GAN_event_generator.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "GAN_event_generator.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/physicistharish/GAN_Event_Generator/blob/main/GAN_event_generator.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6mAFVt4RgNcU"
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import os\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "import glob\n",
        "import imageio\n",
        "import matplotlib.pyplot as plt\n",
        "import PIL\n",
        "from tensorflow.keras import layers\n",
        "import time\n",
        "import io\n",
        "import pickle\n",
        "from IPython import display\n",
        "from google.colab import files"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ht_zCXy4Dhnm"
      },
      "source": [
        "from keras import backend\n",
        "\n",
        "from numpy import cov\n",
        "from numpy import trace\n",
        "from numpy import iscomplexobj\n",
        "from numpy.random import random\n",
        "from scipy.linalg import sqrtm"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P9qodlS4ALYK",
        "outputId": "9cd21e7e-63d9-4d75-9810-646b327f82be"
      },
      "source": [
        "from google.colab import drive\n",
        "\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "path = \"/content/drive/MyDrive/Project_Files/Csv/16349 events XY.csv\"# Change the path directory as per your need\n",
        "\n",
        "df1 = pd.read_csv(path)\n",
        "print(df1)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "       Unnamed: 0         0         1  ...        10        11        12\n",
            "0               0  0.388266  1.066190  ...  -12.1232  -36.5231  -76.7150\n",
            "1               0 -0.450752 -0.833231  ...   24.5506   32.9202   40.7282\n",
            "2               0  1.080110  2.199080  ...   42.8811   61.5420   89.7588\n",
            "3               0 -2.602280 -7.072710  ...  -40.4210  -47.4739  -66.2680\n",
            "4               0  0.085928  0.266413  ...  108.9250  105.9890   95.7408\n",
            "...           ...       ...       ...  ...       ...       ...       ...\n",
            "16344           0  0.424090  0.911776  ...   17.8851   16.7183   18.3112\n",
            "16345           0  0.210162  2.838150  ...   20.9781   24.7027   34.6982\n",
            "16346           0  0.570230  1.158780  ...   69.3984   86.1083  108.1730\n",
            "16347           0  0.645774  2.287450  ...   17.2922   14.9753  -19.0565\n",
            "16348           0  0.895050  3.375180  ...   19.4355   31.8020   33.9234\n",
            "\n",
            "[16349 rows x 14 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GsHEBXIIK_Eu",
        "outputId": "c097115f-36b7-4e36-9484-45386a682b94"
      },
      "source": [
        "df1 = df1.drop(columns='Unnamed: 0') # This line is not required if the input dataset has the proper 13 columns\n",
        "print(df1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "              0         1         2  ...        10        11        12\n",
            "0      0.388266  1.066190   1.65756  ...  -12.1232  -36.5231  -76.7150\n",
            "1     -0.450752 -0.833231  -1.64948  ...   24.5506   32.9202   40.7282\n",
            "2      1.080110  2.199080   4.15661  ...   42.8811   61.5420   89.7588\n",
            "3     -2.602280 -7.072710 -10.71260  ...  -40.4210  -47.4739  -66.2680\n",
            "4      0.085928  0.266413   2.94087  ...  108.9250  105.9890   95.7408\n",
            "...         ...       ...       ...  ...       ...       ...       ...\n",
            "16344  0.424090  0.911776   1.41807  ...   17.8851   16.7183   18.3112\n",
            "16345  0.210162  2.838150   7.73274  ...   20.9781   24.7027   34.6982\n",
            "16346  0.570230  1.158780   1.30524  ...   69.3984   86.1083  108.1730\n",
            "16347  0.645774  2.287450   2.81709  ...   17.2922   14.9753  -19.0565\n",
            "16348  0.895050  3.375180   3.97640  ...   19.4355   31.8020   33.9234\n",
            "\n",
            "[16349 rows x 13 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5nSJF6aOchNx"
      },
      "source": [
        "def z_score(df):\n",
        "    # copy the dataframe\n",
        "    df_std = df.copy()\n",
        "    df_std = df_std.transpose()\n",
        "    # apply the z-score method\n",
        "    for column in df_std.columns:\n",
        "        df_std[column] = (df_std[column] - df_std[column].mean()) / df_std[column].std()\n",
        "    df_std = df_std.transpose()   \n",
        "    return df_std"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z71_tP5flZ4q"
      },
      "source": [
        "df1=z_score(df1)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N2Xou1cAliEK"
      },
      "source": [
        "BUFFER_SIZE = 16300\n",
        "BATCH_SIZE = 1\n",
        "\n",
        "train_dataset_new = tf.data.Dataset.from_tensor_slices(df1).shuffle(BUFFER_SIZE).batch(BATCH_SIZE)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lV_MBL1Jl6h8"
      },
      "source": [
        "def make_generator_model():\n",
        "    model = tf.keras.Sequential()\n",
        "    \n",
        "    model.add(layers.Dense(4, use_bias=False, input_shape=(100,)))\n",
        "    model.add(layers.LeakyReLU())\n",
        "\n",
        "    model.add(layers.Dense(2*64,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "    model.add(layers.Dropout(0.2))\n",
        "\n",
        "    model.add(layers.Dense(2*64,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "    model.add(layers.Dropout(0.2))\n",
        "    \n",
        "    model.add(layers.Dense(13,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "\n",
        "    model.add(layers.Reshape((13,1)))\n",
        "    \n",
        "    return model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YLkweKKemQIX"
      },
      "source": [
        "generator = make_generator_model()\n",
        "# Trying out a sample generation\n",
        "noise = tf.random.normal([1, 100])\n",
        "generated_image = generator(noise, training=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7LY7LFDFtAGV"
      },
      "source": [
        "def make_discriminator_model():\n",
        "    model = tf.keras.Sequential()\n",
        "\n",
        "    model.add(layers.Flatten())\n",
        "    model.add(layers.Dense(13,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "    \n",
        "    model.add(layers.Dense(64,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "    model.add(layers.Dropout(0.2))\n",
        "      \n",
        "    model.add(layers.Dense(64,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "    model.add(layers.Dropout(0.2))\n",
        "   \n",
        "    model.add(layers.Dense(1,use_bias=False))\n",
        "    model.add(layers.LeakyReLU())\n",
        "\n",
        "    return model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qqLVAzQQzXhx"
      },
      "source": [
        "discriminator = make_discriminator_model()\n",
        "decision = discriminator(generated_image)\n",
        "print (decision)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m0OQY27bzlU9"
      },
      "source": [
        "cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hM9dqBUIz0tq"
      },
      "source": [
        "def discriminator_loss(real_output, fake_output):\n",
        "    real_loss = cross_entropy(tf.ones_like(real_output), real_output)\n",
        "    fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)\n",
        "    total_loss = real_loss + fake_loss\n",
        "    return total_loss"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WBrCWW_a0KIK"
      },
      "source": [
        "def generator_loss(fake_output):\n",
        "    return cross_entropy(tf.ones_like(fake_output), fake_output)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJoBUXGV0MQ2"
      },
      "source": [
        "generator_optimizer = tf.keras.optimizers.Adam(1e-4)\n",
        "discriminator_optimizer = tf.keras.optimizers.Adam(1e-4)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X93srOkw0OVe"
      },
      "source": [
        "EPOCHS = 15\n",
        "noise_dim = 100\n",
        "num_examples_to_generate = 10\n",
        "\n",
        "seed = tf.random.normal([num_examples_to_generate, noise_dim])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ri93N4c0WLW"
      },
      "source": [
        "@tf.function\n",
        "def train_step(images,n):\n",
        "    noise = tf.random.normal([BATCH_SIZE, noise_dim])\n",
        "\n",
        "    with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\n",
        "      generated_images = generator(noise, training=True)\n",
        "\n",
        "      real_output = discriminator(images, training=True)\n",
        "      fake_output = discriminator(generated_images, training=True)\n",
        "      \n",
        "      gen_loss = generator_loss(fake_output)\n",
        "      disc_loss = discriminator_loss(real_output, fake_output)\n",
        "      \n",
        "    gradients_of_generator = gen_tape.gradient(gen_loss, generator.trainable_variables)\n",
        "    gradients_of_discriminator = disc_tape.gradient(disc_loss, discriminator.trainable_variables)\n",
        "\n",
        "    generator_optimizer.apply_gradients(zip(gradients_of_generator, generator.trainable_variables))\n",
        "    discriminator_optimizer.apply_gradients(zip(gradients_of_discriminator, discriminator.trainable_variables))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GkEBWU352ZoP"
      },
      "source": [
        "def train(dataset, epochs):\n",
        "  for i in range(epochs):\n",
        "    start = time.time()\n",
        "\n",
        "    for image_batch in dataset:\n",
        "      train_step(image_batch,i)\n",
        "\n",
        "    pddf = generate_and_save_images(generator, i + 1,seed)\n",
        "    print('Time for epoch {} is {} sec'.format(i + 1, time.time()-start))\n",
        "    if i == epochs-1:\n",
        "      pddf.to_csv('file_name.csv'.format(i+1),mode = 'a')\n",
        "      files.download('file_name.csv'.format(i+1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rNEj_0Y05PrC"
      },
      "source": [
        "def generate_and_save_images(model, test_input):\n",
        "  predictions = model(test_input, training=False)\n",
        "   \n",
        "  predictions = tf.reshape(predictions,[num_examples_to_generate,13])\n",
        "  predict_data = predictions.numpy()\n",
        "  panda_df = pd.DataFrame(data = predict_data)\n",
        "\n",
        "  return panda_df"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8WoXKR3n6STa"
      },
      "source": [
        "train(train_dataset_new, EPOCHS)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LV8Sui24QSHl"
      },
      "source": [
        "new_df = generate_and_save_images(generator,tf.random.normal([20000,100]))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tuDz0YJLQj3J"
      },
      "source": [
        "new_df.to_csv('20000 generated posx.csv')"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}