{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Census income classification with Keras\n", "\n", "To download a copy of this notebook visit [github](https://github.com/shap/shap/tree/master/notebooks).\n", "\n", "This notebook showcases the use of kernel and coalition explainer for explaining neural networks. " ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "c:\\programming\\shap\\.venv\\lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n", " from .autonotebook import tqdm as notebook_tqdm\n" ] }, { "data": { "text/html": [ "
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import matplotlib.pyplot as plt\n", "from keras.layers import Dense, Dropout, Embedding, Flatten, Input, concatenate\n", "from keras.models import Model\n", "from sklearn.model_selection import train_test_split\n", "\n", "import shap\n", "\n", "# print the JS visualization code to the notebook\n", "shap.initjs()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Load dataset" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "X, y = shap.datasets.adult()\n", "X_display, y_display = shap.datasets.adult(display=True)\n", "\n", "# normalize data (this is important for model convergence)\n", "dtypes = list(zip(X.dtypes.index, map(str, X.dtypes)))\n", "for k, dtype in dtypes:\n", " if dtype == \"float32\":\n", " X[k] -= X[k].mean()\n", " X[k] /= X[k].std()\n", "\n", "X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.2, random_state=7)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Train Keras model" ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "scrolled": true }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 14ms/step - loss: 3.1297 - val_loss: 0.8234\n", "Epoch 2/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 5ms/step - loss: 1.9760 - val_loss: 1.0105\n", "Epoch 3/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.6813 - val_loss: 0.5041\n", "Epoch 4/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.5226 - val_loss: 0.5412\n", "Epoch 5/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.4729 - val_loss: 0.8485\n", "Epoch 6/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.5344 - val_loss: 0.4860\n", "Epoch 7/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.3546 - val_loss: 0.4602\n", "Epoch 8/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.2509 - val_loss: 0.4429\n", "Epoch 9/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.2643 - val_loss: 0.4513\n", "Epoch 10/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.1453 - val_loss: 0.4645\n", "Epoch 11/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.1067 - val_loss: 0.4369\n", "Epoch 12/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.1187 - val_loss: 0.4549\n", "Epoch 13/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.0626 - val_loss: 0.4600\n", "Epoch 14/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 1.0424 - val_loss: 0.4231\n", "Epoch 15/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.9474 - val_loss: 0.4182\n", "Epoch 16/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.8717 - val_loss: 0.4932\n", "Epoch 17/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - loss: 0.9838 - val_loss: 0.4509\n", "Epoch 18/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.8657 - val_loss: 0.4175\n", "Epoch 19/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - loss: 0.9054 - val_loss: 0.4238\n", "Epoch 20/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.8049 - val_loss: 0.4411\n", "Epoch 21/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.7607 - val_loss: 0.4126\n", "Epoch 22/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6943 - val_loss: 0.4062\n", "Epoch 23/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6739 - val_loss: 0.4163\n", "Epoch 24/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6874 - val_loss: 0.3914\n", "Epoch 25/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6113 - val_loss: 0.3714\n", "Epoch 26/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.5662 - val_loss: 0.3684\n", "Epoch 27/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 8ms/step - loss: 0.6947 - val_loss: 0.6365\n", "Epoch 28/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.8330 - val_loss: 0.4122\n", "Epoch 29/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6455 - val_loss: 0.4116\n", "Epoch 30/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 10ms/step - loss: 0.5741 - val_loss: 0.3864\n", "Epoch 31/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.6159 - val_loss: 0.3896\n", "Epoch 32/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - loss: 0.5781 - val_loss: 0.3807\n", "Epoch 33/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - loss: 0.5281 - val_loss: 0.3799\n", "Epoch 34/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.5337 - val_loss: 0.3780\n", "Epoch 35/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.5242 - val_loss: 0.3794\n", "Epoch 36/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.5320 - val_loss: 0.3818\n", "Epoch 37/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.5084 - val_loss: 0.3808\n", "Epoch 38/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4883 - val_loss: 0.3776\n", "Epoch 39/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4777 - val_loss: 0.3728\n", "Epoch 40/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4945 - val_loss: 0.3745\n", "Epoch 41/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4776 - val_loss: 0.3721\n", "Epoch 42/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4781 - val_loss: 0.3694\n", "Epoch 43/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - loss: 0.4621 - val_loss: 0.3690\n", "Epoch 44/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - loss: 0.4481 - val_loss: 0.3721\n", "Epoch 45/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - loss: 0.4557 - val_loss: 0.3688\n", "Epoch 46/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4404 - val_loss: 0.3702\n", "Epoch 47/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4545 - val_loss: 0.3753\n", "Epoch 48/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - loss: 0.4627 - val_loss: 0.3737\n", "Epoch 49/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4358 - val_loss: 0.3740\n", "Epoch 50/50\n", "\u001b[1m51/51\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 4ms/step - loss: 0.4449 - val_loss: 0.3745\n" ] }, { "data": { "text/plain": [ "" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# build model\n", "input_els = []\n", "encoded_els = []\n", "for k, dtype in dtypes:\n", " input_els.append(Input(shape=(1,)))\n", " if dtype == \"int8\":\n", " e = Flatten()(Embedding(X_train[k].max() + 1, 1)(input_els[-1]))\n", " else:\n", " e = input_els[-1]\n", " encoded_els.append(e)\n", "encoded_els = concatenate(encoded_els)\n", "layer1 = Dropout(0.5)(Dense(100, activation=\"relu\")(encoded_els))\n", "out = Dense(1)(layer1)\n", "\n", "# train model\n", "regression = Model(inputs=input_els, outputs=[out])\n", "regression.compile(optimizer=\"adam\", loss=\"binary_crossentropy\")\n", "regression.fit(\n", " [X_train[k].values for k, t in dtypes],\n", " y_train,\n", " epochs=50,\n", " batch_size=512,\n", " shuffle=True,\n", " validation_data=([X_valid[k].values for k, t in dtypes], y_valid),\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Explain predictions\n", "\n", "Here we take the Keras model trained above and explain why it makes different predictions for different individuals. SHAP expects model functions to take a 2D numpy array as input, so we define a wrapper function around the original Keras predict function." ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "def f(X):\n", " return regression.predict([X[:, i] for i in range(X.shape[1])]).flatten()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Explain a single prediction\n", "\n", "Here we use a selection of 50 samples from the dataset to represent \"typical\" feature values, and then use 500 perturbation samples to estimate the SHAP values for a given prediction. Note that this requires 500 * 50 evaluations of the model." ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/plain": [ "Age -0.042641\n", "Workclass 4.000000\n", "Education-Num -0.420053\n", "Marital Status 0.000000\n", "Occupation 12.000000\n", "Relationship 0.000000\n", "Race 4.000000\n", "Sex 1.000000\n", "Capital Gain -0.145918\n", "Capital Loss -0.216656\n", "Hours per week 3.204112\n", "Country 39.000000\n", "Name: 299, dtype: float64" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X.iloc[299, :]" ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
AgeWorkclassEducation-NumMarital StatusOccupationRelationshipRaceSexCapital GainCapital LossHours per weekCountry
299-0.0426414-0.420053012041-0.145918-0.2166563.20411239
\n", "
" ], "text/plain": [ " Age Workclass Education-Num Marital Status Occupation \\\n", "299 -0.042641 4 -0.420053 0 12 \n", "\n", " Relationship Race Sex Capital Gain Capital Loss Hours per week \\\n", "299 0 4 1 -0.145918 -0.216656 3.204112 \n", "\n", " Country \n", "299 39 " ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X.iloc[299:300, :]" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 244ms/step\n" ] }, { "data": { "text/plain": [ "array([0.23898253], dtype=float32)" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import numpy as np\n", "\n", "f(np.array(X.iloc[299:300, :]))" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 48ms/step \n", "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 63ms/step\n", "\u001b[1m782/782\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 3ms/step\n" ] }, { "data": { "text/html": [ "\n", "
\n", "
\n", " Visualization omitted, Javascript library not loaded!
\n", " Have you run `initjs()` in this notebook? If this notebook was from another\n", " user you must also trust this notebook (File -> Trust notebook). If you are viewing\n", " this notebook on github the Javascript has been stripped for security. If you are using\n", " JupyterLab this error is because a JupyterLab extension has not yet been written.\n", "
\n", " " ], "text/plain": [ "" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "kernel_explainer = shap.KernelExplainer(f, X.iloc[:50, :])\n", "shap_values = kernel_explainer.shap_values(X.iloc[299, :], nsamples=500)\n", "\n", "shap.force_plot(kernel_explainer.expected_value, shap_values, X_display.iloc[299, :])" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ " 0%| | 0/1 [00:00" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots()\n", "shap.plots.waterfall(shap_values[0], show=False)\n", "plt.title(\"PartitionExplainer for instance 0 with the KernelExplainer\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plt.subplots()\n", "shap.plots.waterfall(winter_values[0], show=False)\n", "plt.title(\"PartitionExplainer for instance 0 with the CoalitionExplainer\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "\n", "
\n", "
\n", " Visualization omitted, Javascript library not loaded!
\n", " Have you run `initjs()` in this notebook? If this notebook was from another\n", " user you must also trust this notebook (File -> Trust notebook). If you are viewing\n", " this notebook on github the Javascript has been stripped for security. If you are using\n", " JupyterLab this error is because a JupyterLab extension has not yet been written.\n", "
\n", " " ], "text/plain": [ "" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "shap.initjs()\n", "shap.force_plot(winter_values.base_values, winter_values.values, X_display.iloc[299:300, :])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Explain many predictions\n", "\n", "Here we repeat the above explanation process for 20 individuals. Since we are using a sampling based approximation each explanation can take a couple seconds depending on your machine setup." ] }, { "cell_type": "code", "execution_count": 15, "metadata": { "scrolled": true }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 43ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 61ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 44ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 46ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 53ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 36ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 16ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 40ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 20ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 23ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step \n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 24ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 48ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 42ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 36ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 91ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 39ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 39ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 52ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 44ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 40ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 24ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 24ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 40ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 27ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n", "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 16ms/step\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "CoalitionExplainer explainer: 10%|█ | 1/10 [00:00\n", "
\n", " Visualization omitted, Javascript library not loaded!
\n", " Have you run `initjs()` in this notebook? If this notebook was from another\n", " user you must also trust this notebook (File -> Trust notebook). If you are viewing\n", " this notebook on github the Javascript has been stripped for security. If you are using\n", " JupyterLab this error is because a JupyterLab extension has not yet been written.\n", "
\n", " " ], "text/plain": [ "" ] }, "execution_count": 17, "metadata": {}, "output_type": "execute_result" } ], "source": [ "shap.force_plot(kernel_explainer.expected_value, shap_values50, X_display.iloc[320:330, :])" ] }, { "cell_type": "code", "execution_count": 18, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/html": [ "\n", "
\n", "
\n", " Visualization omitted, Javascript library not loaded!
\n", " Have you run `initjs()` in this notebook? If this notebook was from another\n", " user you must also trust this notebook (File -> Trust notebook). If you are viewing\n", " this notebook on github the Javascript has been stripped for security. If you are using\n", " JupyterLab this error is because a JupyterLab extension has not yet been written.\n", "
\n", " " ], "text/plain": [ "" ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" } ], "source": [ "shap.force_plot(winter_values50.base_values, winter_values50.values, X_display.iloc[320:330, :])" ] } ], "metadata": { "kernelspec": { "display_name": ".venv", "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": "3.10.11" } }, "nbformat": 4, "nbformat_minor": 4 }