Simplify the TensorFlow basics section.

1. Split "TensorFlow basics" from "TensorFlow in depth"
2. Compress `eager.ipynb` to a quick overview of "TF Basics".
3. "Effective TF2" is written for TF1 users.
4. Move everything else to a new section.

PiperOrigin-RevId: 414830812

Author: MarkDaoust

site/en/guide/_toc.yaml

@@ -3,44 +3,48 @@ toc:
   path: /guide/
 
 - heading: "TensorFlow basics"
-- title: "Eager execution"
-  path: /guide/eager
-- title: "Tensor"
+- title: "Overview"
+  path: /guide/basics
+- title: "Tensors"
   path: /guide/tensor
-- title: "Variable"
+- title: "Variables"
   path: /guide/variable
 - title: "Automatic differentiation"
   path: /guide/autodiff
-- title: "Intro to graphs and functions"
+- title: "Graphs and functions"
   path: /guide/intro_to_graphs
-- title: "Intro to modules, layers, and models"
+- title: "Modules, layers, and models"
   path: /guide/intro_to_modules
 - title: "Training loops"
   path: /guide/basic_training_loops
+
+- heading: "Keras"
+- include: /guide/keras/_toc.yaml
+
+- heading: "TensorFlow in depth"
+- title: "Tensor slicing"
+  path: /guide/tensor_slicing
 - title: "Advanced autodiff"
   path: /guide/advanced_autodiff
 - title: "Ragged tensor"
   path: /guide/ragged_tensor
 - title: "Sparse tensor"
   path: /guide/sparse_tensor
+
+- title: "Random number generation"
+  path: /guide/random_numbers
 - title: "NumPy API"
+  status: experimental
   path: /guide/tf_numpy
-- title: "Tensor slicing"
-  path: /guide/tensor_slicing
 - title: "Thinking in TensorFlow 2"
   path: /guide/effective_tf2
 
-- heading: "Keras"
-- include: /guide/keras/_toc.yaml
-
 - heading: "Customization"
 - title: "Create an op"
   path: /guide/create_op
-- title: "Extension type"
+- title: "Extension types"
   path: /guide/extension_type
   status: experimental
-- title: "Random number generation"
-  path: /guide/random_numbers
 
 - heading: "Data input pipelines"
 - title: "tf.data"

site/en/guide/basic_training_loops.ipynb

@@ -90,7 +90,11 @@
       },
       "outputs": [],
       "source": [
-        "import tensorflow as tf"
+        "import tensorflow as tf\n",
+        "\n",
+        "import matplotlib.pyplot as plt\n",
+        "\n",
+        "colors = plt.rcParams['axes.prop_cycle'].by_key()['color']"
       ]
     },
     {
@@ -142,16 +146,20 @@
         "TRUE_W = 3.0\n",
         "TRUE_B = 2.0\n",
         "\n",
-        "NUM_EXAMPLES = 1000\n",
+        "NUM_EXAMPLES = 201\n",
         "\n",
         "# A vector of random x values\n",
-        "x = tf.random.normal(shape=[NUM_EXAMPLES])\n",
+        "x = tf.linspace(-2,2, NUM_EXAMPLES)\n",
+        "x = tf.cast(x, tf.float32)\n",
+        "\n",
+        "def f(x):\n",
+        "  return x * TRUE_W + TRUE_B\n",
         "\n",
         "# Generate some noise\n",
         "noise = tf.random.normal(shape=[NUM_EXAMPLES])\n",
         "\n",
         "# Calculate y\n",
-        "y = x * TRUE_W + TRUE_B + noise"
+        "y = f(x) + noise"
       ]
     },
     {
@@ -163,9 +171,7 @@
       "outputs": [],
       "source": [
         "# Plot all the data\n",
-        "import matplotlib.pyplot as plt\n",
-        "\n",
-        "plt.scatter(x, y, c=\"b\")\n",
+        "plt.plot(x, y, '.')\n",
         "plt.show()"
       ]
     },
@@ -271,8 +277,10 @@
       },
       "outputs": [],
       "source": [
-        "plt.scatter(x, y, c=\"b\")\n",
-        "plt.scatter(x, model(x), c=\"r\")\n",
+        "plt.plot(x, y, '.', label=\"Data\")\n",
+        "plt.plot(x, f(x), label=\"Ground truth\")\n",
+        "plt.plot(x, model(x), label=\"Predictions\")\n",
+        "plt.legend()\n",
         "plt.show()\n",
         "\n",
         "print(\"Current loss: %1.6f\" % loss(y, model(x)).numpy())"
@@ -341,23 +349,37 @@
         "model = MyModel()\n",
         "\n",
         "# Collect the history of W-values and b-values to plot later\n",
-        "Ws, bs = [], []\n",
+        "weights = []\n",
+        "biases = []\n",
         "epochs = range(10)\n",
         "\n",
         "# Define a training loop\n",
+        "def report(model, loss):\n",
+        "  return f\"W = {model.w.numpy():1.2f}, b = {model.b.numpy():1.2f}, loss={current_loss:2.5f}\"\n",
+        "\n",
+        "\n",
         "def training_loop(model, x, y):\n",
         "\n",
         "  for epoch in epochs:\n",
         "    # Update the model with the single giant batch\n",
         "    train(model, x, y, learning_rate=0.1)\n",
         "\n",
         "    # Track this before I update\n",
-        "    Ws.append(model.w.numpy())\n",
-        "    bs.append(model.b.numpy())\n",
+        "    weights.append(model.w.numpy())\n",
+        "    biases.append(model.b.numpy())\n",
         "    current_loss = loss(y, model(x))\n",
         "\n",
-        "    print(\"Epoch %2d: W=%1.2f b=%1.2f, loss=%2.5f\" %\n",
-        "          (epoch, Ws[-1], bs[-1], current_loss))\n"
+        "    print(f\"Epoch {epoch:2d}:\")\n",
+        "    print(\"    \", report(model, current_loss))"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "8dKKLU4KkQEq"
+      },
+      "source": [
+        "Do the training"
       ]
     },
     {
@@ -368,21 +390,50 @@
       },
       "outputs": [],
       "source": [
-        "print(\"Starting: W=%1.2f b=%1.2f, loss=%2.5f\" %\n",
-        "      (model.w, model.b, loss(y, model(x))))\n",
+        "current_loss = loss(y, model(x))\n",
         "\n",
-        "# Do the training\n",
-        "training_loop(model, x, y)\n",
+        "print(f\"Starting:\")\n",
+        "print(\"    \", report(model, current_loss))\n",
         "\n",
-        "# Plot it\n",
-        "plt.plot(epochs, Ws, \"r\",\n",
-        "         epochs, bs, \"b\")\n",
+        "training_loop(model, x, y)"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "JPJgimg8kSA4"
+      },
+      "source": [
+        "Plot the evolution of the weights over time:"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {
+        "id": "ND1fQw8sbTNr"
+      },
+      "outputs": [],
+      "source": [
+        "plt.plot(epochs, weights, label='Weights', color=colors[0])\n",
+        "plt.plot(epochs, [TRUE_W] * len(epochs), '--',\n",
+        "         label = \"True weight\", color=colors[0])\n",
         "\n",
-        "plt.plot([TRUE_W] * len(epochs), \"r--\",\n",
-        "         [TRUE_B] * len(epochs), \"b--\")\n",
+        "plt.plot(epochs, biases, label='bias', color=colors[1])\n",
+        "plt.plot(epochs, [TRUE_B] * len(epochs), \"--\",\n",
+        "         label=\"True bias\", color=colors[1])\n",
         "\n",
-        "plt.legend([\"W\", \"b\", \"True W\", \"True b\"])\n",
-        "plt.show()\n"
+        "plt.legend()\n",
+        "plt.show()"
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {
+        "id": "zhlwj1ojkcUP"
+      },
+      "source": [
+        "Visualize how the trained model performs"
       ]
     },
     {
@@ -393,9 +444,10 @@
       },
       "outputs": [],
       "source": [
-        "# Visualize how the trained model performs\n",
-        "plt.scatter(x, y, c=\"b\")\n",
-        "plt.scatter(x, model(x), c=\"r\")\n",
+        "plt.plot(x, y, '.', label=\"Data\")\n",
+        "plt.plot(x, f(x), label=\"Ground truth\")\n",
+        "plt.plot(x, model(x), label=\"Predictions\")\n",
+        "plt.legend()\n",
         "plt.show()\n",
         "\n",
         "print(\"Current loss: %1.6f\" % loss(model(x), y).numpy())"
@@ -531,8 +583,7 @@
     "colab": {
       "collapsed_sections": [
         "5rmpybwysXGV",
-        "iKD__8kFCKNt",
-        "vPnIVuaSJwWz"
+        "iKD__8kFCKNt"
       ],
       "name": "basic_training_loops.ipynb",
       "toc_visible": true