remove link to script in tutorials and notebooks

docs/notebooks/gan.ipynb

@@ -11,11 +11,8 @@
     "They posit a deep generative model and they enable fast and accurate\n",
     "inferences.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/gan),\n",
-    "or as a script available at\n",
-    "[`examples/gan.py`](https://github.com/blei-lab/edward/blob/master/examples/gan.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/gan."
    ]
   },
   {
@@ -123,7 +120,6 @@
     }
    ],
    "source": [
-    "# DATA. MNIST batches are fed at training time.\n",
     "mnist = input_data.read_data_sets(DATA_DIR, one_hot=True)\n",
     "x_ph = tf.placeholder(tf.float32, [M, 784])"
    ]
@@ -168,7 +164,6 @@
     "  x = slim.fully_connected(h1, 784, activation_fn=tf.sigmoid)\n",
     "  return x\n",
     "\n",
-    "# MODEL\n",
     "with tf.variable_scope(\"Gen\"):\n",
     "  eps = Uniform(a=tf.zeros([M, d]) - 1.0, b=tf.ones([M, d]))\n",
     "  x = generative_network(eps)"

docs/notebooks/latent_space_models.ipynb

@@ -18,11 +18,8 @@
     "their distance in the latent space.\n",
     "\n",
     "We will analyze network data from neuroscience.\n",
-    "A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/latent-space-models),\n",
-    "or as a script available at\n",
-    "[`examples/latent_space_model.py`](https://github.com/blei-lab/edward/blob/master/examples/latent_space_model.py)\n",
-    "in the Github repository."
+    "A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/latent-space-models."
    ]
   },
   {
@@ -119,7 +116,6 @@
    },
    "outputs": [],
    "source": [
-    "# MODEL\n",
     "N = x_train.shape[0]  # number of data points\n",
     "K = 3  # latent dimensionality\n",
     "\n",

docs/notebooks/mixture_density_network.ipynb

@@ -10,11 +10,8 @@
     "of models obtained by combining a conventional neural network with a\n",
     "mixture density model.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/mixture-density-network),\n",
-    "or as a script available at\n",
-    "[`examples/mixture_density_network.py`](https://github.com/blei-lab/edward/blob/master/examples/mixture_density_network.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/mixture-density-network."
    ]
   },
   {
@@ -136,7 +133,6 @@
     "D = 1  # number of features\n",
     "K = 20  # number of mixture components\n",
     "\n",
-    "# DATA\n",
     "X_train, X_test, y_train, y_test = build_toy_dataset(N)\n",
     "print(\"Size of features in training data: {}\".format(X_train.shape))\n",
     "print(\"Size of output in training data: {}\".format(y_train.shape))\n",
@@ -199,7 +195,6 @@
     "  return mus, sigmas, logits\n",
     "\n",
     "\n",
-    "# MODEL\n",
     "mus, sigmas, logits = neural_network(X_ph)\n",
     "cat = Categorical(logits=logits)\n",
     "components = [Normal(mu=mu, sigma=sigma) for mu, sigma\n",
@@ -245,7 +240,6 @@
    },
    "outputs": [],
    "source": [
-    "# INFERENCE\n",
     "# There are no latent variables to infer. Thus inference is concerned\n",
     "# with only training model parameters, which are baked into how we\n",
     "# specify the neural networks.\n",
@@ -351,7 +345,6 @@
    },
    "outputs": [],
    "source": [
-    "# CRITICISM\n",
     "pred_weights, pred_means, pred_std = \\\n",
     "    sess.run([tf.nn.softmax(logits), mus, sigmas], feed_dict={X_ph: X_test})"
    ]

docs/notebooks/probabilistic_pca.ipynb

@@ -12,11 +12,8 @@
     "used when there are missing values in the data or for multidimensional\n",
     "scaling.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/probabilistic-pca),\n",
-    "or as a script available at\n",
-    "[`examples/probabilistic_pca.py`](https://github.com/blei-lab/edward/blob/master/examples/probabilistic_pca.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/probabilistic-pca."
    ]
   },
   {
@@ -89,8 +86,6 @@
     "D = 2  # data dimensionality\n",
     "K = 1  # latent dimensionality\n",
     "\n",
-    "# DATA\n",
-    "\n",
     "x_train = build_toy_dataset(N, D, K)"
    ]
   },
@@ -175,8 +170,6 @@
    },
    "outputs": [],
    "source": [
-    "# MODEL\n",
-    "\n",
     "w = Normal(mu=tf.zeros([D, K]), sigma=2.0 * tf.ones([D, K]))\n",
     "z = Normal(mu=tf.zeros([N, K]), sigma=tf.ones([N, K]))\n",
     "x = Normal(mu=tf.matmul(w, z, transpose_b=True), sigma=tf.ones([D, N]))"

docs/notebooks/supervised_classification.ipynb

@@ -10,11 +10,8 @@
     "labeled data, comprised of training examples $\\{(x_n, y_n)\\}$.\n",
     "Classification means the output $y$ takes discrete values.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/supervised-classification),\n",
-    "or as a script available at\n",
-    "[`examples/gp_classification.py`](https://github.com/blei-lab/edward/blob/master/examples/gp_classification.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/supervised-classification."
    ]
   },
   {
@@ -57,7 +54,6 @@
    "source": [
     "ed.set_seed(42)\n",
     "\n",
-    "# DATA\n",
     "df = np.loadtxt('data/crabs_train.txt', dtype='float32', delimiter=',')\n",
     "df[df[:, 0] == -1, 0] = 0  # replace -1 label with 0 label\n",
     "N = 25  # number of data points\n",

docs/notebooks/supervised_regression.ipynb

@@ -12,11 +12,8 @@
     "labeled data, comprised of training examples $\\{(x_n, y_n)\\}$.\n",
     "Regression typically means the output $y$ takes continuous values.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/supervised-regression),\n",
-    "or as a script available at\n",
-    "[`examples/bayesian_linear_regression.py`](https://github.com/blei-lab/edward/blob/master/examples/bayesian_linear_regression.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/supervised-regression."
    ]
   },
   {
@@ -74,7 +71,6 @@
     "N = 40  # number of data points\n",
     "D = 10  # number of features\n",
     "\n",
-    "# DATA\n",
     "w_true = np.random.randn(D) * 0.5\n",
     "X_train, y_train = build_toy_dataset(N, w_true)\n",
     "X_test, y_test = build_toy_dataset(N, w_true)"
@@ -123,7 +119,6 @@
    },
    "outputs": [],
    "source": [
-    "# MODEL\n",
     "X = tf.placeholder(tf.float32, [N, D])\n",
     "w = Normal(mu=tf.zeros(D), sigma=tf.ones(D))\n",
     "b = Normal(mu=tf.zeros(1), sigma=tf.ones(1))\n",
@@ -157,7 +152,6 @@
    },
    "outputs": [],
    "source": [
-    "# INFERENCE\n",
     "qw = Normal(mu=tf.Variable(tf.random_normal([D])),\n",
     "            sigma=tf.nn.softplus(tf.Variable(tf.random_normal([D]))))\n",
     "qb = Normal(mu=tf.Variable(tf.random_normal([1])),\n",
@@ -231,7 +225,6 @@
    },
    "outputs": [],
    "source": [
-    "# CRITICISM\n",
     "y_post = ed.copy(y, {w: qw, b: qb})\n",
     "# This is equivalent to\n",
     "# y_post = Normal(mu=ed.dot(X, qw) + qb, sigma=tf.ones(N))"

docs/notebooks/unsupervised.ipynb

@@ -9,11 +9,8 @@
     "In unsupervised learning, the task is to infer hidden structure from\n",
     "unlabeled data, comprised of training examples $\\{x_n\\}$.\n",
     "\n",
-    "We demonstrate with an example in Edward. A webpage version is available \n",
-    "[here](http://edwardlib.org/tutorials/unsupervised),\n",
-    "or as a script available at\n",
-    "[`examples/mixture_gaussian_collapsed.py`](https://github.com/blei-lab/edward/blob/master/examples/mixture_gaussian_collapsed.py)\n",
-    "in the Github repository."
+    "We demonstrate with an example in Edward. A webpage version is available at\n",
+    "http://edwardlib.org/tutorials/unsupervised."
    ]
   },
   {
@@ -77,7 +74,6 @@
     "D = 2  # dimensionality of data\n",
     "ed.set_seed(42)\n",
     "\n",
-    "# DATA\n",
     "x_train = build_toy_dataset(N)"
    ]
   },
@@ -194,7 +190,6 @@
    },
    "outputs": [],
    "source": [
-    "# MODEL\n",
     "mu = Normal(mu=tf.zeros([K, D]), sigma=tf.ones([K, D]))\n",
     "sigma = InverseGamma(alpha=tf.ones([K, D]), beta=tf.ones([K, D]))\n",
     "cat = Categorical(logits=tf.zeros([N, K]))\n",
@@ -241,7 +236,6 @@
    },
    "outputs": [],
    "source": [
-    "# INFERENCE\n",
     "qmu = Normal(\n",
     "    mu=tf.Variable(tf.random_normal([K, D])),\n",
     "    sigma=tf.nn.softplus(tf.Variable(tf.zeros([K, D]))))\n",

docs/tex/getting-started.tex

@@ -28,9 +28,7 @@ \subsubsection{Your first Edward program}
 Here we will show a Bayesian neural network. It is a neural network
 with a prior distribution on its weights.
 (This example is abridged; an interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/getting_started.ipynb}{here},
-or as a script available
-\href{https://github.com/blei-lab/edward/blob/master/examples/getting_started_example.py}{here}.)
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/getting_started.ipynb}{here}.)
 
 First, simulate a toy dataset of 50 observations with a cosine relationship.
 

docs/tex/tutorials/gan.tex

@@ -9,10 +9,7 @@ \subsection{Generative Adversarial Networks}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/gan.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/gan.py}
-{\texttt{examples/gan.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/gan.ipynb}{here}.
 
 \begin{lstlisting}[language=Python]
 M = 128  # batch size during training

docs/tex/tutorials/latent-space-models.tex

@@ -15,10 +15,7 @@ \subsection{Latent Space Models for Neural Data}
 
 We will analyze network data from neuroscience.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/latent_space_models.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/latent_space_model.py}
-{\texttt{examples/latent_space_model.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/latent_space_models.ipynb}{here}.
 
 \subsubsection{Data}
 

docs/tex/tutorials/mixture-density-network.tex

@@ -8,10 +8,7 @@ \subsection{Mixture Density Networks}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/mixture_density_network.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/mixture_density_network.py}
-{\texttt{examples/mixture_density_network.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/mixture_density_network.ipynb}{here}.
 
 \subsubsection{Data}
 

docs/tex/tutorials/probabilistic-pca.tex

@@ -10,10 +10,7 @@ \subsection{Probabilistic PCA}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/probabilistic_pca.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/probabilistic_pca.py}
-{\texttt{examples/probabilistic_pca.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/probabilistic_pca.ipynb}{here}.
 
 \subsubsection{Data}
 

docs/tex/tutorials/supervised-classification.tex

@@ -8,10 +8,7 @@ \subsection{Supervised Learning (Classification)}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/supervised_classification.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/gp_classification.py}
-{\texttt{examples/gp_classification.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/supervised_classification.ipynb}{here}.
 
 \subsubsection{Data}
 

docs/tex/tutorials/supervised-regression.tex

@@ -8,10 +8,7 @@ \subsection{Supervised Learning (Regression)}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/supervised_regression.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/bayesian_linear_regression.py}
-{\texttt{examples/bayesian_linear_regression.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/supervised_regression.ipynb}{here}.
 
 \subsubsection{Data}
 
@@ -120,10 +117,10 @@ \subsubsection{Criticism}
 \end{lstlisting}
 
 \begin{lstlisting}
-Mean squared error on test data:
-0.0300492
-Mean absolute error on test data:
-0.123616
+## Mean squared error on test data:
+## 0.0300492
+## Mean absolute error on test data:
+## 0.123616
 \end{lstlisting}
 
 The trained model makes predictions with low error

docs/tex/tutorials/unsupervised.tex

@@ -7,10 +7,7 @@ \subsection{Unsupervised Learning}
 
 We demonstrate with an example in Edward.
 An interactive version with Jupyter notebook is available
-\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/unsupervised.ipynb}{here},
-or as a script available at
-\href{https://github.com/blei-lab/edward/blob/master/examples/mixture_gaussian_collapsed.py}
-{\texttt{examples/mixture_gaussian_collapsed.py}} in the Github repository.
+\href{http://nbviewer.jupyter.org/github/blei-lab/edward/blob/master/docs/notebooks/unsupervised.ipynb}{here}.
 
 \subsubsection{Data}