[ENHANCEMENT] Fixes issue #19 partly

Author: varunp2k

Ch6/01_CNN_RNN_ATIS_intents.ipynb

@@ -1,13 +1,27 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Intent Detection Using CNN & RNN"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
     "colab_type": "text",
     "id": "Wz0jLNgTnMwF"
    },
    "source": [
-    "In this notebook we will demonstrate various CNN and RNN models for the task of intent detection on the ATIS dataset. "
+    "In this notebook we demonstrate various CNN and RNN architectures for the task of intent detection on the ATIS dataset. The ATIS dataset is a standard benchmark dataset for the tast of intent detection. ATIS Stands for Airline Travel Information System. The dataset can we found in the `Data2` folder under the `Data` folder."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Imports"
    ]
   },
   {
@@ -33,11 +47,18 @@
     }
    ],
    "source": [
-    "\n",
-    "#making the necessary imports\n",
+    "#general imports\n",
     "import os\n",
     "import sys\n",
+    "import random\n",
+    "random.seed(0) #for reproducability of results\n",
+    "\n",
+    "#basic imports\n",
     "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "\n",
+    "#NN imports\n",
     "from keras.preprocessing.text import Tokenizer\n",
     "from keras.preprocessing.sequence import pad_sequences\n",
     "from keras.utils import to_categorical\n",
@@ -46,13 +67,10 @@
     "from keras.models import Model, Sequential\n",
     "from keras.initializers import Constant\n",
     "\n",
+    "#encoder\n",
     "from sklearn.preprocessing import LabelEncoder\n",
     "\n",
-    "import random\n",
-    "random.seed(0)#for reproducability of results\n",
-    "\n",
-    "import pandas as pd\n",
-    "import numpy as np"
+    "\n"
    ]
   },
   {
@@ -62,7 +80,9 @@
     "id": "LGFhMpnjo4L1"
    },
    "source": [
-    "Loading the training data"
+    "## Data Loading\n",
+    "We load the data with the help of a few functions from `utils.py` which is included in this repository's Ch6 folder under folder name \"Data\". \n",
+    "### Training Data\n"
    ]
   },
   {
@@ -94,7 +114,6 @@
     }
    ],
    "source": [
-    "#utils is included in this repository'c Ch6 folder under folder name \"Data\"\n",
     "from Data.utils import fetch_data, read_method\n",
     "\n",
     "sents,labels,intents = fetch_data('Data/data2/atis.train.w-intent.iob')\n",
@@ -128,7 +147,7 @@
     "id": "qckNEPKRo8_V"
    },
    "source": [
-    "Loading the test data"
+    "### Testing Data"
    ]
   },
   {
@@ -200,7 +219,7 @@
     "id": "nUZsI3ZmpBA2"
    },
    "source": [
-    "Pre-Processing"
+    "## Data Preprocessing"
    ]
   },
   {
@@ -311,6 +330,15 @@
     "print('Splitting the train data into train and valid is done')"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Modeling\n",
+    "### Embedding Matrix\n",
+    "We need to prepare our embedding."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 12,
@@ -377,6 +405,13 @@
     "print(\"Preparing of embedding matrix is done\")"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### CNN with Pre-Trained Embeddings"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 13,
@@ -457,6 +492,14 @@
     "print('Test accuracy with CNN:', acc)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### CNN-Embedding Layer\n",
+    "Here, we train a CNN model with an embedding layer which is being trained on the fly instead of using the pre-trained embeddings."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 14,
@@ -536,6 +579,14 @@
     "print('Test accuracy with CNN:', acc)\n"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### RNN-Embedding Layer\n",
+    "Here, we train a RNN model with an embedding layer which is being trained on the fly instead of using the pre-trained embeddings."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 15,
@@ -601,6 +652,13 @@
     "print('Test accuracy with RNN:', acc)\n"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### LSTM with Pre-Trained Embeddings"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 16,
@@ -698,7 +756,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.8.3"
+   "version": "3.6.12"
   },
   "toc": {
    "base_numbering": 1,