diff --git a/section-4-unstructured-data-projects/end-to-end-dog-vision-v2.ipynb b/section-4-unstructured-data-projects/end-to-end-dog-vision-v2.ipynb
index b68c4892e..177d21df6 100644
--- a/section-4-unstructured-data-projects/end-to-end-dog-vision-v2.ipynb
+++ b/section-4-unstructured-data-projects/end-to-end-dog-vision-v2.ipynb
@@ -2851,7 +2851,7 @@
     "id": "WYwlNwfjdMEp"
    },
    "source": [
-    "## TK - 4. Creating training and test data split directories\n",
+    "## 4. Creating training and test data split directories\n",
     "\n",
     "After exploring the data, one of the next best things you can do is create experimental data splits.\n",
     "\n",
@@ -3294,7 +3294,7 @@
     "id": "QUYK0GNLS7on"
    },
    "source": [
-    "### TK - Making a 10% training dataset split\n",
+    "### Making a 10% training dataset split\n",
     "\n",
     "We've already split the data into training and test sets, so why might we want to make another split?\n",
     "\n",
@@ -3966,7 +3966,7 @@
     "id": "TUlSbEnIh-PO"
    },
    "source": [
-    "## TK - 5. Turning datasets into TensorFlow Dataset(s)\n",
+    "## 5. Turning datasets into TensorFlow Dataset(s)\n",
     "\n",
     "Alright, we've spent a bunch of time getting our dog images into different folders.\n",
     "\n",
@@ -4298,7 +4298,7 @@
     "id": "pclcbSv3VPv2"
    },
    "source": [
-    "### TK - Visualizing images from our TensorFlow Dataset\n",
+    "### Visualizing images from our TensorFlow Dataset\n",
     "\n",
     "Let's follow the data explorer's motto once again and *visualize, visualize, visualize!*\n",
     "\n",
@@ -4399,7 +4399,7 @@
     "id": "07e-LqpXwu5K"
    },
    "source": [
-    "### TK - Getting labels from our TensorFlow Dataset\n",
+    "### Getting labels from our TensorFlow Dataset\n",
     "\n",
     "Since our data is now in `tf.data.Dataset` format, there are a couple of important attributes we can pull from it if necessary.\n",
     "\n",
@@ -4508,7 +4508,7 @@
     "id": "vTE3j1OKnJ6k"
    },
    "source": [
-    "### TK - Configuring our datasets for performance\n",
+    "### Configuring our datasets for performance\n",
     "\n",
     "There's one last step we're going to do before we build our first TensorFlow model.\n",
     "\n",
@@ -4587,7 +4587,7 @@
     "id": "teWeV33en5Cq"
    },
    "source": [
-    "## TK - 6. Creating a neural network with TensorFlow\n",
+    "## 6. Creating a neural network with TensorFlow\n",
     "\n",
     "We've spent lots of time preparing the data.\n",
     "\n",
@@ -4605,9 +4605,11 @@
     "\n",
     "A neural network often follows the structure of:\n",
     "\n",
-    "Input layer -> Middle layer(s) -> Output layer\n",
+    "Input layer -> Middle layer(s) -> Output layer.\n",
     "\n",
-    "TK - image of neural network with example\n",
+    "<img src=\"https://github.com/mrdbourke/zero-to-mastery-ml/blob/master/images/unstructured-data-anatomy-of-a-neural-network.png?raw=true\" width=750 alt=\"This image illustrates a simple neural network diagram with one input layer, one hidden layer, and one output layer. The input layer has 2 neurons, the hidden layer has 3 neurons, and the output layer has 1 neuron. The diagram shows connections between neurons across layers. There are annotations indicating that the neural network can be customized for almost any kind of input and output, and although the number and type of hidden layers are infinitely customizable, there is often a practical upper limit. The output layer represents learned data or prediction probabilities. The overall theme conveys the flexibility and adaptability of neural network architectures.\"/>\n",
+    "\n",
+    "*General anatomy of a neural network. Neural networks are almost infinitely customisable. The main premise is that data goes in one end, gets manipulated by many small functions in an attempt to learn patterns/weights which represent the data to produce useful outputs. Note that \"patterns\" is an arbitrary term, you’ll often hear \"embedding\", \"weights\", \"feature representation\", \"representation\" all referring to similar things.*\n",
     "\n",
     "Where the input layer takes in the data, the middle layer(s) perform calculations on the data and (hopefully) learn patterns (also called weights/biases) to represent the data and the output layer performs a final transformation on the learned patterns to make them usable in human applications.\n",
     "\n",
@@ -4640,6 +4642,8 @@
    "source": [
     "### TK - The magic of transfer learning\n",
     "\n",
+    "UPTOHERE\n",
+    "\n",
     "**Transfer learning** is the process of getting an existing working model and adjusting it to your own problem.\n",
     "\n",
     "This works particularly well for neural networks.\n",
@@ -4664,7 +4668,7 @@
     "* [`Hugging Face Models Hub`](https://huggingface.co/models) - A large collection of pretrained models on a wide range on tasks, from computer vision to natural language processing to audio processing.  \n",
     "* [`Kaggle Models`](https://www.kaggle.com/models) - A huge collection of different pretrained models for many different tasks.\n",
     "\n",
-    "TK - image of where to find pretrained models\n",
+    "TK - image/table of where to find pretrained models\n",
     "\n",
     "> **Note:** For most new machine learning problems, if you're looking to get good results quickly, you should generally look for a pretrained model similar to your problem and use transfer learning to adapt it to your own domain.\n",
     "\n",