train_ranking.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "6ekaJkcuHsiY"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import pandas as pd\n",
        "from typing import List\n",
        "\n",
        "import tensorflow as tf\n",
        "import tensorflow_ranking as tfr\n",
        "import tensorflow_recommenders as tfrs\n",
        "\n",
        "from src.model.tower import Tower\n",
        "from src.model.embedding import Embedding\n",
        "from src.model.ranking import ListwiseRanking\n",
        "from src.model.utils.utilities import (\n",
        "    sample_listwise,\n",
        "    train_test_split,\n",
        "    plot_history,\n",
        ")\n",
        "\n",
        "RANDOM_STATE = 42\n",
        "tf.random.set_seed(RANDOM_STATE)\n",
        "os.environ['TF_USE_LEGACY_KERAS'] = '1'"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Load Datasets"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "k-VF30hJn5-3"
      },
      "outputs": [],
      "source": [
        "# Dataset size (options: 100k, 1m)\n",
        "DATASET_SIZE: str = '100k'\n",
        "N_RATINGS: int    = 100_000\n",
        "\n",
        "# Load each dataset\n",
        "users_df    = pd.read_parquet(f'data/{DATASET_SIZE}-users.parquet')\n",
        "movies_df   = pd.read_parquet(f'data/{DATASET_SIZE}-movies.parquet')\n",
        "ratings_df  = pd.read_parquet(f'data/{DATASET_SIZE}-ratings.parquet')\n",
        "\n",
        "ratings_df = ratings_df.iloc[:N_RATINGS]\n",
        "\n",
        "# Remove lists\n",
        "movies_df.drop(columns=['movie_genres'], inplace=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user_id</th>\n",
              "      <th>movie_id</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>user_rating</th>\n",
              "      <th>user_gender</th>\n",
              "      <th>user_zip_code</th>\n",
              "      <th>user_bucketized_age</th>\n",
              "      <th>user_occupation_label</th>\n",
              "      <th>movie_title</th>\n",
              "      <th>movie_release_year</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>138</td>\n",
              "      <td>357</td>\n",
              "      <td>879024327</td>\n",
              "      <td>4.0</td>\n",
              "      <td>1</td>\n",
              "      <td>53211</td>\n",
              "      <td>45.0</td>\n",
              "      <td>4</td>\n",
              "      <td>One Flew Over the Cuckoo's Nest</td>\n",
              "      <td>1975</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>92</td>\n",
              "      <td>709</td>\n",
              "      <td>875654590</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1</td>\n",
              "      <td>80525</td>\n",
              "      <td>25.0</td>\n",
              "      <td>5</td>\n",
              "      <td>Strictly Ballroom</td>\n",
              "      <td>1992</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>301</td>\n",
              "      <td>412</td>\n",
              "      <td>882075110</td>\n",
              "      <td>4.0</td>\n",
              "      <td>1</td>\n",
              "      <td>55439</td>\n",
              "      <td>18.0</td>\n",
              "      <td>17</td>\n",
              "      <td>Very Brady Sequel, A</td>\n",
              "      <td>1996</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>60</td>\n",
              "      <td>56</td>\n",
              "      <td>883326919</td>\n",
              "      <td>4.0</td>\n",
              "      <td>1</td>\n",
              "      <td>06472</td>\n",
              "      <td>50.0</td>\n",
              "      <td>4</td>\n",
              "      <td>Pulp Fiction</td>\n",
              "      <td>1994</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>197</td>\n",
              "      <td>895</td>\n",
              "      <td>891409199</td>\n",
              "      <td>3.0</td>\n",
              "      <td>1</td>\n",
              "      <td>75094</td>\n",
              "      <td>50.0</td>\n",
              "      <td>18</td>\n",
              "      <td>Scream 2</td>\n",
              "      <td>1997</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "  user_id movie_id  timestamp  user_rating  user_gender user_zip_code  \\\n",
              "0     138      357  879024327          4.0            1         53211   \n",
              "1      92      709  875654590          2.0            1         80525   \n",
              "2     301      412  882075110          4.0            1         55439   \n",
              "3      60       56  883326919          4.0            1         06472   \n",
              "4     197      895  891409199          3.0            1         75094   \n",
              "\n",
              "   user_bucketized_age  user_occupation_label  \\\n",
              "0                 45.0                      4   \n",
              "1                 25.0                      5   \n",
              "2                 18.0                     17   \n",
              "3                 50.0                      4   \n",
              "4                 50.0                     18   \n",
              "\n",
              "                       movie_title movie_release_year  \n",
              "0  One Flew Over the Cuckoo's Nest               1975  \n",
              "1                Strictly Ballroom               1992  \n",
              "2             Very Brady Sequel, A               1996  \n",
              "3                     Pulp Fiction               1994  \n",
              "4                         Scream 2               1997  "
            ]
          },
          "execution_count": 21,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Handle missing values\n",
        "movies_df.fillna(value=-1, inplace=True)\n",
        "\n",
        "# Merge DataFrames\n",
        "ratings_df = ratings_df.merge(users_df, on = 'user_id')\n",
        "ratings_df = ratings_df.merge(movies_df, on = 'movie_id')\n",
        "\n",
        "ratings_df.head()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {},
      "outputs": [
        {
          "data": {
            "text/plain": [
              "{'user_id': b'138',\n",
              " 'movie_id': b'357',\n",
              " 'timestamp': 879024327,\n",
              " 'user_rating': 4.0,\n",
              " 'user_gender': 1,\n",
              " 'user_zip_code': b'53211',\n",
              " 'user_bucketized_age': 45.0,\n",
              " 'user_occupation_label': 4,\n",
              " 'movie_title': b\"One Flew Over the Cuckoo's Nest\",\n",
              " 'movie_release_year': b'1975'}"
            ]
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "TRAIN_RATIO: float = 0.8\n",
        "\n",
        "# Convert the DataFrame to TensorFlow Dataset\n",
        "ratings_dataset = tf.data.Dataset.from_tensor_slices(dict(ratings_df.astype(str)))\n",
        "movies_dataset  = tf.data.Dataset.from_tensor_slices(dict(movies_df.astype(str)))\n",
        "users_dataset  = tf.data.Dataset.from_tensor_slices(dict(users_df.astype(str)))\n",
        "\n",
        "ratings_dataset = ratings_dataset.map(lambda x: {\n",
        "    'user_id':                  x['user_id'],\n",
        "    'movie_id':                 x['movie_id'],\n",
        "    'timestamp':                int(x['timestamp']),\n",
        "    'user_rating':              float(x['user_rating']),\n",
        "    'user_gender':              int(x['user_gender']),\n",
        "    'user_zip_code':            x['user_zip_code'],\n",
        "    'user_bucketized_age':      float(x['user_bucketized_age']),\n",
        "    'user_occupation_label':    int(x['user_occupation_label']),\n",
        "    'movie_title':              x['movie_title'],\n",
        "    'movie_release_year':       x['movie_release_year'],\n",
        "})\n",
        "\n",
        "movies_dataset = movies_dataset.map(lambda x: {\n",
        "    'movie_id':                 x['movie_id'],\n",
        "    'movie_title':              x['movie_title'],\n",
        "    'movie_release_year':       x['movie_release_year'],\n",
        "})\n",
        "\n",
        "users_dataset = users_dataset.map(lambda x: {\n",
        "    'user_id':                  x['user_id'],\n",
        "    'user_gender':              int(x['user_gender']),\n",
        "    'user_zip_code':            x['user_zip_code'],\n",
        "    'user_bucketized_age':      float(x['user_bucketized_age']),\n",
        "    'user_occupation_label':    int(x['user_occupation_label']),\n",
        "})\n",
        "\n",
        "ratings_dataset.take(1).as_numpy_iterator().next()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training data:\t 8000\n",
            "Validation data: 2000\n"
          ]
        }
      ],
      "source": [
        "# Split into train and validation sets\n",
        "ratings_trainset, ratings_validset = train_test_split(\n",
        "    dataset = ratings_dataset,\n",
        "    train_size = TRAIN_RATIO,\n",
        "    random_state = RANDOM_STATE,\n",
        "    buffer_size = 10_000\n",
        ")\n",
        "\n",
        "print(f\"Training data:\\t\", ratings_trainset.__len__().numpy())\n",
        "print(f\"Validation data:\", ratings_validset.__len__().numpy())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {},
      "outputs": [],
      "source": [
        "ratings_trainset = sample_listwise(\n",
        "    ratings_trainset,\n",
        "    num_list_per_user = 50,\n",
        "    num_examples_per_list = 5,\n",
        "    seed = RANDOM_STATE\n",
        ")\n",
        "\n",
        "ratings_validset = sample_listwise(\n",
        "    ratings_validset,\n",
        "    num_list_per_user = 1,\n",
        "    num_examples_per_list = 5,\n",
        "    seed = RANDOM_STATE\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Create the Embedding Models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### User Embeddings"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {},
      "outputs": [],
      "source": [
        "# For either integer or string categorical features, the\n",
        "# `StringLookup` layer will create a vocabulary that maps\n",
        "# each string value to an integer index followed by an\n",
        "# embedding layer:\n",
        "USER_STR_FEATURES: List[str] = ['user_id', 'user_zip_code']\n",
        "USER_INT_FEATURES: List[str] = ['user_gender', 'user_bucketized_age', 'user_occupation_label']\n",
        "\n",
        "# For text features, the `TextVectorization` layer will\n",
        "# create a vocabulary that maps each token to an integer\n",
        "# index followed by an embedding layer:\n",
        "USER_TEXT_FEATURES: List[str] = []\n",
        "\n",
        "# Timestamp features will be discretized into buckets\n",
        "# and the `Discretization` layer will create a vocabulary\n",
        "# for the embedding layer. The value will finally be\n",
        "# normalized between 0 and 1:\n",
        "USER_TIMESTAMP_FEATURES: List[str] = []\n",
        "\n",
        "# Embedding layers will be applied to each of the specified \n",
        "# features, with all layers having the same dimensionality \n",
        "# as defined below:\n",
        "USER_EMBEDDING_DIM: int = 64"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {},
      "outputs": [],
      "source": [
        "user_embedding_model = Embedding(\n",
        "    dataset             = users_dataset.batch(1_000),\n",
        "    str_features        = USER_STR_FEATURES,\n",
        "    int_features        = USER_INT_FEATURES,\n",
        "    text_features       = USER_TEXT_FEATURES,\n",
        "    timestamp_features  = USER_TIMESTAMP_FEATURES,\n",
        "    embedding_dim       = USER_EMBEDDING_DIM,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Movie Embeddings"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {},
      "outputs": [],
      "source": [
        "# For either integer or string categorical features, the\n",
        "# `StringLookup` layer will create a vocabulary that maps\n",
        "# each string value to an integer index followed by an\n",
        "# embedding layer:\n",
        "MOVIE_STR_FEATURES: List[str] = ['movie_release_year', 'movie_title']\n",
        "MOVIE_INT_FEATURES: List[str] = []\n",
        "\n",
        "# For text features, the `TextVectorization` layer will\n",
        "# create a vocabulary that maps each token to an integer\n",
        "# index followed by an embedding layer:\n",
        "MOVIE_TEXT_FEATURES: List[str] = []\n",
        "\n",
        "# Timestamp features will be discretized into buckets\n",
        "# and the `Discretization` layer will create a vocabulary\n",
        "# for the embedding layer. The value will finally be\n",
        "# normalized between 0 and 1:\n",
        "MOVIE_TIMESTAMP_FEATURES: List[str] = []\n",
        "\n",
        "# Embedding layers will be applied to each of the specified \n",
        "# features, with all layers having the same dimensionality \n",
        "# as defined below:\n",
        "MOVIE_EMBEDDING_DIM: int  = 32"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {},
      "outputs": [],
      "source": [
        "movie_embedding_model = Embedding(\n",
        "    dataset             = movies_dataset.batch(1_000),\n",
        "    str_features        = MOVIE_STR_FEATURES,\n",
        "    int_features        = MOVIE_INT_FEATURES,\n",
        "    text_features       = MOVIE_TEXT_FEATURES,\n",
        "    timestamp_features  = MOVIE_TIMESTAMP_FEATURES,\n",
        "    embedding_dim       = MOVIE_EMBEDDING_DIM,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Create the Towers Models"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Query Tower"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {},
      "outputs": [],
      "source": [
        "# The projection dimension determines the size of the feature \n",
        "# space for learning cross-features. By employing low-rank \n",
        "# techniques to approximate the weight matrices of the DCN, \n",
        "# we can effectively reduce both training and serving costs.\n",
        "# More: https://www.tensorflow.org/recommenders/examples/dcn\n",
        "QT_CROSS_PROJECTION_DIM: int = user_embedding_model.embeddings_output_dim / 4\n",
        "\n",
        "# Larger and more complex models, though often delivering better \n",
        "# performance, generally require careful tuning. Nevertheless, it's \n",
        "# essential to include at least one dense layer in the tower to unify \n",
        "# the outputs from the embedding layers. This is necessary because \n",
        "# each tower may generate embeddings of varying sizes, depending on \n",
        "# the different features it processes:\n",
        "# More: https://www.tensorflow.org/recommenders/examples/deep_recommenders\n",
        "QT_DENSE_LAYERS: List[int] = [32]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Deep & Cross Network\n",
        "qt_cross_layer = tfrs.layers.dcn.Cross(\n",
        "    projection_dim     = QT_CROSS_PROJECTION_DIM,\n",
        "    kernel_initializer = \"glorot_uniform\"\n",
        ")\n",
        "\n",
        "query_tower = Tower(\n",
        "    embedding_model = user_embedding_model,\n",
        "    cross_layer     = qt_cross_layer,\n",
        "    dense_layers    = QT_DENSE_LAYERS\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "### Candidate Tower"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {},
      "outputs": [],
      "source": [
        "# The projection dimension determines the size of the feature \n",
        "# space for learning cross-features. By employing low-rank \n",
        "# techniques to approximate the weight matrices of the DCN, \n",
        "# we can effectively reduce both training and serving costs.\n",
        "# More: https://www.tensorflow.org/recommenders/examples/dcn\n",
        "CT_CROSS_PROJECTION_DIM: int = movie_embedding_model.embeddings_output_dim / 2\n",
        "\n",
        "# Larger and more complex models, though often delivering better \n",
        "# performance, generally require careful tuning. Nevertheless, it's \n",
        "# essential to include at least one dense layer in the tower to unify \n",
        "# the outputs from the embedding layers. This is necessary because \n",
        "# each tower may generate embeddings of varying sizes, depending on \n",
        "# the different features it processes:\n",
        "# More: https://www.tensorflow.org/recommenders/examples/deep_recommenders\n",
        "CT_DENSE_LAYERS: List[int] = [32]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Deep & Cross Network\n",
        "ct_cross_layer = tfrs.layers.dcn.Cross(\n",
        "    projection_dim     = CT_CROSS_PROJECTION_DIM,\n",
        "    kernel_initializer = \"glorot_uniform\"\n",
        ")\n",
        "\n",
        "candidate_tower = Tower(\n",
        "    embedding_model = movie_embedding_model,\n",
        "    cross_layer     = ct_cross_layer,\n",
        "    dense_layers    = CT_DENSE_LAYERS\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Define the Ranking Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {},
      "outputs": [],
      "source": [
        "ranking_task = tfrs.tasks.Ranking(\n",
        "    loss = tf.keras.losses.MeanSquaredError(),      # MSE\n",
        "    # loss = tfr.keras.losses.PairwiseHingeLoss(),\n",
        "    # loss = tfr.keras.losses.ListMLELoss(),\n",
        "    metrics = [\n",
        "        tf.keras.metrics.RootMeanSquaredError(),    # RMSE\n",
        "        tfr.keras.metrics.NDCGMetric(),             # NGCG\n",
        "    ],\n",
        ")\n",
        "\n",
        "# We will use a ranking model to help train the \n",
        "# retrieval system. We will use the pointwise approach\n",
        "# that assesses the relevance of individual items \n",
        "# or documents independently.\n",
        "ranking_model = ListwiseRanking(\n",
        "    query_tower     = query_tower,\n",
        "    candidate_tower = candidate_tower,\n",
        "    task            = ranking_task\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "## Recommender"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 39,
      "metadata": {},
      "outputs": [],
      "source": [
        "RUN_EAGERLY: bool       = True     # Better debugging, slower training\n",
        "LEARNING_RATE: float    = 0.1\n",
        "\n",
        "model = ranking_model\n",
        "\n",
        "optimizer = tf.keras.optimizers.Adagrad(learning_rate=LEARNING_RATE)\n",
        "model.compile(optimizer = optimizer, run_eagerly=RUN_EAGERLY)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 40,
      "metadata": {},
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "26/26 [==============================] - 3s 115ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 7.4846 - regularization_loss: 0.0000e+00 - total_loss: 7.4846 - val_root_mean_squared_error: 1.0589 - val_ndcg_metric_1: 0.8749 - val_loss: 1.1213 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1213\n",
            "Epoch 2/50\n",
            "26/26 [==============================] - 3s 107ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.8437 - regularization_loss: 0.0000e+00 - total_loss: 0.8437 - val_root_mean_squared_error: 1.0843 - val_ndcg_metric_1: 0.8927 - val_loss: 1.1757 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1757\n",
            "Epoch 3/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.7729 - regularization_loss: 0.0000e+00 - total_loss: 0.7729 - val_root_mean_squared_error: 1.0327 - val_ndcg_metric_1: 0.8934 - val_loss: 1.0664 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.0664\n",
            "Epoch 4/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6931 - regularization_loss: 0.0000e+00 - total_loss: 0.6931 - val_root_mean_squared_error: 1.0387 - val_ndcg_metric_1: 0.8933 - val_loss: 1.0789 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.0789\n",
            "Epoch 5/50\n",
            "26/26 [==============================] - 3s 102ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6924 - regularization_loss: 0.0000e+00 - total_loss: 0.6924 - val_root_mean_squared_error: 1.0502 - val_ndcg_metric_1: 0.8918 - val_loss: 1.1029 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1029\n",
            "Epoch 6/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6842 - regularization_loss: 0.0000e+00 - total_loss: 0.6842 - val_root_mean_squared_error: 1.0421 - val_ndcg_metric_1: 0.8902 - val_loss: 1.0860 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.0860\n",
            "Epoch 7/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6641 - regularization_loss: 0.0000e+00 - total_loss: 0.6641 - val_root_mean_squared_error: 1.0435 - val_ndcg_metric_1: 0.8903 - val_loss: 1.0888 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.0888\n",
            "Epoch 8/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6592 - regularization_loss: 0.0000e+00 - total_loss: 0.6592 - val_root_mean_squared_error: 1.0466 - val_ndcg_metric_1: 0.8895 - val_loss: 1.0953 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.0953\n",
            "Epoch 9/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6569 - regularization_loss: 0.0000e+00 - total_loss: 0.6569 - val_root_mean_squared_error: 1.0489 - val_ndcg_metric_1: 0.8891 - val_loss: 1.1002 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1002\n",
            "Epoch 10/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6521 - regularization_loss: 0.0000e+00 - total_loss: 0.6521 - val_root_mean_squared_error: 1.0499 - val_ndcg_metric_1: 0.8879 - val_loss: 1.1023 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1023\n",
            "Epoch 11/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6456 - regularization_loss: 0.0000e+00 - total_loss: 0.6456 - val_root_mean_squared_error: 1.0511 - val_ndcg_metric_1: 0.8906 - val_loss: 1.1049 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1049\n",
            "Epoch 12/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6396 - regularization_loss: 0.0000e+00 - total_loss: 0.6396 - val_root_mean_squared_error: 1.0521 - val_ndcg_metric_1: 0.8888 - val_loss: 1.1068 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1068\n",
            "Epoch 13/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6324 - regularization_loss: 0.0000e+00 - total_loss: 0.6324 - val_root_mean_squared_error: 1.0518 - val_ndcg_metric_1: 0.8880 - val_loss: 1.1063 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1063\n",
            "Epoch 14/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6234 - regularization_loss: 0.0000e+00 - total_loss: 0.6234 - val_root_mean_squared_error: 1.0525 - val_ndcg_metric_1: 0.8886 - val_loss: 1.1077 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1077\n",
            "Epoch 15/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6136 - regularization_loss: 0.0000e+00 - total_loss: 0.6136 - val_root_mean_squared_error: 1.0541 - val_ndcg_metric_1: 0.8882 - val_loss: 1.1112 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1112\n",
            "Epoch 16/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.6019 - regularization_loss: 0.0000e+00 - total_loss: 0.6019 - val_root_mean_squared_error: 1.0571 - val_ndcg_metric_1: 0.8880 - val_loss: 1.1175 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1175\n",
            "Epoch 17/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5881 - regularization_loss: 0.0000e+00 - total_loss: 0.5881 - val_root_mean_squared_error: 1.0604 - val_ndcg_metric_1: 0.8890 - val_loss: 1.1245 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1245\n",
            "Epoch 18/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5692 - regularization_loss: 0.0000e+00 - total_loss: 0.5692 - val_root_mean_squared_error: 1.0636 - val_ndcg_metric_1: 0.8895 - val_loss: 1.1313 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1313\n",
            "Epoch 19/50\n",
            "26/26 [==============================] - 3s 103ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5516 - regularization_loss: 0.0000e+00 - total_loss: 0.5516 - val_root_mean_squared_error: 1.0674 - val_ndcg_metric_1: 0.8893 - val_loss: 1.1393 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1393\n",
            "Epoch 20/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5359 - regularization_loss: 0.0000e+00 - total_loss: 0.5359 - val_root_mean_squared_error: 1.0715 - val_ndcg_metric_1: 0.8916 - val_loss: 1.1480 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1480\n",
            "Epoch 21/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5204 - regularization_loss: 0.0000e+00 - total_loss: 0.5204 - val_root_mean_squared_error: 1.0750 - val_ndcg_metric_1: 0.8926 - val_loss: 1.1556 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1556\n",
            "Epoch 22/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.5074 - regularization_loss: 0.0000e+00 - total_loss: 0.5074 - val_root_mean_squared_error: 1.0773 - val_ndcg_metric_1: 0.8932 - val_loss: 1.1605 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1605\n",
            "Epoch 23/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4934 - regularization_loss: 0.0000e+00 - total_loss: 0.4934 - val_root_mean_squared_error: 1.0805 - val_ndcg_metric_1: 0.8931 - val_loss: 1.1674 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1674\n",
            "Epoch 24/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4806 - regularization_loss: 0.0000e+00 - total_loss: 0.4806 - val_root_mean_squared_error: 1.0840 - val_ndcg_metric_1: 0.8919 - val_loss: 1.1750 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1750\n",
            "Epoch 25/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4717 - regularization_loss: 0.0000e+00 - total_loss: 0.4717 - val_root_mean_squared_error: 1.0862 - val_ndcg_metric_1: 0.8921 - val_loss: 1.1798 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1798\n",
            "Epoch 26/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4592 - regularization_loss: 0.0000e+00 - total_loss: 0.4592 - val_root_mean_squared_error: 1.0887 - val_ndcg_metric_1: 0.8924 - val_loss: 1.1853 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1853\n",
            "Epoch 27/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4525 - regularization_loss: 0.0000e+00 - total_loss: 0.4525 - val_root_mean_squared_error: 1.0907 - val_ndcg_metric_1: 0.8964 - val_loss: 1.1896 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1896\n",
            "Epoch 28/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4397 - regularization_loss: 0.0000e+00 - total_loss: 0.4397 - val_root_mean_squared_error: 1.0933 - val_ndcg_metric_1: 0.8956 - val_loss: 1.1952 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.1952\n",
            "Epoch 29/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4327 - regularization_loss: 0.0000e+00 - total_loss: 0.4327 - val_root_mean_squared_error: 1.0958 - val_ndcg_metric_1: 0.8970 - val_loss: 1.2008 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2008\n",
            "Epoch 30/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4195 - regularization_loss: 0.0000e+00 - total_loss: 0.4195 - val_root_mean_squared_error: 1.0992 - val_ndcg_metric_1: 0.8951 - val_loss: 1.2083 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2083\n",
            "Epoch 31/50\n",
            "26/26 [==============================] - 3s 103ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.4118 - regularization_loss: 0.0000e+00 - total_loss: 0.4118 - val_root_mean_squared_error: 1.1025 - val_ndcg_metric_1: 0.8929 - val_loss: 1.2155 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2155\n",
            "Epoch 32/50\n",
            "26/26 [==============================] - 3s 104ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3994 - regularization_loss: 0.0000e+00 - total_loss: 0.3994 - val_root_mean_squared_error: 1.1071 - val_ndcg_metric_1: 0.8910 - val_loss: 1.2257 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2257\n",
            "Epoch 33/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3912 - regularization_loss: 0.0000e+00 - total_loss: 0.3912 - val_root_mean_squared_error: 1.1123 - val_ndcg_metric_1: 0.8901 - val_loss: 1.2373 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2373\n",
            "Epoch 34/50\n",
            "26/26 [==============================] - 3s 105ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3805 - regularization_loss: 0.0000e+00 - total_loss: 0.3805 - val_root_mean_squared_error: 1.1166 - val_ndcg_metric_1: 0.8895 - val_loss: 1.2469 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2469\n",
            "Epoch 35/50\n",
            "26/26 [==============================] - 3s 110ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3706 - regularization_loss: 0.0000e+00 - total_loss: 0.3706 - val_root_mean_squared_error: 1.1209 - val_ndcg_metric_1: 0.8915 - val_loss: 1.2564 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2564\n",
            "Epoch 36/50\n",
            "26/26 [==============================] - 3s 110ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3613 - regularization_loss: 0.0000e+00 - total_loss: 0.3613 - val_root_mean_squared_error: 1.1260 - val_ndcg_metric_1: 0.8922 - val_loss: 1.2680 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2680\n",
            "Epoch 37/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3523 - regularization_loss: 0.0000e+00 - total_loss: 0.3523 - val_root_mean_squared_error: 1.1301 - val_ndcg_metric_1: 0.8930 - val_loss: 1.2772 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2772\n",
            "Epoch 38/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3432 - regularization_loss: 0.0000e+00 - total_loss: 0.3432 - val_root_mean_squared_error: 1.1330 - val_ndcg_metric_1: 0.8915 - val_loss: 1.2838 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2838\n",
            "Epoch 39/50\n",
            "26/26 [==============================] - 3s 106ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3347 - regularization_loss: 0.0000e+00 - total_loss: 0.3347 - val_root_mean_squared_error: 1.1365 - val_ndcg_metric_1: 0.8897 - val_loss: 1.2915 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2915\n",
            "Epoch 40/50\n",
            "26/26 [==============================] - 3s 115ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3283 - regularization_loss: 0.0000e+00 - total_loss: 0.3283 - val_root_mean_squared_error: 1.1401 - val_ndcg_metric_1: 0.8885 - val_loss: 1.2997 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.2997\n",
            "Epoch 41/50\n",
            "26/26 [==============================] - 3s 116ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3196 - regularization_loss: 0.0000e+00 - total_loss: 0.3196 - val_root_mean_squared_error: 1.1435 - val_ndcg_metric_1: 0.8885 - val_loss: 1.3076 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3076\n",
            "Epoch 42/50\n",
            "26/26 [==============================] - 3s 112ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3145 - regularization_loss: 0.0000e+00 - total_loss: 0.3145 - val_root_mean_squared_error: 1.1480 - val_ndcg_metric_1: 0.8898 - val_loss: 1.3180 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3180\n",
            "Epoch 43/50\n",
            "26/26 [==============================] - 3s 112ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3062 - regularization_loss: 0.0000e+00 - total_loss: 0.3062 - val_root_mean_squared_error: 1.1498 - val_ndcg_metric_1: 0.8906 - val_loss: 1.3221 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3221\n",
            "Epoch 44/50\n",
            "26/26 [==============================] - 3s 110ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3040 - regularization_loss: 0.0000e+00 - total_loss: 0.3040 - val_root_mean_squared_error: 1.1548 - val_ndcg_metric_1: 0.8904 - val_loss: 1.3336 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3336\n",
            "Epoch 45/50\n",
            "26/26 [==============================] - 3s 109ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.2905 - regularization_loss: 0.0000e+00 - total_loss: 0.2905 - val_root_mean_squared_error: 1.1564 - val_ndcg_metric_1: 0.8875 - val_loss: 1.3373 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3373\n",
            "Epoch 46/50\n",
            "26/26 [==============================] - 3s 124ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3012 - regularization_loss: 0.0000e+00 - total_loss: 0.3012 - val_root_mean_squared_error: 1.1628 - val_ndcg_metric_1: 0.8863 - val_loss: 1.3520 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3520\n",
            "Epoch 47/50\n",
            "26/26 [==============================] - 3s 120ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.2760 - regularization_loss: 0.0000e+00 - total_loss: 0.2760 - val_root_mean_squared_error: 1.1694 - val_ndcg_metric_1: 0.8863 - val_loss: 1.3674 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3674\n",
            "Epoch 48/50\n",
            "26/26 [==============================] - 3s 122ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.3047 - regularization_loss: 0.0000e+00 - total_loss: 0.3047 - val_root_mean_squared_error: 1.1546 - val_ndcg_metric_1: 0.8857 - val_loss: 1.3331 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3331\n",
            "Epoch 49/50\n",
            "26/26 [==============================] - 3s 113ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.2721 - regularization_loss: 0.0000e+00 - total_loss: 0.2721 - val_root_mean_squared_error: 1.1727 - val_ndcg_metric_1: 0.8883 - val_loss: 1.3752 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3752\n",
            "Epoch 50/50\n",
            "26/26 [==============================] - 3s 110ms/step - root_mean_squared_error: 0.0000e+00 - ndcg_metric_1: 0.0000e+00 - loss: 0.2752 - regularization_loss: 0.0000e+00 - total_loss: 0.2752 - val_root_mean_squared_error: 1.1738 - val_ndcg_metric_1: 0.8877 - val_loss: 1.3779 - val_regularization_loss: 0.0000e+00 - val_total_loss: 1.3779\n",
            "{'root_mean_squared_error': 1.1738265752792358, 'ndcg_metric_1': 0.8876756429672241, 'loss': 1.3778690099716187, 'regularization_loss': 0, 'total_loss': 1.3778690099716187}\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x700 with 6 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "EPOCHS: int          = 50\n",
        "VALIDATION_FREQ: int = 1\n",
        "\n",
        "cached_ratings_trainset = ratings_trainset.shuffle(100_000).batch(1024).cache()\n",
        "cached_ratings_validset = ratings_validset.batch(1024).cache()\n",
        "\n",
        "early_stopping_callback = tf.keras.callbacks.EarlyStopping(\n",
        "    monitor = 'total_loss',\n",
        "    patience = 10,\n",
        ")\n",
        "\n",
        "history = model.fit(\n",
        "    cached_ratings_trainset,\n",
        "    validation_data = cached_ratings_validset,\n",
        "    validation_freq = VALIDATION_FREQ,\n",
        "    epochs          = EPOCHS,\n",
        "    callbacks = [\n",
        "        early_stopping_callback\n",
        "    ]\n",
        ")\n",
        "\n",
        "print(model.evaluate(cached_ratings_validset, return_dict=True, verbose=False))\n",
        "plot_history(history = history, plot_training=False)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "listwise_ranking.ipynb",
      "private_outputs": true,
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.10.12"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}