            A processed [84, 84] state representing grayscale values.\n",
     "        \"\"\"\n",
     "        return sess.run(self.output, { self.input_state: state })"
    ]

DQN/dqn.py

@@ -39,7 +39,7 @@ def process(self, sess, state):
             state: A [210, 160, 3] Atari RGB State
 
         Returns:
-            A processed [84, 84, 1] state representing grayscale values.
+            A processed [84, 84] state representing grayscale values.
         """
         return sess.run(self.output, { self.input_state: state })
 
@@ -95,7 +95,7 @@ def _build_model(self):
         gather_indices = tf.range(batch_size) * tf.shape(self.predictions)[1] + self.actions_pl
         self.action_predictions = tf.gather(tf.reshape(self.predictions, [-1]), gather_indices)
 
-        # Calcualte the loss
+        # Calculate the loss
         self.losses = tf.squared_difference(self.y_pl, self.action_predictions)
         self.loss = tf.reduce_mean(self.losses)