35-RecommendationSystem-tf-data.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "import pandas as pd\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "dataset = pd.read_csv('https://github.com/Alireza-Akhavan/datasets_and_models/raw/main/ratings.csv')\n",
        "train, validation = train_test_split(dataset, test_size=0.2, random_state=42)\n",
        "\n",
        "n_users = len(dataset.user_id.unique())\n",
        "n_books = len(dataset.book_id.unique())\n",
        "\n",
        "\n",
        "def create_dataset(df):\n",
        "    user_inputs = tf.data.Dataset.from_tensor_slices(df['user_id'])\n",
        "    book_inputs = tf.data.Dataset.from_tensor_slices(df['book_id'])\n",
        "    ratings = tf.data.Dataset.from_tensor_slices(df['rating'])\n",
        "    dataset = tf.data.Dataset.zip(((user_inputs, book_inputs), ratings))\n",
        "    dataset = dataset.shuffle(len(df))\n",
        "    dataset = dataset.batch(32)\n",
        "    dataset = dataset.prefetch(tf.data.experimental.AUTOTUNE)\n",
        "    return dataset\n",
        "\n",
        "\n",
        "train_dataset = create_dataset(train)\n",
        "validation_dataset = create_dataset(validation)"
      ],
      "metadata": {
        "id": "jU7FWGKjOptV"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "book_input = tf.keras.layers.Input(shape=[1], name=\"Book-Input\")\n",
        "book_embedding = tf.keras.layers.Embedding(n_books+1, 5, name=\"Book-Embedding\")(book_input)\n",
        "book_vec = tf.keras.layers.Flatten(name=\"Flatten-Books\")(book_embedding)\n",
        "\n",
        "user_input = tf.keras.layers.Input(shape=[1], name=\"User-Input\")\n",
        "user_embedding = tf.keras.layers.Embedding(n_users+1, 5, name=\"User-Embedding\")(user_input)\n",
        "user_vec = tf.keras.layers.Flatten(name=\"Flatten-Users\")(user_embedding)\n",
        "\n",
        "prod = tf.keras.layers.Dot(name=\"Dot-Product\", axes=1)([book_vec, user_vec])\n",
        "model = tf.keras.models.Model([user_input, book_input], prod)\n",
        "model.compile('adam', 'mean_squared_error')"
      ],
      "metadata": {
        "id": "r93R2EglOubH"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "keras.utils.plot_model(model, show_shapes=True)"
      ],
      "metadata": {
        "id": "xCHA-kDXrUSV",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 422
        },
        "outputId": "4f10e051-164e-4ad4-9c48-90d534604e6a"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.callbacks import ReduceLROnPlateau\n",
        "\n",
        "# Define the callback\n",
        "reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=5, min_lr=0.0001)\n",
        "\n",
        "# Add the callback to the fit function\n",
        "history = model.fit(train_dataset, epochs=10,\n",
        "                    callbacks=[reduce_lr],\n",
        "                    validation_data=validation_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6LwdA-egOwVv",
        "outputId": "9fa244c7-0871-4855-c83f-51b10335f028"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "24544/24544 [==============================] - 122s 5ms/step - loss: 13.9469 - val_loss: 8.9091 - lr: 0.0010\n",
            "Epoch 2/10\n",
            "24544/24544 [==============================] - 105s 4ms/step - loss: 4.6770 - val_loss: 2.6496 - lr: 0.0010\n",
            "Epoch 3/10\n",
            "24544/24544 [==============================] - 115s 5ms/step - loss: 1.8080 - val_loss: 1.6492 - lr: 0.0010\n",
            "Epoch 4/10\n",
            "24544/24544 [==============================] - 104s 4ms/step - loss: 1.1996 - val_loss: 1.3241 - lr: 0.0010\n",
            "Epoch 5/10\n",
            "24544/24544 [==============================] - 113s 4ms/step - loss: 0.9620 - val_loss: 1.1684 - lr: 0.0010\n",
            "Epoch 6/10\n",
            "24544/24544 [==============================] - 108s 4ms/step - loss: 0.8424 - val_loss: 1.0777 - lr: 0.0010\n",
            "Epoch 7/10\n",
            "24544/24544 [==============================] - 105s 4ms/step - loss: 0.7727 - val_loss: 1.0228 - lr: 0.0010\n",
            "Epoch 8/10\n",
            "24544/24544 [==============================] - 107s 4ms/step - loss: 0.7299 - val_loss: 0.9842 - lr: 0.0010\n",
            "Epoch 9/10\n",
            "24544/24544 [==============================] - 104s 4ms/step - loss: 0.7016 - val_loss: 0.9580 - lr: 0.0010\n",
            "Epoch 10/10\n",
            "24544/24544 [==============================] - 105s 4ms/step - loss: 0.6814 - val_loss: 0.9403 - lr: 0.0010\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "bimport matplotlib.pyplot as plt\n",
        "\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.legend([\"train\", \"val\"])"
      ],
      "metadata": {
        "id": "CeADjhqDOxM-",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        },
        "outputId": "6e00d233-706b-4d2f-e78c-eb19d1a9c003"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7d1e29510610>"
            ]
          },
          "metadata": {},
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.fit(train_dataset, epochs=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ijpo9Jzq0mL1",
        "outputId": "717186c0-9461-4a5c-de6b-149884290bee"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "24544/24544 [==============================] - 100s 4ms/step - loss: 0.6671\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7d1d5c746bc0>"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    }
  ],
  "metadata": {
    "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": "3.6.6"
    },
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "accelerator": "GPU"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}