Skip to content

Instantly share code, notes, and snippets.

@mirrornerror

mirrornerror/auto-keras-mnist.ipynb

Created November 5, 2018 03:10
Show Gist options
  • Save mirrornerror/6a2172b2d45bb5b22fdd5b7691b4e321 to your computer and use it in GitHub Desktop.
Save mirrornerror/6a2172b2d45bb5b22fdd5b7691b4e321 to your computer and use it in GitHub Desktop.
Download ZIP
Kaggle: Digit Recognizer, Auto-Keras
Display the source blob
Display the rendered blob
Raw
auto-keras-mnist.ipynb
{
"cells": [
{
"metadata": {},
"cell_type": "markdown",
"source": "## Kaggle: Digit Recognizer by Auto-Keras \ndata: https://www.kaggle.com/c/digit-recognizer \nlibrary: https://autokeras.com\n \nreference: https://www.simonwenkel.com/2018/08/29/introduction-to-autokeras.html"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T22:27:25.426155Z",
"start_time": "2018-11-04T22:27:24.026022Z"
},
"trusted": true
},
"cell_type": "code",
"source": "from autokeras.image_supervised import ImageClassifier\nimport matplotlib.pyplot as plt\nimport pandas as pd\n%matplotlib inline\nseed = 123",
"execution_count": 12,
"outputs": []
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Load MNIST(from Kaggle) data\n* split data to train data and validation data \n* no normalization "
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T22:27:30.134433Z",
"start_time": "2018-11-04T22:27:25.428126Z"
},
"trusted": true
},
"cell_type": "code",
"source": "train = pd.read_csv('../train.csv')\ntest = pd.read_csv('../test.csv')\n\nlabel = train.label\ntest_index = test.index\ntrain = train.drop(['label'], axis=1)\n\ntrain = train.values.reshape(-1,28,28,1)\ntest = test.values.reshape(-1,28,28,1)\nprint(test.dtype)\nlabel = label.astype('uint8')\ntrain = train.astype('uint8')\ntest = test.astype('uint8')\n\nfrom sklearn.model_selection import train_test_split\nx_train, x_val, y_train, y_val = train_test_split(train, label, test_size = 0.2, random_state=seed)\nprint(x_train.shape, y_train.shape, x_val.shape, y_val.shape)",
"execution_count": 2,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": "int64\n(33600, 28, 28, 1) (33600,) (8400, 28, 28, 1) (8400,)\n"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### visualize a sample image from the data:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T22:27:30.263687Z",
"start_time": "2018-11-04T22:27:30.135899Z"
},
"trusted": true
},
"cell_type": "code",
"source": "plt.imshow(x_val[0,:,:,0], cmap='gray')",
"execution_count": 3,
"outputs": [
{
"data": {
"text/plain": "<matplotlib.image.AxesImage at 0x7f93afd8e390>"
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
"image/png": "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\n",
"text/plain": "<matplotlib.figure.Figure at 0x7f93b032a048>"
},
"metadata": {},
"output_type": "display_data"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Train:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:29:34.604361Z",
"start_time": "2018-11-04T22:27:30.266562Z"
},
"trusted": true
},
"cell_type": "code",
"source": "clf = ImageClassifier(verbose=True, augment=True, searcher_args={'trainer_args':{'max_iter_num':5}})\nclf.fit(x_train, y_train, time_limit=1 * 60 * 60) # 1 hours",
"execution_count": 4,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": "\nInitializing search.\nInitialization finished.\n\n\n+----------------------------------------------+\n| Training model 0 |\n+----------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 0 | 4.164044760167599 | 0.9410714285714287 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 1 |\n+----------------------------------------------+\n"
},
{
"name": "stderr",
"output_type": "stream",
"text": "/home/mirrornerror/.pyenv/versions/anaconda3-5.1.0/envs/py36/lib/python3.6/site-packages/autokeras/bayesian.py:151: UserWarning: Predicted variances smaller than 0. Setting those variances to 0.\n warnings.warn(\"Predicted variances smaller than 0. \"\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| 0 | ('to_concat_skip_model', 1, 5) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 1 | 2.8610854782164097 | 0.9596428571428571 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 2 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| 1 | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 2 | 2.0033180736005307 | 0.9714285714285715 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 3 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_concat_skip_model', 1, 5) |\n| | ('to_conv_deeper_model', 5, 3) |\n| 2 | ('to_add_skip_model', 1, 5) |\n| | ('to_add_skip_model', 1, 23) |\n| | ('to_add_skip_model', 5, 23) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 3 | 1.5857322109863161 | 0.9770000000000001 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 4 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| 3 | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 4 | 1.6691380865871905 | 0.973857142857143 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 5 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| 3 | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_add_skip_model', 44, 47) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_add_skip_model', 18, 5) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 5 | 1.2643032357096673 | 0.9827142857142857 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 6 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| 5 | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 47, 3) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 47, 18) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 6 | 1.5100525204092263 | 0.9800000000000001 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 7 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| 5 | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 7 | 1.269026187248528 | 0.9820714285714285 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 8 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| 5 | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 8 | 1.2396936539560557 | 0.9833571428571428 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 9 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| 8 | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_concat_skip_model', 62, 9) |\n| | ('to_add_skip_model', 62, 44) |\n| | ('to_add_skip_model', 47, 24) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 9 | 1.311048859730363 | 0.9823571428571428 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 10 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| 8 | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n| | ('to_conv_deeper_model', 27, 3) |\n| | ('to_concat_skip_model', 47, 18) |\n| | ('to_concat_skip_model', 47, 9) |\n| | ('to_add_skip_model', 1, 18) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 10 | 1.53233876619488 | 0.9795 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 11 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| 8 | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_wider_model', 14, 128) |\n| | ('to_add_skip_model', 1, 5) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_wider_model', 14, 256) |\n| | ('to_add_skip_model', 18, 62) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 11 | 1.3345525920391084 | 0.9814999999999999 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 12 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| 8 | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n| | ('to_dense_deeper_model', 14) |\n| | ('to_conv_deeper_model', 47, 3) |\n| | ('to_wider_model', 18, 64) |\n| | ('to_add_skip_model', 47, 18) |\n| | ('to_concat_skip_model', 1, 18) |\n| | ('to_add_skip_model', 64, 24) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 12 | 1.3271637907251717 | 0.9812857142857144 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 13 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| 5 | ('to_add_skip_model', 47, 44) |\n| | ('to_dense_deeper_model', 14) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 47, 9) |\n| | ('to_add_skip_model', 5, 9) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_add_skip_model', 18, 5) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_concat_skip_model', 1, 5) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 13 | 1.394709468819201 | 0.9802857142857142 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 14 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| 9 | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 14 | 1.3999419607222081 | 0.9801428571428572 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 15 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| 8 | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_wider_model', 14, 64) |\n| | ('to_wider_model', 44, 64) |\n| | ('to_add_skip_model', 21, 5) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_concat_skip_model', 44, 24) |\n| | ('to_add_skip_model', 1, 24) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 15 | 1.2699033791199326 | 0.9835714285714285 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 16 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| 15 | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 74, 3) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 16 | 1.1803107248619198 | 0.9840714285714286 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 17 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| 15 | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 17 | 1.2850757971405984 | 0.9824999999999999 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 18 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| 15 | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 47, 3) |\n| | ('to_concat_skip_model', 77, 24) |\n| | ('to_concat_skip_model', 24, 74) |\n| | ('to_wider_model', 77, 64) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 18 | 1.1741155210882426 | 0.985 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 19 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| 18 | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_wider_model', 9, 64) |\n| | ('to_add_skip_model', 18, 74) |\n| | ('to_concat_skip_model', 1, 80) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 19 | 1.2402092449367046 | 0.9835714285714285 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 20 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| 18 | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 74, 3) |\n| | ('to_wider_model', 18, 64) |\n| | ('to_concat_skip_model', 5, 57) |\n| | ('to_wider_model', 74, 64) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 20 | 1.1882534926757216 | 0.9848571428571429 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 21 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| 18 | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 77, 3) |\n| | ('to_add_skip_model', 24, 80) |\n| | ('to_concat_skip_model', 18, 57) |\n| | ('to_wider_model', 87, 64) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 21 | 1.223137110285461 | 0.9833571428571428 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 22 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| 18 | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_dense_deeper_model', 14) |\n| | ('to_conv_deeper_model', 80, 3) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_wider_model', 9, 64) |\n| | ('to_concat_skip_model', 1, 89) |\n| | ('to_wider_model', 87, 64) |\n| | ('to_wider_model', 80, 128) |\n| | ('to_add_skip_model', 47, 24) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 22 | 1.0686696421355009 | 0.9869285714285713 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 23 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| 22 | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 77, 3) |\n| | ('to_add_skip_model', 24, 80) |\n| | ('to_concat_skip_model', 18, 57) |\n| | ('to_wider_model', 87, 64) |\n| | ('to_conv_deeper_model', 74, 3) |\n| | ('to_concat_skip_model', 77, 80) |\n| | ('to_concat_skip_model', 1, 57) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 23 | 1.3905318457633258 | 0.9829285714285714 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 24 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| | ('to_concat_skip_model', 18, 24) |\n| 22 | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 77, 3) |\n| | ('to_add_skip_model', 24, 80) |\n| | ('to_concat_skip_model', 18, 57) |\n| | ('to_wider_model', 87, 64) |\n| | ('to_conv_deeper_model', 47, 3) |\n| | ('to_add_skip_model', 21, 77) |\n| | ('to_concat_skip_model', 24, 80) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 24 | 1.328992754034698 | 0.9837142857142857 |\n+--------------------------------------------------------------------------+\n"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\n\n+----------------------------------------------+\n| Training model 25 |\n+----------------------------------------------+\n\n+--------------------------------------------------------------------------+\n| Father Model ID | Added Operation |\n+--------------------------------------------------------------------------+\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 1, 3) |\n| | ('to_conv_deeper_model', 5, 3) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_add_skip_model', 18, 24) |\n| | ('to_add_skip_model', 24, 27) |\n| | ('to_conv_deeper_model', 18, 3) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_add_skip_model', 47, 44) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_add_skip_model', 5, 54) |\n| | ('to_wider_model', 54, 64) |\n| | ('to_wider_model', 5, 64) |\n| 22 | ('to_concat_skip_model', 18, 24) |\n| | ('to_add_skip_model', 1, 18) |\n| | ('to_conv_deeper_model', 24, 3) |\n| | ('to_conv_deeper_model', 44, 3) |\n| | ('to_wider_model', 47, 64) |\n| | ('to_wider_model', 24, 64) |\n| | ('to_conv_deeper_model', 57, 3) |\n| | ('to_add_skip_model', 21, 18) |\n| | ('to_wider_model', 80, 64) |\n| | ('to_conv_deeper_model', 9, 3) |\n| | ('to_conv_deeper_model', 77, 3) |\n| | ('to_add_skip_model', 24, 80) |\n| | ('to_concat_skip_model', 18, 57) |\n| | ('to_wider_model', 87, 64) |\n| | ('to_conv_deeper_model', 21, 3) |\n| | ('to_conv_deeper_model', 74, 3) |\n+--------------------------------------------------------------------------+\n\nSaving model.\n+--------------------------------------------------------------------------+\n| Model ID | Loss | Metric Value |\n+--------------------------------------------------------------------------+\n| 25 | 1.3192439060658216 | 0.9845714285714285 |\n+--------------------------------------------------------------------------+\n\n\n+----------------------------------------------+\n| Training model 26 |\n+----------------------------------------------+\n\nTime limit for model search is reached. Ending the model search.\n"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Search the Best model:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:35:49.020783Z",
"start_time": "2018-11-04T23:29:34.611571Z"
},
"trusted": true
},
"cell_type": "code",
"source": "clf.final_fit(x_train, y_train, x_val, y_val, retrain=False, trainer_args={'max_iter_num':10})",
"execution_count": 5,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": "\nLoading and training the best model recorded from the search.\n"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-1, Current Metric - 0', layout=Layout(flex='2'), max=26…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-2, Current Metric - 0.9876190476190476', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-3, Current Metric - 0.9901190476190476', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-4, Current Metric - 0.9851190476190477', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-5, Current Metric - 0.9876190476190476', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-6, Current Metric - 0.9870238095238095', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-7, Current Metric - 0.9852380952380952', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-8, Current Metric - 0.9866666666666667', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\r"
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": "HBox(children=(IntProgress(value=0, description='Epoch-9, Current Metric - 0.9889285714285714', layout=Layout(…"
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": "\nNo loss decrease after 5 epochs.\n\n"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Evaluate the result:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:37:18.351462Z",
"start_time": "2018-11-04T23:35:49.022963Z"
},
"scrolled": true,
"trusted": true
},
"cell_type": "code",
"source": "y = clf.evaluate(x_val, y_val)\nprint(y) \n# 0.9873809523809524(result from one-hour-training)\n# 0.9896428571428572(result from five-hour-training)",
"execution_count": 6,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": "0.9853571428571428\n"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Predict the submission test data:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:42:15.687771Z",
"start_time": "2018-11-04T23:37:18.353119Z"
},
"trusted": true
},
"cell_type": "code",
"source": "pred = clf.predict(test)\npred.shape",
"execution_count": 7,
"outputs": [
{
"data": {
"text/plain": "(28000,)"
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Best model: \noutput the number of layers "
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:42:15.698325Z",
"start_time": "2018-11-04T23:42:15.689495Z"
},
"trusted": true
},
"cell_type": "code",
"source": "best_model = clf.load_searcher().load_best_model()\nbest_model.n_layers",
"execution_count": 8,
"outputs": [
{
"data": {
"text/plain": "101"
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Save the best model:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:42:23.238163Z",
"start_time": "2018-11-04T23:42:15.699687Z"
},
"code_folding": [],
"scrolled": true,
"trusted": true
},
"cell_type": "code",
"source": "clf.load_searcher().load_best_model().produce_keras_model().save('autokeras_mnist_model.h5')",
"execution_count": 9,
"outputs": []
},
{
"metadata": {},
"cell_type": "markdown",
"source": "## Load the best model:\nshow the model summary"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:42:26.178889Z",
"start_time": "2018-11-04T23:42:23.239424Z"
},
"trusted": true
},
"cell_type": "code",
"source": "from keras.models import load_model\nmodel = load_model('autokeras_mnist_model.h5')\nmodel.summary()",
"execution_count": 10,
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": "__________________________________________________________________________________________________\nLayer (type) Output Shape Param # Connected to \n==================================================================================================\ninput_1 (InputLayer) (None, 28, 28, 1) 0 \n__________________________________________________________________________________________________\nactivation_1 (Activation) (None, 28, 28, 1) 0 input_1[0][0] \n__________________________________________________________________________________________________\nconv2d_1 (Conv2D) (None, 28, 28, 64) 640 activation_1[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_1 (BatchNor (None, 28, 28, 64) 256 conv2d_1[0][0] \n__________________________________________________________________________________________________\nactivation_6 (Activation) (None, 28, 28, 64) 0 batch_normalization_1[0][0] \n__________________________________________________________________________________________________\nconv2d_5 (Conv2D) (None, 28, 28, 64) 36928 activation_6[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_5 (BatchNor (None, 28, 28, 64) 256 conv2d_5[0][0] \n__________________________________________________________________________________________________\nactivation_13 (Activation) (None, 28, 28, 64) 0 batch_normalization_5[0][0] \n__________________________________________________________________________________________________\nconv2d_12 (Conv2D) (None, 28, 28, 128) 73856 activation_13[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_12 (BatchNo (None, 28, 28, 128) 512 conv2d_12[0][0] \n__________________________________________________________________________________________________\nactivation_5 (Activation) (None, 28, 28, 128) 0 batch_normalization_12[0][0] \n__________________________________________________________________________________________________\nactivation_23 (Activation) (None, 28, 28, 64) 0 batch_normalization_5[0][0] \n__________________________________________________________________________________________________\nconv2d_4 (Conv2D) (None, 28, 28, 64) 73792 activation_5[0][0] \n__________________________________________________________________________________________________\nconv2d_22 (Conv2D) (None, 28, 28, 64) 4160 activation_23[0][0] \n__________________________________________________________________________________________________\nactivation_19 (Activation) (None, 28, 28, 64) 0 batch_normalization_1[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_4 (BatchNor (None, 28, 28, 64) 256 conv2d_4[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_22 (BatchNo (None, 28, 28, 64) 256 conv2d_22[0][0] \n__________________________________________________________________________________________________\nconv2d_18 (Conv2D) (None, 28, 28, 64) 4160 activation_19[0][0] \n__________________________________________________________________________________________________\nadd_7 (Add) (None, 28, 28, 64) 0 batch_normalization_4[0][0] \n batch_normalization_22[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_18 (BatchNo (None, 28, 28, 64) 256 conv2d_18[0][0] \n__________________________________________________________________________________________________\nadd_6 (Add) (None, 28, 28, 64) 0 add_7[0][0] \n batch_normalization_18[0][0] \n__________________________________________________________________________________________________\nactivation_12 (Activation) (None, 28, 28, 64) 0 add_6[0][0] \n__________________________________________________________________________________________________\nconv2d_11 (Conv2D) (None, 28, 28, 64) 36928 activation_12[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_11 (BatchNo (None, 28, 28, 64) 256 conv2d_11[0][0] \n__________________________________________________________________________________________________\nactivation_21 (Activation) (None, 28, 28, 64) 0 batch_normalization_11[0][0] \n__________________________________________________________________________________________________\nconv2d_20 (Conv2D) (None, 28, 28, 64) 36928 activation_21[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_20 (BatchNo (None, 28, 28, 64) 256 conv2d_20[0][0] \n__________________________________________________________________________________________________\nactivation_25 (Activation) (None, 28, 28, 64) 0 batch_normalization_20[0][0] \n__________________________________________________________________________________________________\nactivation_14 (Activation) (None, 28, 28, 128) 0 batch_normalization_12[0][0] \n__________________________________________________________________________________________________\nconv2d_24 (Conv2D) (None, 28, 28, 64) 36928 activation_25[0][0] \n__________________________________________________________________________________________________\nconv2d_13 (Conv2D) (None, 28, 28, 64) 8256 activation_14[0][0] \n__________________________________________________________________________________________________\nactivation_9 (Activation) (None, 28, 28, 64) 0 batch_normalization_1[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_24 (BatchNo (None, 28, 28, 64) 256 conv2d_24[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_13 (BatchNo (None, 28, 28, 64) 256 conv2d_13[0][0] \n__________________________________________________________________________________________________\nconv2d_8 (Conv2D) (None, 28, 28, 64) 4160 activation_9[0][0] \n__________________________________________________________________________________________________\nadd_4 (Add) (None, 28, 28, 64) 0 batch_normalization_24[0][0] \n batch_normalization_13[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_8 (BatchNor (None, 28, 28, 64) 256 conv2d_8[0][0] \n__________________________________________________________________________________________________\nadd_1 (Add) (None, 28, 28, 64) 0 add_4[0][0] \n batch_normalization_8[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_1 (MaxPooling2D) (None, 14, 14, 64) 0 add_1[0][0] \n__________________________________________________________________________________________________\nactivation_2 (Activation) (None, 14, 14, 64) 0 max_pooling2d_1[0][0] \n__________________________________________________________________________________________________\nconv2d_2 (Conv2D) (None, 14, 14, 128) 73856 activation_2[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_2 (BatchNor (None, 14, 14, 128) 512 conv2d_2[0][0] \n__________________________________________________________________________________________________\nactivation_7 (Activation) (None, 14, 14, 128) 0 batch_normalization_2[0][0] \n__________________________________________________________________________________________________\nconv2d_6 (Conv2D) (None, 14, 14, 128) 147584 activation_7[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_6 (BatchNor (None, 14, 14, 128) 512 conv2d_6[0][0] \n__________________________________________________________________________________________________\nactivation_20 (Activation) (None, 14, 14, 128) 0 batch_normalization_6[0][0] \n__________________________________________________________________________________________________\nconv2d_19 (Conv2D) (None, 14, 14, 64) 73792 activation_20[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_19 (BatchNo (None, 14, 14, 64) 256 conv2d_19[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_7 (MaxPooling2D) (None, 14, 14, 64) 0 add_6[0][0] \n__________________________________________________________________________________________________\nconcatenate_1 (Concatenate) (None, 14, 14, 128) 0 batch_normalization_19[0][0] \n max_pooling2d_7[0][0] \n__________________________________________________________________________________________________\nactivation_18 (Activation) (None, 14, 14, 128) 0 concatenate_1[0][0] \n__________________________________________________________________________________________________\nconv2d_17 (Conv2D) (None, 14, 14, 64) 8256 activation_18[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_17 (BatchNo (None, 14, 14, 64) 256 conv2d_17[0][0] \n__________________________________________________________________________________________________\nactivation_16 (Activation) (None, 14, 14, 64) 0 batch_normalization_17[0][0] \n__________________________________________________________________________________________________\nconv2d_15 (Conv2D) (None, 14, 14, 64) 36928 activation_16[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_15 (BatchNo (None, 14, 14, 64) 256 conv2d_15[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_8 (MaxPooling2D) (None, 14, 14, 64) 0 batch_normalization_4[0][0] \n__________________________________________________________________________________________________\nconcatenate_2 (Concatenate) (None, 14, 14, 128) 0 batch_normalization_15[0][0] \n max_pooling2d_8[0][0] \n__________________________________________________________________________________________________\nactivation_27 (Activation) (None, 14, 14, 128) 0 concatenate_2[0][0] \n__________________________________________________________________________________________________\nconv2d_26 (Conv2D) (None, 14, 14, 64) 8256 activation_27[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_26 (BatchNo (None, 14, 14, 64) 256 conv2d_26[0][0] \n__________________________________________________________________________________________________\nactivation_22 (Activation) (None, 14, 14, 64) 0 batch_normalization_26[0][0] \n__________________________________________________________________________________________________\nactivation_26 (Activation) (None, 14, 14, 128) 0 batch_normalization_6[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_4 (MaxPooling2D) (None, 14, 14, 64) 0 add_4[0][0] \n__________________________________________________________________________________________________\nconv2d_21 (Conv2D) (None, 14, 14, 128) 73856 activation_22[0][0] \n__________________________________________________________________________________________________\nconv2d_25 (Conv2D) (None, 14, 14, 128) 16512 activation_26[0][0] \n__________________________________________________________________________________________________\nactivation_10 (Activation) (None, 14, 14, 64) 0 max_pooling2d_4[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_21 (BatchNo (None, 14, 14, 128) 512 conv2d_21[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_25 (BatchNo (None, 14, 14, 128) 512 conv2d_25[0][0] \n__________________________________________________________________________________________________\nconv2d_9 (Conv2D) (None, 14, 14, 128) 8320 activation_10[0][0] \n__________________________________________________________________________________________________\nadd_8 (Add) (None, 14, 14, 128) 0 batch_normalization_21[0][0] \n batch_normalization_25[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_9 (BatchNor (None, 14, 14, 128) 512 conv2d_9[0][0] \n__________________________________________________________________________________________________\nadd_2 (Add) (None, 14, 14, 128) 0 add_8[0][0] \n batch_normalization_9[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_2 (MaxPooling2D) (None, 7, 7, 128) 0 add_2[0][0] \n__________________________________________________________________________________________________\nactivation_3 (Activation) (None, 7, 7, 128) 0 max_pooling2d_2[0][0] \n__________________________________________________________________________________________________\nconv2d_3 (Conv2D) (None, 7, 7, 64) 73792 activation_3[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_3 (BatchNor (None, 7, 7, 64) 256 conv2d_3[0][0] \n__________________________________________________________________________________________________\nactivation_24 (Activation) (None, 7, 7, 64) 0 batch_normalization_3[0][0] \n__________________________________________________________________________________________________\nconv2d_23 (Conv2D) (None, 7, 7, 128) 73856 activation_24[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_23 (BatchNo (None, 7, 7, 128) 512 conv2d_23[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_6 (MaxPooling2D) (None, 7, 7, 128) 0 batch_normalization_2[0][0] \n__________________________________________________________________________________________________\nactivation_15 (Activation) (None, 7, 7, 128) 0 batch_normalization_23[0][0] \n__________________________________________________________________________________________________\nactivation_17 (Activation) (None, 7, 7, 128) 0 max_pooling2d_6[0][0] \n__________________________________________________________________________________________________\nconv2d_14 (Conv2D) (None, 7, 7, 128) 147584 activation_15[0][0] \n__________________________________________________________________________________________________\nconv2d_16 (Conv2D) (None, 7, 7, 128) 16512 activation_17[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_14 (BatchNo (None, 7, 7, 128) 512 conv2d_14[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_16 (BatchNo (None, 7, 7, 128) 512 conv2d_16[0][0] \n__________________________________________________________________________________________________\nadd_5 (Add) (None, 7, 7, 128) 0 batch_normalization_14[0][0] \n batch_normalization_16[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_5 (MaxPooling2D) (None, 7, 7, 128) 0 add_8[0][0] \n__________________________________________________________________________________________________\nactivation_8 (Activation) (None, 7, 7, 128) 0 add_5[0][0] \n__________________________________________________________________________________________________\nactivation_11 (Activation) (None, 7, 7, 128) 0 max_pooling2d_5[0][0] \n__________________________________________________________________________________________________\nconv2d_7 (Conv2D) (None, 7, 7, 64) 73792 activation_8[0][0] \n__________________________________________________________________________________________________\nconv2d_10 (Conv2D) (None, 7, 7, 64) 8256 activation_11[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_7 (BatchNor (None, 7, 7, 64) 256 conv2d_7[0][0] \n__________________________________________________________________________________________________\nbatch_normalization_10 (BatchNo (None, 7, 7, 64) 256 conv2d_10[0][0] \n__________________________________________________________________________________________________\nadd_3 (Add) (None, 7, 7, 64) 0 batch_normalization_7[0][0] \n batch_normalization_10[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_3 (MaxPooling2D) (None, 3, 3, 64) 0 add_3[0][0] \n__________________________________________________________________________________________________\nflatten_1 (Flatten) (None, 576) 0 max_pooling2d_3[0][0] \n__________________________________________________________________________________________________\ndropout_1 (Dropout) (None, 576) 0 flatten_1[0][0] \n__________________________________________________________________________________________________\ndense_1 (Dense) (None, 64) 36928 dropout_1[0][0] \n__________________________________________________________________________________________________\nactivation_4 (Activation) (None, 64) 0 dense_1[0][0] \n__________________________________________________________________________________________________\ndense_2 (Dense) (None, 10) 650 activation_4[0][0] \n==================================================================================================\nTotal params: 1,204,426\nTrainable params: 1,199,946\nNon-trainable params: 4,480\n__________________________________________________________________________________________________\n"
},
{
"name": "stderr",
"output_type": "stream",
"text": "/home/mirrornerror/.pyenv/versions/anaconda3-5.1.0/envs/py36/lib/python3.6/site-packages/keras/engine/saving.py:269: UserWarning: No training configuration found in save file: the model was *not* compiled. Compile it manually.\n warnings.warn('No training configuration found in save file: '\n"
}
]
},
{
"metadata": {},
"cell_type": "markdown",
"source": "### Show the model diagram:"
},
{
"metadata": {
"ExecuteTime": {
"end_time": "2018-11-04T23:42:26.816078Z",
"start_time": "2018-11-04T23:42:26.180356Z"
},
"trusted": true
},
"cell_type": "code",
"source": "# from keras.utils import plot_model\n# plot_model(model, to_file='autokeras_mnist_model.png')",
"execution_count": 1,
"outputs": []
},
{
"metadata": {
"trusted": true
},
"cell_type": "code",
"source": "",
"execution_count": null,
"outputs": []
}
],
"metadata": {
"_draft": {
"nbviewer_url": "https://gist.github.com/c2a31a283fac5c409799ad215b9445da"
},
"gist": {
"id": "c2a31a283fac5c409799ad215b9445da",
"data": {
"description": "Kaggle: Digit Recognizer, Auto-Keras",
"public": true
}
},
"kernelspec": {
"name": "py36",
"display_name": "py36",
"language": "python"
},
"language_info": {
"name": "python",
"version": "3.6.4",
"mimetype": "text/x-python",
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"pygments_lexer": "ipython3",
"nbconvert_exporter": "python",
"file_extension": ".py"
}
},
"nbformat": 4,
"nbformat_minor": 2
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment