diff --git a/TGAN.ipynb b/TGAN.ipynb
new file mode 100644
index 0000000000000000000000000000000000000000..2a6038a592fb521e58da9fcbecb216ec9be822c1
--- /dev/null
+++ b/TGAN.ipynb
@@ -0,0 +1,226 @@
+{
+ "cells": [
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "fefebe4b",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#https://github.com/sdv-dev/TGAN/tree/master\n",
+ "import pandas as pd\n",
+ "from tgan.data import load_demo_data\n",
+ "from tgan.model import TGANModel\n",
+ "import tensorflow as tf\n",
+ "import numpy as np"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "a345cefc",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#loading datasets\n",
+ "def load_data(location):\n",
+ " data = pd.read_csv(location)\n",
+ " data_columns = data.columns\n",
+ " data = preprocessing(data)\n",
+ " return data, data_columns\n",
+ "\n",
+ "\n",
+ "#Dataset preprocessing\n",
+ "def preprocessing(data):\n",
+ " \n",
+ " \"\"\"\n",
+ " dropping duplicate values\n",
+ " changing timeformat to d/m/Y H:M:S and then to Unix fomrat that starts from 1970/1/ 00:00:00\n",
+ " making sure that numeric columns only have numeric values and if not numeric then to NaN\n",
+ " dropping all NaN values \n",
+ " \"\"\"\n",
+ " print(\"Shape of data before preprocessing:\", data.shape)\n",
+ " data.drop_duplicates(inplace=True) #dropping duplicated\n",
+ " data.replace([np.inf, -np.inf], np.nan, inplace=True)# changing inf and -inf to nan\n",
+ "\n",
+ " data['Timestamp'] = pd.to_datetime(data['Timestamp'], format='%d/%m/%Y %H:%M:%S', errors='coerce')\n",
+ " data['Timestamp'] = (data['Timestamp'] - pd.Timestamp(\"1970-01-01\")) // pd.Timedelta('1s') \n",
+ "\n",
+ " for col in data.columns: #changing columns to numeric if not, then to NaN\n",
+ " if data[col].dtype == 'object' and col != 'Label':\n",
+ " data[col] = pd.to_numeric(data[col], errors='coerce')\n",
+ " \n",
+ " \n",
+ " \n",
+ " data.dropna(inplace=True) #droping Na\n",
+ " \n",
+ " print(\"Shape of data after preprocessing:\", data.shape)\n",
+ " \n",
+ " return data"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "41dfaeff",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def tGAN(data, continuous_columns,max_epoch=5, steps_per_epoch=10000, batch_size=200, z_dim=200, noise=0.2, l2norm=0.00001, \n",
+ " learning_rate=0.001, num_gen_rnn=100, num_gen_feature=100, num_dis_layers=1, num_dis_hidden=100, \n",
+ " optimizer='AdamOptimizer'):\n",
+ " \n",
+ " \"\"\"\n",
+ " Required arguments to be passed:\n",
+ " -data: dataframe with rows and columns\n",
+ " -continuous_columns: a list containing all the columns that are continuous \n",
+ " \"\"\"\n",
+ "\n",
+ " print(data.shape)\n",
+ "\n",
+ " tgan = TGANModel(continuous_columns=continuous_columns, max_epoch=max_epoch, steps_per_epoch=steps_per_epoch, \n",
+ " batch_size=batch_size, z_dim=z_dim, noise=noise, l2norm=l2norm, learning_rate=learning_rate, \n",
+ " num_gen_rnn=num_gen_rnn, num_gen_feature=num_gen_feature, num_dis_layers=num_dis_layers, \n",
+ " num_dis_hidden=num_dis_hidden, optimizer=optimizer)\n",
+ " \n",
+ " tgan.fit(data)\n",
+ "\n",
+ " return tgan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "07714e0c",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "continuous_columns = [2, 3, 17, 18, 20, 21, 38, 39]\n",
+ "\n",
+ "data, data_columns = load_data(\"C:\\\\Users\\\\sayed\\\\Desktop\\\\Dataset\\\\02-14-2018.csv\")\n",
+ "\n",
+ "data = data[data[\"Label\"] == \"FTP-BruteForce\"]\n",
+ "data.columns = [None] * len(data.columns) # revoming column names"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "95f316b5",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#fitting the TGAN model\n",
+ "tgan= tGAN(data=data, continuous_columns=continuous_columns, batch_size=150, max_epoch = 15)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "d0b5a40e",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#Saving the model\n",
+ "model_path = 'C:\\\\Users\\\\sayed\\\\Desktop\\\\Dataset\\\\models\\\\tGAN_model_firstrun.pkl'\n",
+ "tgan.save(model_path)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "d7343348",
+ "metadata": {
+ "scrolled": true
+ },
+ "outputs": [],
+ "source": [
+ "num_samples = 8000\n",
+ "new_tgan = TGANModel.load(model_path)\n",
+ "samples = new_tgan.sample(num_samples)\n",
+ "samples.head()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "20667896",
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "#assinging back the column names\n",
+ "samples.columns = data_columns\n",
+ "data.columns = data_columns"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "4bdf2761",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "8423e0de",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "9ae7fce1",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "d2fd753c",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "06a0123c",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "ebf1c5cd",
+ "metadata": {},
+ "outputs": [],
+ "source": []
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "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": "3.7.16"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}