notebook/modeling/05_cnn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CNN"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Initially I started with a simple CNN in order to find out where the base performance of a CNN could be around."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "%load_ext tensorboard"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pandas import read_parquet\n",
    "from data import file\n",
    "from reporting.evaluation import plot_confusion_matrix\n",
    "import json"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_train = read_parquet(file.news_articles_cleaned_train)\n",
    "data_test = read_parquet(file.news_articles_cleaned_test)\n",
    "data_val = read_parquet(file.news_articles_cleaned_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Dense, Conv1D, GlobalMaxPooling1D, InputLayer\n",
    "from tensorflow.keras import Sequential\n",
    "import os\n",
    "import datetime\n",
    "from tensorflow import keras\n",
    "from keras.callbacks import TensorBoard\n",
    "from preprocessing.text import extract_vocabulary\n",
    "from models.text.layers import create_text_vectorization, create_word_embedding\n",
    "from sklearn.preprocessing import LabelBinarizer\n",
    "    \n",
    "    \n",
    "def build_model(X_train, y_train, conv_num_filters=128, conv_kernel_size=7):\n",
    "\n",
    "    vocabulary, embedding_length = extract_vocabulary(X_train, verbose=True)\n",
    "    vectorize_layer = create_text_vectorization(vocabulary)\n",
    "    embedding_layer = create_word_embedding(vocabulary, embedding_length)\n",
    "\n",
    "    output_classes = len(label_binarizer.classes_)\n",
    "    \n",
    "    model = Sequential(name=\"cnn\")\n",
    "    model.add(InputLayer(input_shape=(1,), dtype=tf.string, name=\"text_input\"))\n",
    "    model.add(vectorize_layer)\n",
    "    model.add(embedding_layer)\n",
    "    model.add(Conv1D(conv_num_filters, conv_kernel_size, activation=\"relu\", strides=1, padding=\"valid\", name=\"conv_1\"))\n",
    "    model.add(GlobalMaxPooling1D(name=\"global_max_pool_1\"))\n",
    "    model.add(Dense(output_classes, activation=tf.nn.softmax, name=\"prediction\"))\n",
    "\n",
    "    return (model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = data_train.text_tokenized_stemmed\n",
    "y_train = data_train.label\n",
    "\n",
    "X_test = data_test.text_tokenized_stemmed\n",
    "y_test = data_test.label\n",
    "\n",
    "X_val = data_val.text_tokenized_stemmed\n",
    "y_val = data_val.label"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Median sequence length:       : 171\n",
      "Percentil                     : 0.98)\n",
      "Cutoff sequence length        : 588\n",
      "Max sequence length           : 2590\n",
      "Embedding length              : 588\n",
      "Vocabulary length             : 139291\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2021-11-23 05:30:05.918033: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:06.022624: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:06.023568: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:06.026438: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2021-11-23 05:30:06.027991: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:06.028809: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:06.029479: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:08.091662: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:08.092484: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:08.093122: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
      "2021-11-23 05:30:08.094344: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 13839 MB memory:  -> device: 0, name: Tesla T4, pci bus id: 0000:00:04.0, compute capability: 7.5\n",
      "Warning : `load_model` does not return WordVectorModel or SupervisedModel any more, but a `FastText` object which is very similar.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"cnn\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "text_vectorization (TextVect (None, None)              0         \n",
      "_________________________________________________________________\n",
      "embedding (Embedding)        (None, None, 300)         41787300  \n",
      "_________________________________________________________________\n",
      "conv_1 (Conv1D)              (None, None, 128)         268928    \n",
      "_________________________________________________________________\n",
      "global_max_pool_1 (GlobalMax (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "prediction (Dense)           (None, 9)                 1161      \n",
      "=================================================================\n",
      "Total params: 42,057,389\n",
      "Trainable params: 270,089\n",
      "Non-trainable params: 41,787,300\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from preprocessing.categorical import binarizer\n",
    "label_binarizer = binarizer(y_train)\n",
    "model = build_model(X_train, y_train)\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "from keras.losses import CategoricalCrossentropy\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "\n",
    "learning_rate=0.01\n",
    "model_metric = [\"accuracy\"]\n",
    "loss_function = CategoricalCrossentropy()\n",
    "optimizer = Adam(learning_rate=learning_rate)\n",
    "\n",
    "model.compile(loss=loss_function, optimizer=optimizer, metrics=model_metric)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['International', 'Inland', 'Kultur', 'Panorama', 'Web', 'Sport',\n",
       "       'Wirtschaft', 'Wissenschaft', 'Etat'], dtype=object)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_train.label.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [],
   "source": [
    "\n",
    "x_train = data_train['text_stem']\n",
    "y_train_bin = label_binarizer.transform(y_train)\n",
    "\n",
    "x_test = data_test['text_stem']\n",
    "y_test_bin = label_binarizer.transform(y_test)\n",
    "\n",
    "x_val = data_val['text_stem']\n",
    "y_val_bin = label_binarizer.transform(y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false,
    "jupyter": {
     "outputs_hidden": false
    },
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<PrefetchDataset shapes: ((None,), (None, 9)), types: (tf.string, tf.int64)>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "batch_size = 32\n",
    "train_input = tf.data.Dataset.from_tensor_slices((x_train, y_train_bin)).batch(batch_size).prefetch(tf.data.AUTOTUNE)\n",
    "test_input = tf.data.Dataset.from_tensor_slices((x_test, y_test_bin)).batch(batch_size).prefetch(tf.data.AUTOTUNE)\n",
    "train_input\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2021-11-23 05:31:13.220556: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n",
      "2021-11-23 05:31:14.387161: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "226/226 [==============================] - 65s 226ms/step - loss: 1.7248 - accuracy: 0.4056 - val_loss: 1.4763 - val_accuracy: 0.4903\n",
      "Epoch 2/5\n",
      "226/226 [==============================] - 34s 152ms/step - loss: 1.2039 - accuracy: 0.5945 - val_loss: 1.4238 - val_accuracy: 0.5291\n",
      "Epoch 3/5\n",
      "226/226 [==============================] - 35s 155ms/step - loss: 0.8033 - accuracy: 0.7345 - val_loss: 1.5678 - val_accuracy: 0.5388\n",
      "Epoch 4/5\n",
      "226/226 [==============================] - 38s 167ms/step - loss: 0.4985 - accuracy: 0.8358 - val_loss: 2.0620 - val_accuracy: 0.4981\n",
      "Epoch 5/5\n",
      "226/226 [==============================] - 35s 155ms/step - loss: 0.3332 - accuracy: 0.8880 - val_loss: 1.9736 - val_accuracy: 0.5369\n"
     ]
    }
   ],
   "source": [
    "#%tensorboard --logdir logs/fit\n",
    "\n",
    "#callbacks = [TensorBoard(\"logs/fit\", histogram_freq=1)]\n",
    "callbacks = []\n",
    "history = model.fit(train_input, validation_data=test_input, callbacks=callbacks, epochs=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from reporting.training import plot_history\n",
    "    \n",
    "plot_history(history)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "pandas.core.series.Series"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "count = 100\n",
    "expected = y_val[0:count]\n",
    "predicted = label_binarizer.inverse_transform(model.predict(x_val[0:count]))\n",
    "plot_confusion_matrix(expected, predicted)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Write files for reporting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "reporting_data = {\n",
    "    \"expected\": expected.tolist(),\n",
    "    \"predicted\": predicted.tolist(),\n",
    "}\n",
    "with open(file.reporting_data_report_cnn, 'w') as f:\n",
    "    json.dump(reporting_data, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "environment": {
   "kernel": "python3",
   "name": "tf2-gpu.2-6.m84",
   "type": "gcloud",
   "uri": "gcr.io/deeplearning-platform-release/tf2-gpu.2-6:m84"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}