{ "nbformat": 4, "nbformat_minor": 0, "metadata": { "colab": { "provenance": [], "gpuType": "T4" }, "kernelspec": { "name": "python3", "display_name": "Python 3" }, "language_info": { "name": "python" }, "accelerator": "GPU" }, "cells": [ { "cell_type": "code", "execution_count": null, "metadata": { "id": "xqTOxz_hO69H" }, "outputs": [], "source": [ "import tensorflow as tf\n", "import tensorflow_datasets as tfds\n", "import numpy as np\n", "import matplotlib.pyplot as plt" ] }, { "cell_type": "code", "source": [ "dataset = tfds.load('imdb_reviews', as_supervised=True)\n", "train_dataset, test_dataset = dataset['train'], dataset['test']\n", "\n", "batch_size = 32\n", "\n", "train_dataset = train_dataset.shuffle(10000).batch(batch_size)\n", "test_dataset = test_dataset.batch(batch_size)" ], "metadata": { "id": "mrt_kDGBO_1_" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "example, label = next(iter(train_dataset))\n", "print('Text:\\n', example.numpy()[0])\n", "print('\\nLabel: ', label.numpy()[0])" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "pykFQMilPBVK", "outputId": "f5bfe4b0-4b26-4f5a-a953-87ac8ef7efa3" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Text:\n", " b'Very funny to watch \"Beretta\\'s Island\" as kind of natural trash-film.It is like answer to Jess Franko\\'s type of b-movie.Bodybuilders strikes back (!face to face!) to pushers.The very very very stupid strike!Action: unbelievably bad directed firing(shooting) scenes look even better than hand-to-hand fighting.Chasing scenes ridiculous.Saving beauties scenes incredibly stupid.Erotic scenes are very unerotic.The main luck of film is pretty landscapes and festival scenes.Don\\'t miss:Arnold Schwarzenegger\\'s joke at start of film and list of Franco Columbu\\'s kin at the end. Special attraction: naked bosom.Almoust forgot - Franco can sing!'\n", "\n", "Label: 0\n" ] } ] }, { "cell_type": "code", "source": [ "vectorize_layer = tf.keras.layers.TextVectorization(output_mode='int', output_sequence_length=100)\n", "\n", "vectorize_layer.adapt(train_dataset.map(lambda x, y: x))" ], "metadata": { "id": "ZLSsiXquPCt_" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "model = tf.keras.Sequential([\n", " vectorize_layer,\n", " tf.keras.layers.Embedding(len(vectorize_layer.get_vocabulary()), 32, mask_zero=True),\n", " tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32, return_sequences=True)),\n", " tf.keras.layers.Dropout(0.4),\n", " tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(16)),\n", " tf.keras.layers.Dropout(0.4),\n", " tf.keras.layers.Dense(1)\n", "])\n", "\n", "model.build(input_shape=(None,))\n", "\n", "model.summary()\n", "\n", "model.compile(\n", " loss=tf.keras.losses.BinaryCrossentropy(from_logits=True),\n", " optimizer=tf.keras.optimizers.Adam(),\n", " metrics=['accuracy']\n", ")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 425 }, "id": "WZ7IMLD2PEv-", "outputId": "4617a678-35c8-4174-a133-bcce02c81245" }, "execution_count": null, "outputs": [ { "output_type": "display_data", "data": { "text/plain": [ "\u001b[1mModel: \"sequential_14\"\u001b[0m\n" ], "text/html": [ "
Model: \"sequential_14\"\n",
              "
\n" ] }, "metadata": {} }, { "output_type": "display_data", "data": { "text/plain": [ "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n", "┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n", "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n", "│ text_vectorization_2 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m100\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n", "│ (\u001b[38;5;33mTextVectorization\u001b[0m) │ │ │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ embedding_14 (\u001b[38;5;33mEmbedding\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m100\u001b[0m, \u001b[38;5;34m32\u001b[0m) │ \u001b[38;5;34m3,900,608\u001b[0m │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ bidirectional_28 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m100\u001b[0m, \u001b[38;5;34m64\u001b[0m) │ \u001b[38;5;34m16,640\u001b[0m │\n", "│ (\u001b[38;5;33mBidirectional\u001b[0m) │ │ │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ dropout_18 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m100\u001b[0m, \u001b[38;5;34m64\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ bidirectional_29 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m) │ \u001b[38;5;34m10,368\u001b[0m │\n", "│ (\u001b[38;5;33mBidirectional\u001b[0m) │ │ │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ dropout_19 (\u001b[38;5;33mDropout\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m) │ \u001b[38;5;34m0\u001b[0m │\n", "├─────────────────────────────────┼────────────────────────┼───────────────┤\n", "│ dense_28 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m) │ \u001b[38;5;34m33\u001b[0m │\n", "└─────────────────────────────────┴────────────────────────┴───────────────┘\n" ], "text/html": [ "
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
              "┃ Layer (type)                    ┃ Output Shape           ┃       Param # ┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
              "│ text_vectorization_2            │ (None, 100)            │             0 │\n",
              "│ (TextVectorization)             │                        │               │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ embedding_14 (Embedding)        │ (None, 100, 32)        │     3,900,608 │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ bidirectional_28                │ (None, 100, 64)        │        16,640 │\n",
              "│ (Bidirectional)                 │                        │               │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ dropout_18 (Dropout)            │ (None, 100, 64)        │             0 │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ bidirectional_29                │ (None, 32)             │        10,368 │\n",
              "│ (Bidirectional)                 │                        │               │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ dropout_19 (Dropout)            │ (None, 32)             │             0 │\n",
              "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
              "│ dense_28 (Dense)                │ (None, 1)              │            33 │\n",
              "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
              "
\n" ] }, "metadata": {} }, { "output_type": "display_data", "data": { "text/plain": [ "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m3,927,649\u001b[0m (14.98 MB)\n" ], "text/html": [ "
 Total params: 3,927,649 (14.98 MB)\n",
              "
\n" ] }, "metadata": {} }, { "output_type": "display_data", "data": { "text/plain": [ "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m3,927,649\u001b[0m (14.98 MB)\n" ], "text/html": [ "
 Trainable params: 3,927,649 (14.98 MB)\n",
              "
\n" ] }, "metadata": {} }, { "output_type": "display_data", "data": { "text/plain": [ "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n" ], "text/html": [ "
 Non-trainable params: 0 (0.00 B)\n",
              "
\n" ] }, "metadata": {} } ] }, { "cell_type": "code", "source": [ "history = model.fit(\n", " train_dataset,\n", " epochs=3,\n", " validation_data=test_dataset,\n", ")" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "2Rk-AfPlPIiZ", "outputId": "10eb5c4c-f01e-4169-d8f2-7995812d8b0f" }, "execution_count": null, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Epoch 1/3\n", "\u001b[1m782/782\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m30s\u001b[0m 34ms/step - accuracy: 0.6638 - loss: 0.5573 - val_accuracy: 0.7794 - val_loss: 0.4273\n", "Epoch 2/3\n", "\u001b[1m782/782\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m28s\u001b[0m 36ms/step - accuracy: 0.9093 - loss: 0.2319 - val_accuracy: 0.7772 - val_loss: 0.4602\n", "Epoch 3/3\n", "\u001b[1m782/782\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m26s\u001b[0m 33ms/step - accuracy: 0.9645 - loss: 0.1010 - val_accuracy: 0.7841 - val_loss: 0.6840\n" ] } ] }, { "cell_type": "code", "source": [], "metadata": { "id": "cw3MBi32POgy" }, "execution_count": null, "outputs": [] } ] }