VALIDATION_explainable-model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Explaination on Mitotic stage classification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References\n",
    "\n",
    "**Summary of Explainable methods**\n",
    "- A Brief History of Machine Learning Models Explainability. ([link](https://medium.com/@Zelros/a-brief-history-of-machine-learning-models-explainability-f1c3301be9dc))\n",
    "- https://notebooks.azure.com/denfromufa/projects/SHAP\n",
    "\n",
    "**LIME Algorithm**\n",
    "- USING LIME FOR IMAGE CLASSIFICATION EXPLAINABILITY \n",
    "    - [Part I](https://www.braincreators.com/2018/06/image-classification-explainability-lime/)\n",
    "- Python LIME library - A Tutorial ([link](https://github.com/marcotcr/lime/blob/master/doc/notebooks/Tutorial%20-%20Image%20Classification%20Keras.ipynb))\n",
    "- Explainable model. Youtube 李弘毅. ([link](https://www.youtube.com/watch?v=OjqIVSwly4k&t=1s), [link](https://www.youtube.com/watch?v=K1mWgthGS-A))\n",
    "\n",
    "**SHAP Alogrithm**\n",
    "- SHAP Library ([link](https://github.com/slundberg/shap))\n",
    "- SHAP Tutorial ([link](https://slundberg.github.io/shap/notebooks/deep_explainer/Front%20Page%20DeepExplainer%20MNIST%20Example.html))\n",
    "- SHAP Paper ([link](https://arxiv.org/pdf/1705.07874.pdf))\n",
    "\n",
    "**Compare SHAP with LIME**\n",
    "- SHAP and LIME Python Libraries. ([link](https://blog.dominodatalab.com/shap-lime-python-libraries-part-1-great-explainers-pros-cons/))\n",
    "\n",
    "**Class activation maps (CAM)**\n",
    "- CAM in keras. ([link](https://raghakot.github.io/keras-vis/visualizations/class_activation_maps/))\n",
    "- Saliency map. ([link](https://raghakot.github.io/keras-vis/visualizations/saliency/))\n",
    "\n",
    "**Saliency Map**\n",
    "- Notebook of neuroimage. ([link](https://colab.research.google.com/drive/1EgdnWZeNqmzqEmnSR9PUnYXlTjeu1wAU#scrollTo=RWyJftrVDY9-&forceEdit=true&offline=true&sandboxMode=true))\n",
    "\n",
    "## Data \n",
    "- Model and data: [link](https://drive.google.com/open?id=1CotCCn9tnp0LS0RmSIqfQhr1LrqJt4OZ&authuser=b03901045@g.ntu.edu.tw&usp=drive_fs)\n",
    "- This notebook(.ipynb): [link](https://drive.google.com/open?id=1JijG5trdClL3TQv0SWvWB30u-frVKy4Z&authuser=b03901045@g.ntu.edu.tw&usp=drive_fs)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os \n",
    "import pickle\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from skimage import io\n",
    "from skimage import data, color\n",
    "from skimage import measure\n",
    "from skimage.color import gray2rgb\n",
    "from keras.models import load_model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def setpath():\n",
    "    path = {\n",
    "            'model': '../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95.h5',\n",
    "            'test_x': '../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95_test_x.pickle',\n",
    "            'test_y': '../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95_test_y.pickle',\n",
    "            'train_x':'../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95_train_x.pickle',\n",
    "            'train_y':'../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95_train_y.pickle',\n",
    "            'le':'../Report_2_1_Automatic-Cell-Stage-Classification_nucleus_only/model/seq_acu_95_encoder.pickle',\n",
    "    }\n",
    "    \n",
    "    for key in path.keys(): # Check the paths are correct\n",
    "        if os.path.isfile(path[key]):\n",
    "            continue\n",
    "        elif os.path.isdir(path[key]):\n",
    "            continue\n",
    "        else:\n",
    "            raise ValueError\n",
    "    \n",
    "    return path\n",
    "\n",
    "def get_data(pickle_path):\n",
    "    assert os.path.isfile(pickle_path)\n",
    "    data= []\n",
    "    if '.pickle' in pickle_path:     \n",
    "        with open(pickle_path,\"rb\") as pickle_in:\n",
    "            data = pickle.load(pickle_in)\n",
    "        return data\n",
    "    else:\n",
    "        raise ValueError"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [],
   "source": [
    "path = setpath()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/stevenchiu/miniconda3/envs/keras/lib/python3.6/site-packages/sklearn/base.py:306: UserWarning: Trying to unpickle estimator LabelEncoder from version 0.20.2 when using version 0.21.1. This might lead to breaking code or invalid results. Use at your own risk.\n",
      "  UserWarning)\n"
     ]
    }
   ],
   "source": [
    "# load model\n",
    "model = load_model(path['model'])\n",
    "\n",
    "# Load data\n",
    "test_x = get_data(path['test_x'])\n",
    "test_y = get_data(path['test_y'])\n",
    "le = get_data(path['le'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3726, 1, 200, 230)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import shap\n",
    "import numpy as np\n",
    "\n",
    "# select a set of background examples to take an expectation over\n",
    "background = test_x[np.random.choice(test_x.shape[0], 100, replace=False)]\n",
    "\n",
    "# explain predictions of the model on three images\n",
    "e = shap.DeepExplainer(model, background)\n",
    "# ...or pass tensors directly\n",
    "# e = shap.DeepExplainer((model.layers[0].input, model.layers[-1].output), background)\n",
    "shap_values = e.shap_values(test_x[1:5])\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SHAP Map\n",
    "\n",
    "Should be like in [SHAP Library](https://github.com/slundberg/shap)\n",
    "<img src=\"https://raw.githubusercontent.com/slundberg/shap/master/docs/artwork/mnist_image_plot.png\">\n",
    "\n",
    "However, the results shows that the model focuses on the edge.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1296x900 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the feature attributions\n",
    "shap.image_plot(shap_values, test_x[100:104,0,:,:])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confirm model test accuracy again"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3726/3726 [==============================] - 60s 16ms/step\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "[0.11311475236030816, 0.954106280097255]"
      ],
      "text/plain": [
       "[0.11311475236030816, 0.954106280097255]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Confirm the model test accuracy\n",
    "model.evaluate(test_x, test_y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## LIME "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [],
   "source": [
    "import lime\n",
    "from lime import lime_image\n",
    "from lime.wrappers.scikit_image import SegmentationAlgorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [],
   "source": [
    "explainer = lime_image.LimeImageExplainer(verbose = False)\n",
    "segmenter = SegmentationAlgorithm('quickshift', kernel_size=1, max_dist=200, ratio=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [],
   "source": [
    "test_x = get_data(path['test_x'])\n",
    "model = load_model(path['model'])\n",
    "le = get_data(path['le'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [],
   "source": [
    "preds = model.predict(test_x[0:1,:,:,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false,
    "inputHidden": false,
    "outputHidden": false
   },
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Error when checking input: expected conv2d_1_input to have shape (1, 200, 230) but got array with shape (200, 230, 3)",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-18-02b96a8de542>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m                                          \u001b[0mhide_color\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m                                          \u001b[0mnum_samples\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m10000\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m                                          segmentation_fn=segmenter)\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/lime/lime_image.py\u001b[0m in \u001b[0;36mexplain_instance\u001b[0;34m(self, image, classifier_fn, labels, hide_color, top_labels, num_features, num_samples, batch_size, segmentation_fn, distance_metric, model_regressor, random_seed)\u001b[0m\n\u001b[1;32m    192\u001b[0m         data, labels = self.data_labels(image, fudged_image, segments,\n\u001b[1;32m    193\u001b[0m                                         \u001b[0mclassifier_fn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_samples\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 194\u001b[0;31m                                         batch_size=batch_size)\n\u001b[0m\u001b[1;32m    195\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    196\u001b[0m         distances = sklearn.metrics.pairwise_distances(\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/lime/lime_image.py\u001b[0m in \u001b[0;36mdata_labels\u001b[0;34m(self, image, fudged_image, segments, classifier_fn, num_samples, batch_size)\u001b[0m\n\u001b[1;32m    253\u001b[0m             \u001b[0mimgs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtemp\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    254\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimgs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 255\u001b[0;31m                 \u001b[0mpreds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mclassifier_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimgs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    256\u001b[0m                 \u001b[0mlabels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpreds\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    257\u001b[0m                 \u001b[0mimgs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/keras/engine/sequential.py\u001b[0m in \u001b[0;36mpredict_proba\u001b[0;34m(self, x, batch_size, verbose)\u001b[0m\n\u001b[1;32m    243\u001b[0m             \u001b[0mA\u001b[0m \u001b[0mNumpy\u001b[0m \u001b[0marray\u001b[0m \u001b[0mof\u001b[0m \u001b[0mprobability\u001b[0m \u001b[0mpredictions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    244\u001b[0m         \"\"\"\n\u001b[0;32m--> 245\u001b[0;31m         \u001b[0mpreds\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    246\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mpreds\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m<\u001b[0m \u001b[0;36m0.\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mpreds\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1.\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    247\u001b[0m             warnings.warn('Network returning invalid probability values. '\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mpredict\u001b[0;34m(self, x, batch_size, verbose, steps)\u001b[0m\n\u001b[1;32m   1147\u001b[0m                              'argument.')\n\u001b[1;32m   1148\u001b[0m         \u001b[0;31m# Validate user data.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1149\u001b[0;31m         \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_standardize_user_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1150\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstateful\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1151\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0mbatch_size\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0mbatch_size\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36m_standardize_user_data\u001b[0;34m(self, x, y, sample_weight, class_weight, check_array_lengths, batch_size)\u001b[0m\n\u001b[1;32m    749\u001b[0m             \u001b[0mfeed_input_shapes\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    750\u001b[0m             \u001b[0mcheck_batch_axis\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m  \u001b[0;31m# Don't enforce the batch size.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 751\u001b[0;31m             exception_prefix='input')\n\u001b[0m\u001b[1;32m    752\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    753\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0my\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/miniconda3/envs/keras/lib/python3.6/site-packages/keras/engine/training_utils.py\u001b[0m in \u001b[0;36mstandardize_input_data\u001b[0;34m(data, names, shapes, check_batch_axis, exception_prefix)\u001b[0m\n\u001b[1;32m    136\u001b[0m                             \u001b[0;34m': expected '\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mnames\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m' to have shape '\u001b[0m \u001b[0;34m+\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    137\u001b[0m                             \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m' but got array with shape '\u001b[0m \u001b[0;34m+\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 138\u001b[0;31m                             str(data_shape))\n\u001b[0m\u001b[1;32m    139\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    140\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Error when checking input: expected conv2d_1_input to have shape (1, 200, 230) but got array with shape (200, 230, 3)"
     ]
    }
   ],
   "source": [
    "image = test_x[0,0,:,:]\n",
    "explanation = explainer.explain_instance(image, \n",
    "                                         classifier_fn = model.predict_proba, \n",
    "                                         top_labels=5, \n",
    "                                         hide_color=0, \n",
    "                                         num_samples=10000, \n",
    "                                         segmentation_fn=segmenter)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model2 = model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<keras.layers.core.Dense at 0x11f251f60>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2.layers.pop()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "'Sequential' object does not support item assignment",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-19-07388b3f3522>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0minput_layer\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mInputLayer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput_shape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m200\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m230\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"input\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mmodel2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0mmodel2\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minput_layer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: 'Sequential' object does not support item assignment"
     ]
    }
   ],
   "source": [
    "from keras.engine import InputLayer\n",
    "\n",
    "input_layer = InputLayer(input_shape=(None,200, 230, 3), name=\"input\")\n",
    "model2[0] = input_layer\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "from numpy import moveaxis\n",
    "from tensorflow.image import grayscale_to_rgb\n",
    "test_x3 = moveaxis(test_x, 1, 3)\n",
    "test_x3 = grayscale_to_rgb(test_x3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([Dimension(3726), Dimension(200), Dimension(230), Dimension(3)])"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_x3.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.imshow(tf.Session().run(test_x3[0,:,:,:]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Saliency map - Keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from vis.visualization import visualize_saliency"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "visualize_saliency(model=model, \n",
    "                   layer_idx=-3, \n",
    "                   filter_indices=None, \n",
    "                   seed_input=test_x[0:1,:,:,:]\n",
    ")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install git+https://github.com/raghakot/keras-vis.git"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernel_info": {
   "name": "keras"
  },
  "kernelspec": {
   "display_name": "keras",
   "language": "python",
   "name": "keras"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  },
  "nteract": {
   "version": "0.12.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}