{
"nbformat": 4,
"nbformat_minor": 0,
"metadata": {
"accelerator": "GPU",
"colab": {
"name": "Female_names_generator_using_charRNN.ipynb",
"provenance": [],
"collapsed_sections": []
},
"kernelspec": {
"display_name": "Python 3",
"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": "3.6.9"
}
},
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "SF2XaOyEqjcU",
"colab_type": "text"
},
"source": [
"# CS568:Deep Learning Spring 2020 "
]
},
{
"cell_type": "code",
"metadata": {
"id": "MweShd4nqqMd",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 122
},
"outputId": "153374a2-e92e-4b4c-e10f-d4cf03b07703"
},
"source": [
"from google.colab import drive\n",
"drive.mount('/content/drive')"
],
"execution_count": 1,
"outputs": [
{
"output_type": "stream",
"text": [
"Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
"\n",
"Enter your authorization code:\n",
"··········\n",
"Mounted at /content/drive\n"
],
"name": "stdout"
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "SirtjPxJqjcV",
"colab_type": "text"
},
"source": [
"## Load Data"
]
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "HXA8rCMRJEBH",
"colab": {}
},
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"from keras.models import Sequential\n",
"from keras import layers\n",
"from keras.optimizers import Adam\n",
"import sys\n",
"\n",
"max_length = 2\n",
"learning_rate = 0.01\n",
"num_epochs = 100\n",
"batch_size = 1000\n",
"\n",
"def load_data(filename):\n",
" df = pd.read_csv(filename)\n",
" df = df.filter(['Name'])\n",
" df = np.array(df)\n",
" return df\n",
"\n",
"def preprocess_data(data): \n",
" inputs = []\n",
" targets = [] \n",
" vocab = '' \n",
" for item in data: \n",
" item = str(np.squeeze(item)) \n",
" \n",
" # track all possible characters to generate\n",
" vocab += item\n",
" \n",
" # create tokens from each name\n",
" for i in range(len(item) - max_length):\n",
" inputs.append(item[i : i + max_length])\n",
" targets.append(item[i + max_length])\n",
"\n",
" # get list of unique characters to generate from\n",
" chars = sorted(list(set(vocab)))\n",
" data_size, chars_size = len(data), len(chars)\n",
" print(\"Data has {} characters, {} unique\".format(str(data_size), str(chars_size)))\n",
"\n",
" char_indices = dict((ch, chars.index(ch)) for ch in chars)\n",
" \n",
" # create empty numpy arrays for X and y\n",
" X = np.zeros((len(inputs), max_length, chars_size), dtype=np.bool)\n",
" t = np.zeros((len(inputs), chars_size), dtype=np.bool)\n",
"\n",
" # one-hot encode selections\n",
" for inp, indiv_input in enumerate(inputs):\n",
" for tar, indiv_char in enumerate(indiv_input):\n",
" X[inp, tar, char_indices[indiv_char]] = 1\n",
" t[inp, char_indices[targets[inp]]] = 1\n",
" \n",
" return X, t, chars, char_indices"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"colab_type": "text",
"id": "mgs62W3NMbr1"
},
"source": [
"## Define model"
]
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "82ABbbwyL5eZ",
"colab": {}
},
"source": [
"def build_model(chars_length):\n",
" model = Sequential()\n",
" model.add(layers.LSTM(128, input_shape=(max_length, chars_length)))\n",
" model.add(layers.Dense(chars_length, activation='softmax')) \n",
" model.compile(loss='categorical_crossentropy', optimizer= Adam(lr = learning_rate)) \n",
" return model"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "BMFcVFdZOX-e",
"colab": {}
},
"source": [
"def sample(output, total):\n",
" output = np.asarray(output).astype('float64')\n",
" output = np.log(output) / total\n",
" exp_output = np.exp(output)\n",
" output = exp_output / np.sum(exp_output)\n",
" probas = np.random.multinomial(1, output, 1)\n",
" out = np.argmax(probas)\n",
" return out\n",
" \n",
"def generate_names(seed, length, chars, char_indices, char_length):\n",
" generated_text = seed\n",
" name = seed\n",
"\n",
" for i in range(length - len(seed)):\n",
" sampled = np.zeros((1, max_length, char_length))\n",
"\n",
" for tar, char in enumerate(generated_text): \n",
" # print(generated_text) \n",
" # print(char, char_indices[char], tar) \n",
" sampled[0, tar, char_indices[char]] = 1. \n",
"\n",
" preds = model.predict(sampled, verbose=0)[0]\n",
" next_index = sample(preds, 0.5)\n",
" next_char = chars[next_index]\n",
"\n",
" generated_text += next_char\n",
" generated_text = generated_text[1:]\n",
"\n",
" name += next_char\n",
" \n",
" return name\n"
],
"execution_count": 0,
"outputs": []
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "DDsA2NgKNYB0",
"outputId": "f312858d-b83b-40ee-9157-da88c6fa9007",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 1000
}
},
"source": [
"data = load_data(\"/content/drive/My Drive/muslim_girls_names.csv\")\n",
"print(data.shape, data.dtype)\n",
"X, t, chars, char_indices = preprocess_data(data)\n",
"print(\"X.shape \", X.shape)\n",
"print(\"t.shape \", t.shape)\n",
"\n",
"model = build_model(len(chars))\n",
"model.summary()\n",
"model.fit(X, t, epochs=num_epochs, batch_size=batch_size)"
],
"execution_count": 6,
"outputs": [
{
"output_type": "stream",
"text": [
"(4442, 1) object\n",
"Data has 4442 characters, 52 unique\n",
"X.shape (19500, 2, 52)\n",
"t.shape (19500, 52)\n",
"Model: \"sequential_1\"\n",
"_________________________________________________________________\n",
"Layer (type) Output Shape Param # \n",
"=================================================================\n",
"lstm_1 (LSTM) (None, 128) 92672 \n",
"_________________________________________________________________\n",
"dense_1 (Dense) (None, 52) 6708 \n",
"=================================================================\n",
"Total params: 99,380\n",
"Trainable params: 99,380\n",
"Non-trainable params: 0\n",
"_________________________________________________________________\n",
"Epoch 1/100\n",
"19500/19500 [==============================] - 2s 105us/step - loss: 3.0127\n",
"Epoch 2/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 2.2989\n",
"Epoch 3/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 2.1260\n",
"Epoch 4/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 2.0619\n",
"Epoch 5/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 2.0289\n",
"Epoch 6/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 2.0117\n",
"Epoch 7/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9948\n",
"Epoch 8/100\n",
"19500/19500 [==============================] - 0s 9us/step - loss: 1.9815\n",
"Epoch 9/100\n",
"19500/19500 [==============================] - 0s 9us/step - loss: 1.9688\n",
"Epoch 10/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9593\n",
"Epoch 11/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9502\n",
"Epoch 12/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.9409\n",
"Epoch 13/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.9341\n",
"Epoch 14/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.9243\n",
"Epoch 15/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9186\n",
"Epoch 16/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9122\n",
"Epoch 17/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.9065\n",
"Epoch 18/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8993\n",
"Epoch 19/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8987\n",
"Epoch 20/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8901\n",
"Epoch 21/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8880\n",
"Epoch 22/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8842\n",
"Epoch 23/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8794\n",
"Epoch 24/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8791\n",
"Epoch 25/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8728\n",
"Epoch 26/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8672\n",
"Epoch 27/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8680\n",
"Epoch 28/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8627\n",
"Epoch 29/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8590\n",
"Epoch 30/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8584\n",
"Epoch 31/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8529\n",
"Epoch 32/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8527\n",
"Epoch 33/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8516\n",
"Epoch 34/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8460\n",
"Epoch 35/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8450\n",
"Epoch 36/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8438\n",
"Epoch 37/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8414\n",
"Epoch 38/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8409\n",
"Epoch 39/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8400\n",
"Epoch 40/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8382\n",
"Epoch 41/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8348\n",
"Epoch 42/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8321\n",
"Epoch 43/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8323\n",
"Epoch 44/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8294\n",
"Epoch 45/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8307\n",
"Epoch 46/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8257\n",
"Epoch 47/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8269\n",
"Epoch 48/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8253\n",
"Epoch 49/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8243\n",
"Epoch 50/100\n",
"19500/19500 [==============================] - 0s 9us/step - loss: 1.8226\n",
"Epoch 51/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8244\n",
"Epoch 52/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8210\n",
"Epoch 53/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8198\n",
"Epoch 54/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8186\n",
"Epoch 55/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8189\n",
"Epoch 56/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8169\n",
"Epoch 57/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8177\n",
"Epoch 58/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8151\n",
"Epoch 59/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8130\n",
"Epoch 60/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8129\n",
"Epoch 61/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8131\n",
"Epoch 62/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8118\n",
"Epoch 63/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8120\n",
"Epoch 64/100\n",
"19500/19500 [==============================] - 0s 9us/step - loss: 1.8132\n",
"Epoch 65/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8110\n",
"Epoch 66/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8089\n",
"Epoch 67/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8094\n",
"Epoch 68/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8080\n",
"Epoch 69/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8089\n",
"Epoch 70/100\n",
"19500/19500 [==============================] - 0s 9us/step - loss: 1.8094\n",
"Epoch 71/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8095\n",
"Epoch 72/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8076\n",
"Epoch 73/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8052\n",
"Epoch 74/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8069\n",
"Epoch 75/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8050\n",
"Epoch 76/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8043\n",
"Epoch 77/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8035\n",
"Epoch 78/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8027\n",
"Epoch 79/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8027\n",
"Epoch 80/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8040\n",
"Epoch 81/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8027\n",
"Epoch 82/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8024\n",
"Epoch 83/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8016\n",
"Epoch 84/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8004\n",
"Epoch 85/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7987\n",
"Epoch 86/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8002\n",
"Epoch 87/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.8012\n",
"Epoch 88/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.7994\n",
"Epoch 89/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.8012\n",
"Epoch 90/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.7982\n",
"Epoch 91/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.7992\n",
"Epoch 92/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7991\n",
"Epoch 93/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7975\n",
"Epoch 94/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7972\n",
"Epoch 95/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7991\n",
"Epoch 96/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7972\n",
"Epoch 97/100\n",
"19500/19500 [==============================] - 0s 8us/step - loss: 1.7973\n",
"Epoch 98/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7965\n",
"Epoch 99/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7970\n",
"Epoch 100/100\n",
"19500/19500 [==============================] - 0s 7us/step - loss: 1.7955\n"
],
"name": "stdout"
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
""
]
},
"metadata": {
"tags": []
},
"execution_count": 6
}
]
},
{
"cell_type": "code",
"metadata": {
"colab_type": "code",
"id": "pKFXJqYIOwSf",
"outputId": "df92996d-1a0e-4114-9487-9e38bf89157a",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
}
},
"source": [
"generate_names(\"Ab\", 7, chars, char_indices, len(chars))"
],
"execution_count": 11,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Abeerah'"
]
},
"metadata": {
"tags": []
},
"execution_count": 11
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "jAh4eO25qjcs",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "b8e6fe4f-04fd-4002-d8ec-076e612758d7"
},
"source": [
"generate_names(\"An\", 7, chars, char_indices, len(chars))"
],
"execution_count": 17,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Anaheen'"
]
},
"metadata": {
"tags": []
},
"execution_count": 17
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "Yp-IT69XrUvU",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "22bc5b8f-a665-4f61-fe2f-b337e6c91af4"
},
"source": [
"generate_names(\"Mi\", 7, chars, char_indices, len(chars))"
],
"execution_count": 18,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Mishadi'"
]
},
"metadata": {
"tags": []
},
"execution_count": 18
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "bTAkewJtrYIu",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "eb55284e-c90e-4c04-96a6-c2c67844ad59"
},
"source": [
"generate_names(\"Ma\", 5, chars, char_indices, len(chars))"
],
"execution_count": 31,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Mahid'"
]
},
"metadata": {
"tags": []
},
"execution_count": 31
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "II-QYNrGrjqB",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "5f567fa4-6165-4ac3-95b7-605a86f2cddf"
},
"source": [
"generate_names(\"Z\", 6, chars, char_indices, len(chars))"
],
"execution_count": 34,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Zairam'"
]
},
"metadata": {
"tags": []
},
"execution_count": 34
}
]
},
{
"cell_type": "code",
"metadata": {
"id": "CV5BpbumrpSJ",
"colab_type": "code",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 34
},
"outputId": "1beb1a06-c1b7-404e-b29a-42e4ffc7005c"
},
"source": [
"generate_names(\"Za\", 8, chars, char_indices, len(chars))"
],
"execution_count": 38,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"'Zareenah'"
]
},
"metadata": {
"tags": []
},
"execution_count": 38
}
]
}
]
}