Merge branch 'master' of https://github.com/lexfridman/mit-deep-learning

amit140381 · Feb 17, 2019 · 68089a7 · 68089a7
2 parents a4a6642 + 5e58a70
commit 68089a7
Showing 1 changed file with 29 additions and 30 deletions.
diff --git a/tutorial_deep_learning_basics/deep_learning_basics.ipynb b/tutorial_deep_learning_basics/deep_learning_basics.ipynb
@@ -11,7 +11,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "S5Uhzt6vVIB2"
+    "id": ""
    },
    "source": [
     "<table align=\"center\">\n",
@@ -33,7 +33,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "FbVhjPpzn6BM"
+    "id": ""
    },
    "source": [
     "# Deep Learning Basics\n",
@@ -80,7 +80,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "dzLKpmZICaWN"
+    "id": ""
    },
    "outputs": [
     {
@@ -149,7 +149,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "59veuiEZCaW4"
+    "id": ""
    },
    "source": [
     "### Build the model\n",
@@ -169,7 +169,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "Lhan11blCaW7"
+    "id": ""
    },
    "outputs": [],
    "source": [
@@ -189,7 +189,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "qKF6uW-BCaW-"
+    "id": ""
    },
    "source": [
     "### Train the model\n",
@@ -209,7 +209,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "xvwvpA64CaW_"
+    "id": ""
    },
    "outputs": [
     {
@@ -295,7 +295,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "oEw4bZgGCaXB"
+    "id": ""
    },
    "source": [
     "Next, compare how the model performs on the test dataset:"
@@ -307,7 +307,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "VflXLEeECaXC"
+    "id": ""
    },
    "outputs": [
     {
@@ -330,7 +330,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "yWfgsmVXCaXG"
+    "id": ""
    },
    "source": [
     "Compare the RMSE measure you get to the [Kaggle leaderboard](https://www.kaggle.com/c/boston-housing/leaderboard). An RMSE of 2.651 puts us in 5th place."
@@ -360,7 +360,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "DLdCchMdCaWQ"
+    "id": ""
    },
    "source": [
     "The MNIST dataset containss 70,000 grayscale images of handwritten digits at a resolution of 28 by 28 pixels. The task is to take one of these images as input and predict the most likely digit contained in the image (along with a relative confidence in this prediction):\n",
@@ -376,7 +376,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "7MqDQO0KCaWS"
+    "id": ""
    },
    "outputs": [],
    "source": [
@@ -391,7 +391,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "Wz7l27Lz9S1P"
+    "id": ""
    },
    "source": [
     "We scale these values to a range of 0 to 1 before feeding to the neural network model. For this, we divide the values by 255. It's important that the *training set* and the *testing set* are preprocessed in the same way:"
@@ -403,7 +403,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "bW5WzIPlCaWv"
+    "id": ""
    },
    "outputs": [],
    "source": [
@@ -420,7 +420,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "Ee638AlnCaWz"
+    "id": ""
    },
    "source": [
     "Display the first 5 images from the *training set* and display the class name below each image. Verify that the data is in the correct format and we're ready to build and train the network."
@@ -432,7 +432,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "oZTImqg_CaW1"
+    "id": ""
    },
    "outputs": [
     {
@@ -461,7 +461,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "59veuiEZCaW4"
+    "id": ""
    },
    "source": [
     "### Build the model\n",
@@ -475,7 +475,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "9ODch-OFCaW4"
+    "id": ""
    },
    "outputs": [],
    "source": [
@@ -502,7 +502,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "gut8A_7rCaW6"
+    "id": ""
    },
    "source": [
     "Before the model is ready for training, it needs a few more settings. These are added during the model's *compile* step:\n",
@@ -518,7 +518,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "Lhan11blCaW7"
+    "id": ""
    },
    "outputs": [],
    "source": [
@@ -531,7 +531,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "qKF6uW-BCaW-"
+    "id": ""
    },
    "source": [
     "### Train the model\n",
@@ -551,7 +551,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "xvwvpA64CaW_"
+    "id": ""
    },
    "outputs": [
     {
@@ -579,7 +579,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "W3ZVOhugCaXA"
+    "id": ""
    },
    "source": [
     "As the model trains, the loss and accuracy metrics are displayed. This model reaches an accuracy of about 98.68% on the training data."
@@ -589,7 +589,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "oEw4bZgGCaXB"
+    "id": ""
    },
    "source": [
     "### Evaluate accuracy\n",
@@ -603,7 +603,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "VflXLEeECaXC"
+    "id": ""
    },
    "outputs": [
     {
@@ -627,7 +627,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "yWfgsmVXCaXG"
+    "id": ""
    },
    "source": [
     "Often times, the accuracy on the test dataset is a little less than the accuracy on the training dataset. This gap between training accuracy and test accuracy is an example of *overfitting*. In our case, the accuracy is better at 99.19%! This is, in part, due to successful regularization accomplished with the Dropout layers."
@@ -637,7 +637,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "xsoS7CPDCaXH"
+    "id": ""
    },
    "source": [
     "### Make predictions\n",
@@ -653,7 +653,7 @@
    "metadata": {
     "colab": {},
     "colab_type": "code",
-    "id": "Gl91RPhdCaXI"
+    "id": ""
    },
    "outputs": [
     {
@@ -771,7 +771,7 @@
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
-    "id": "x9Kk1voUCaXJ"
+    "id": ""
    },
    "source": [
     "The above shows the prediction of the network by choosing the neuron with the highest output. While the output layer values add 1 to one, these do not reflect well-calibrated measures of \"uncertainty\". Often, the network is overly confident about the top choice that does not reflect a learned measure of probability. If everything ran correctly you should get an animation like this:\n",
@@ -793,7 +793,6 @@
   "colab": {
    "collapsed_sections": [],
    "name": "tutorial_deep_learning_basics.ipynb",
-   "private_outputs": true,
    "provenance": [],
    "toc_visible": true,
    "version": "0.3.2"