{ "cells": [ { "cell_type": "code", "execution_count": 3, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "import matplotlib\n", "import matplotlib.pyplot as plt\n", "matplotlib.pyplot.gray()\n", "\n", "%matplotlib inline\n", "\n", "import numpy as np\n", "\n", "import theano\n", "import theano.tensor as T\n", "rng = np.random.RandomState(42)\n", "\n", "import sys\n", "import time\n", "\n", "theano.config.floatX = 'float32'" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# ComputeFest 2016 - Deep Learning Workshop Part 1 - Fully Connected Networks\n", "\n", "Today we will focus on fully connected networks. We follow the traditional scheme and tackle the traditional task of hand written digit classification. First we use a shallow classifier, simple logistic regression, to get started. Then we add layers to our model to make it more powerful. But, first let's start by loading and looking at the data. \n", "We will make heavy use of the resources in the Theano deep learning tutorial. We have it integrated into our git repository as a submodule. You can clone the git repository by doing the following steps:\n", "\n", "`git clone --recursive https://github.com/vkaynig/IACS_ComputeFest_DeepLearning.git`\n", "\n", "If you already cloned the repository, but the `DeepLearningTutorial` folder is empty, you need to checkout the submodule. Make sure you are in the folder `IACS_ComputeFest_DeepLearning` and then execute the following command:\n", "\n", "`git submodule update --init --recursive`\n", "\n", "\n", "Now we have to add this directory to the PythonPath. Depending on the location of your git repository you might have to change this path.\n", "\n", "\n", "\n" ] }, { "cell_type": "code", "execution_count": 20, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/plain": [ "['',\n", " './DeepLearningTutorials/code',\n", " './DeepLearningTutorials/code',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\pyautodiff-0.0.1-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\autodiff-0.4-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\meta-_-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\distribute-0.6.19-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\pycuda-2012.1-py2.7-win32.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\decorator-3.4.0-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\pytest-2.6.4-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\pytools-2014.3.5-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\colorama-0.3.3-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\py-1.4.26-py2.7.egg',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\appdirs-1.4.0-py2.7.egg',\n", " 'C:\\\\Users\\\\vkaynig\\\\Documents\\\\AM207\\\\webpage_pelican_setup',\n", " 'C:\\\\Windows\\\\SYSTEM32\\\\python27.zip',\n", " 'C:\\\\Python27\\\\DLLs',\n", " 'C:\\\\Python27\\\\lib',\n", " 'C:\\\\Python27\\\\lib\\\\plat-win',\n", " 'C:\\\\Python27\\\\lib\\\\lib-tk',\n", " 'C:\\\\Python27',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages',\n", " 'C:\\\\Python27\\\\lib\\\\site-packages\\\\IPython\\\\extensions']" ] }, "execution_count": 20, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sys.path.insert(1,'./DeepLearningTutorials/code')\n", "sys.path" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## The Task: Handwritten Digit Recognition \n", "Classifying handwritten digits is like the 'hello world' for deep learning methods. It is how everyone normally starts to get used to the models and how to best train them. Interstingly though, it is not just for beginners. There are still papers published nowadays that use MNIST to evaluate new methods. There is some debate on whether this is still useful or not, but we won't care about this discussion for now. \n", "\n", "MNIST consists of 70 000 small image patches, each showing a handwritten digit character in white on a black background. There are 10 different classes (the digits from 0-9). Let's load the data and have a look." ] }, { "cell_type": "code", "execution_count": 21, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "... loading data\n" ] } ], "source": [ "from logistic_sgd import load_data\n", "dataset='mnist.pkl.gz'\n", "\n", "## If not already cached this function actually downloads the data\n", "datasets = load_data(dataset)\n", "\n", "train_set_x, train_set_y = datasets[0]\n", "valid_set_x, valid_set_y = datasets[1]\n", "test_set_x, test_set_y = datasets[2]\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Theano Shared Variables" ] }, { "cell_type": "code", "execution_count": 22, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n" ] } ], "source": [ "print train_set_x\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now that looks a bit odd, why can't we see the values of our data matrix? The `load_data` function provided by the deep learning tutorial downloads the data and loads it into a shared variable. Shared variables are an important and genious concept in Theano. What it means is that the data can be shared between the CPU and the GPU without you having to write a single line of code! You basically do not have to care at all where your data lives, unless you are running out of memory, which is another story we can talk about later. For now let us only worry about how we can access the actual values of the data matrix:" ] }, { "cell_type": "code", "execution_count": 23, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[[ 0. 0. 0. ..., 0. 0. 0.]\n", " [ 0. 0. 0. ..., 0. 0. 0.]\n", " [ 0. 0. 0. ..., 0. 0. 0.]\n", " ..., \n", " [ 0. 0. 0. ..., 0. 0. 0.]\n", " [ 0. 0. 0. ..., 0. 0. 0.]\n", " [ 0. 0. 0. ..., 0. 0. 0.]]\n" ] } ], "source": [ "print train_set_x.get_value()" ] }, { "cell_type": "code", "execution_count": 24, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Size of the training data matrix: (50000, 784)\n" ] } ], "source": [ "print \"Size of the training data matrix: \", train_set_x.get_value().shape" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The data is stored with one training sample per row. The original image patches are $28 \\times 28$ pixels, hence we have $28 \\cdot 28 = 784$ feature columns in our data matrix. We can reshape each row of our data matrix back to the original $28 \\times 28$ image patch to visualize what the data looks like." ] }, { "cell_type": "code", "execution_count": 25, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACFCAYAAAB/o8JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from utils import tile_raster_images\n", "\n", "samples = tile_raster_images(train_set_x.get_value(), img_shape=(28,28), \n", " tile_shape=(3,10), tile_spacing=(0, 0),\n", " scale_rows_to_unit_interval=True,\n", " output_pixel_vals=True)\n", "\n", "plt.imshow(samples)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "So each image patch contains a hand written digit in white on a black background. The strokes vary in thickness and there is quite some variability in the hand writing. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The labels of the different image patches are stored in our variable `train_set_y`. When we print this variable it looks different than for our training data, and there is no `get_value()` function that we can call to access the labels." ] }, { "cell_type": "code", "execution_count": 26, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Elemwise{Cast{int32}}.0\n" ] } ], "source": [ "print train_set_y" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The reason is that the `load_data` function casted our labels to integers. Why would it do that? The GPU can only store float variables, so all shared variables are stored as float. Our labels are integers though and later on we want to use them as indices, which requires them to be integers. As a workaround `load_data` first stores the labels as floats and then casts them to integer. As you can see from the `print` statement, Theano does not execute the casting right away. Instead it keeps a record of all the computations that are to be performed on the shared variable in a computational graph. This is another genious concept of Theano, because when these graphs get more complicated, it optimizes the whole graph for computational efficiency and numerical stability. If you want to know more about the computational graph structures and their optimization in Theano you can look [here](http://deeplearning.net/software/theano/tutorial/symbolic_graphs.html).\n", "\n", "To access the labels we need to tell Theano to exectue all computations from the corresponding graph. We do this by calling `eval()`:" ] }, { "cell_type": "code", "execution_count": 27, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "[5 0 4 1 9 2 1 3 1 4]\n" ] } ], "source": [ "print train_set_y[:10].eval()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Exercise: Data Normalization\n", "\n", "Deep learning algorithms typically work best when you normalize your data. A standard procedure is to normalize the mean and the standard deviation of each feature column.\n", "Use the training data to estimate the mean normalization, and then use these values to normralize the training, the validation and the test data. Then use the `tile_raster_images` fuction as above to visualize some normalized training patches. Hint: you can basically deal with the shared variables as if they were numpy arrays. " ] }, { "cell_type": "code", "execution_count": 28, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# estimate the mean and std dev from the training data\n", "# then use these estimates to normalize the data" ] }, { "cell_type": "code", "execution_count": 29, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAACFCAYAAAB/o8JHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nq5wAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# visualize the result\n", "samples = tile_raster_images(train_set_x.eval(), img_shape=(28,28), \n", " tile_shape=(3,10), tile_spacing=(0, 0),\n", " scale_rows_to_unit_interval=True,\n", " output_pixel_vals=True)\n", "\n", "plt.imshow(samples)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "We have to do quite some computations using our training data, so we can just compute the whole graph on the shared variables now and re-assign our shared variables to the new values. " ] }, { "cell_type": "code", "execution_count": 30, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/plain": [ "Elemwise{true_div,no_inplace}.0" ] }, "execution_count": 30, "metadata": {}, "output_type": "execute_result" } ], "source": [ "train_set_x" ] }, { "cell_type": "code", "execution_count": 31, "metadata": { "collapsed": false }, "outputs": [], "source": [ "train_set_x = theano.shared(train_set_x.eval())\n", "valid_set_x = theano.shared(valid_set_x.eval())\n", "test_set_x = theano.shared(test_set_x.eval())" ] }, { "cell_type": "code", "execution_count": 32, "metadata": { "collapsed": false }, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 32, "metadata": {}, "output_type": "execute_result" } ], "source": [ "train_set_x" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Logistic Regression\n", "\n", "We have seen the basics of logistic regression in the theory part of the workshop. Now it is time to actually apply the method to some data using Theano. We start with this `shallow` classifier and then will add more layers later on. \n", "\n", "From the slides we remember that the logistic regression model has the weight parameters W and the biases b. We need to intialize these as Theano variables. In addition we create placeholders for our data by declaring a data matrix `x`, and a label vector `y`. Note that we do not assign our actual training data to the variables `x` and `y`. Theano can create the whole computational graph for `x` and `y` without knowing the actual values. " ] }, { "cell_type": "code", "execution_count": 33, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# Step 1. Declare Theano variables\n", "x = T.fmatrix()\n", "y = T.ivector()\n", "# size of our image patches, and number of output units (classes)\n", "W = theano.shared(value=np.zeros((28*28, 10), \n", " dtype=theano.config.floatX),\n", " name='W',\n", " borrow=True\n", " )\n", "b = theano.shared(value=np.zeros((10,), # number of output units (classes)\n", " dtype=theano.config.floatX),\n", " name='b',\n", " borrow=True\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we create all necessary functions for the computation. We need the likelihood to make predictions, the negative log likelihood as our cost function to optimize, the actual conversion of the class probabilities to class labels, and the classification error. " ] }, { "cell_type": "code", "execution_count": 34, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# Step 2. Construct Theano expression graph\n", "\n", "p_y_given_x = T.nnet.softmax(T.dot(x, W) + b)\n", "y_pred = T.argmax(p_y_given_x, axis=1)\n", "\n", "error = T.mean(T.neq(y_pred, y))\n", "cost = -T.mean(T.log(p_y_given_x)[T.arange(y.shape[0]), y])\n", "#cost = -T.mean(T.log(p_y_given_x)[:, y])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now comes the really cool part. We do not have to bother to compute the gradient ourselves! Theano is capable of computing the gradient for us. We just call the `T.grad()` function on our cost function and specify with respect to which parameter the derivative should be computed. Then we can use the computed gradient to perform an iteration of gradient decent. We just update the parameters by taking a small step into the right direction on the cost function surface. The size of the step is specified by the learning rate. If you are interested in the best performance for your network, the learning rate is a very important parameter to tune. Typically smaller values work better, but take longer. We will talk more about this in the tips and tricks section of the workshop. " ] }, { "cell_type": "code", "execution_count": 35, "metadata": { "collapsed": false }, "outputs": [], "source": [ "g_W = T.grad(cost=cost, wrt=W)\n", "g_b = T.grad(cost=cost, wrt=b)\n", "\n", "learning_rate = 0.1\n", "\n", "updates = [(W, W - learning_rate * g_W),\n", " (b, b - learning_rate * g_b)]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Theano Functions\n", "\n", "Now we are ready to compile everything we defined so far into a callable function to update our parameters and train our logistic regression classifier. How do we do this? Above we defined our placeholders `x` and `y` for the training data. And we defined `cost` as a function (actually its a Theano computational graph) depending on these placeholders. We now use the `function` interface provided by Theano to create a function that we can call directly on our training data, e.g., binding the training data to the placeholders `x` and `y`. To create this function we need to specify the `inputs` and the `outputs`. In addition the `function` interface has a convenient option to apply our gradient decent updates to the parameters. All we do is pass along the `updates` list we defined in the cell above. " ] }, { "cell_type": "code", "execution_count": null, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# Step 3. Compile expressions to functions\n", "train = theano.function(inputs=[x, y],\n", " outputs=cost,\n", " updates=updates)\n", "\n", "validate = theano.function(inputs=[x, y],\n", " outputs=error)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As you can see the inputs are our placeholders `x` and `y`. When we call our function `validate` it will ask for two parameters. When we call it with our training samples and their labels, the matrix `train_set_x` gets bound to `x` and `train_set_y` to `y`. The function then traverses the computational graph specified by the outputs, in this case `error`. Thus, the validation function will give us our training error when we call it on the training data. We can also call it on the test data to obtain the test error. \n", "\n", "Now we really have everything in place to train our classifier!" ] }, { "cell_type": "code", "execution_count": 20, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iteration: 0\n", "Training error, validation error: 0.26888 0.2467\n", "Iteration: 2\n", "Training error, validation error: 0.21946 0.2003\n", "Iteration: 4\n", "Training error, validation error: 0.18914 0.171\n", "Iteration: 6\n", "Training error, validation error: 0.1698 0.153\n", "Iteration: 8\n", "Training error, validation error: 0.15726 0.1396\n", "The code ran for 5.713457 seconds\n" ] } ], "source": [ "# Step 4. Perform computation\n", "train_monitor = []\n", "val_monitor = []\n", "\n", "n_epochs = 10\n", "\n", "start_time = time.clock()\n", "for epoch in range(n_epochs):\n", " loss = train(train_set_x.eval(), train_set_y.eval())\n", " train_monitor.append(validate(train_set_x.eval(), train_set_y.eval()))\n", " val_monitor.append(validate(valid_set_x.eval(), valid_set_y.eval()))\n", " if epoch%2 == 0:\n", " print \"Iteration: \", epoch\n", " print \"Training error, validation error: \", train_monitor[-1], val_monitor[-1]\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "That is ok, but could be faster. All those graph evaluations cost time. Unfortunately Theano functions do not allow shared variables as direct input (more about this soon). So let's at least just do the evaluation once and save it in a normal non-shared variable that we can call our functions with. " ] }, { "cell_type": "code", "execution_count": 21, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iteration: 0\n", "Training error, validation error: 0.26888 0.2467\n", "Iteration: 2\n", "Training error, validation error: 0.21946 0.2003\n", "Iteration: 4\n", "Training error, validation error: 0.18914 0.171\n", "Iteration: 6\n", "Training error, validation error: 0.1698 0.153\n", "Iteration: 8\n", "Training error, validation error: 0.15726 0.1396\n", "Iteration: 10\n", "Training error, validation error: 0.14868 0.1314\n", "Iteration: 12\n", "Training error, validation error: 0.14178 0.1263\n", "Iteration: 14\n", "Training error, validation error: 0.13726 0.1226\n", "Iteration: 16\n", "Training error, validation error: 0.13316 0.1192\n", "Iteration: 18\n", "Training error, validation error: 0.12968 0.1165\n", "The code ran for 11.016269 seconds\n" ] } ], "source": [ "# Step 4. Perform computation\n", "# reset W and b to zero\n", "W.set_value(np.zeros((28*28, 10),\n", " dtype=theano.config.floatX)) \n", "b.set_value(np.zeros((10,),\n", " dtype=theano.config.floatX))\n", "\n", "train_monitor = []\n", "val_monitor = []\n", "\n", "n_epochs = 20\n", "\n", "trainx = train_set_x.eval()\n", "trainy = train_set_y.eval()\n", "valx = valid_set_x.eval()\n", "valy = valid_set_y.eval()\n", "testx = test_set_x.eval()\n", "testy = test_set_y.eval()\n", "\n", "start_time = time.clock()\n", "for epoch in range(n_epochs):\n", " loss = train(trainx, trainy)\n", " train_monitor.append(validate(trainx, trainy))\n", " val_monitor.append(validate(valx,valy))\n", " if epoch%2 == 0:\n", " print \"Iteration: \", epoch\n", " print \"Training error, validation error: \", train_monitor[-1], val_monitor[-1]\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "That looks better. As we already know the number of epochs we are going to train for, we could also fix the size of our monitor variables beforehand and then fill them during training. " ] }, { "cell_type": "code", "execution_count": 22, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAEKCAYAAAAMzhLIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS9/gaiEptPAAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(train_monitor, c='r')\n", "plt.plot(val_monitor, c='b')\n", "plt.xlabel('Number of Epochs')\n", "plt.ylabel('Missclassification')\n", "plt.show()\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Exercise Theano Functions\n", "\n", "Write a Theano function that computes the actual predictions for a data set." ] }, { "cell_type": "code", "execution_count": null, "metadata": { "collapsed": false }, "outputs": [], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Theano Functions with Shared Variable Input\n", "### Introducing batch training\n", "\n", "It seems quite limiting that we are not allowed to just use our shared variables directly as input in the defined Theano functions. After all the whole idea behind the shared variables is to have our data potentially on the GPU, and we don't want to copy it back and forth all the time. There is another mechanism in Theano functions to achieve shared variable input: the `givens` parameter. When I first enocuntered it, I found it very confusing, which is why we are going to look at the conversion of our functions in two steps: First we are going to redefine the training and validation function to work with shared variables, and then we are going to introduce batch training for stochastic gradient decent. For batch training we want to use only part of the data to already compute an update. An epoch then consists of several runs over all batches, until all training data has been seen once, then we start all over again like before. \n", "\n", "First let's look at our definition of the train function and what they look like for shared variables:" ] }, { "cell_type": "code", "execution_count": 25, "metadata": { "collapsed": false }, "outputs": [], "source": [ "train = theano.function(inputs=[],\n", " outputs=[cost, error],\n", " updates=updates,\n", " givens={x: train_set_x, y: train_set_y}\n", " )\n", "\n", "validate = theano.function(inputs=[],\n", " outputs=error,\n", " givens={x: valid_set_x, y: valid_set_y}\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Remember that we defined our `cost` and `error` dependent on the placeholder variables `x` and `y`. Now we are using thing `givens` parameter to substitute `x` and `y` with our training and validation data. This is all the `givens` parameter does. It can be used to replace any symbolic variable, not just shared ones. You can even replace constants or whole expressions. Careful though: Expressions should not be co-dependent, as the order of execution is not defined, so the replacements have to work in any order. \n", "There is a small price attached to this way of defining our functions. The data is now hardcoded into the definition. This means we either have to define a training_error function, or we can add the error to the output of our training function. " ] }, { "cell_type": "code", "execution_count": 26, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iteration: 0\n", "Training error, validation error: 0.90136 0.9009\n", "Iteration: 10\n", "Training error, validation error: 0.1526 0.1348\n", "Iteration: 20\n", "Training error, validation error: 0.12796 0.1154\n", "Iteration: 30\n", "Training error, validation error: 0.118 0.1062\n", "Iteration: 40\n", "Training error, validation error: 0.11202 0.101\n", "The code ran for 3.131449 seconds\n" ] } ], "source": [ "# Step 4. Perform computation\n", "# reset W and b to zero\n", "W.set_value(np.zeros((28*28, 10),\n", " dtype=theano.config.floatX)) \n", "b.set_value(np.zeros((10,),\n", " dtype=theano.config.floatX))\n", "\n", "train_monitor = []\n", "val_monitor = []\n", "\n", "n_epochs = 50\n", "\n", "start_time = time.clock()\n", "for epoch in range(n_epochs):\n", " val_monitor.append(validate())\n", " loss, err = train()\n", " train_monitor.append(err)\n", " if epoch%10 == 0:\n", " print \"Iteration: \", epoch\n", " print \"Training error, validation error: \", train_monitor[-1], val_monitor[-1]\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "And we got even faster, nice! Something weird is going on though. Note that the training error of the first iteration is really high and different from what we have seen before. The reason is that the error is computed by the training function before the actual update. To get a fair comparison of train and validation error, we can just call the `validate` function before we do the training step. This means we are missing the final performance of our classifier, so you would have to run an additional evaluation round in the end. Preferably with entirely new test data that you haven't looked at before. \n", "\n", "Let's go further and introduce batch training. The idea is to not estimate the gradient each time over the whole dataset. The gradient estimation typically is already stable after averaging over a few training samples, so using the whole dataset is a waste of time, and it can actually help to have noisy gradient estimations. The reason is that noisy estimations add an exploration factor to our optimization scheme. We basically jump around more, but might find a good local minimum of our energy function on one of our detours. \n", "\n", "To define a training function using batches, we introduce a variable for the batch size and resulting number of batches. Both variables are fixed and defined outside of our function. We also introduce a new input argument `index`. The `index` refers to the index of the specific batch of training data for the next iteration. Note that we now do several updates per epoch. If we have 100 training batches, one epoch now corresponds to 100 updates, instead of one update as before. " ] }, { "cell_type": "code", "execution_count": 28, "metadata": { "collapsed": false }, "outputs": [], "source": [ "batch_size = 20\n", "n_train_batches = train_set_x.shape[0].eval() / batch_size\n", "index = T.iscalar() \n", "\n", "train_batch = theano.function(inputs=[index],\n", " outputs=cost, \n", " updates=updates,\n", " givens={\n", " x: train_set_x[index * batch_size: (index + 1) * batch_size],\n", " y: train_set_y[index * batch_size: (index + 1) * batch_size]\n", " }\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "For simplicity I just omitted the classification error from the output. The error provided by one call of this function is just the average missclassification rate for this particular batch. If you want you can add the error output in again and then during training compute the average training error over a whole epoch. " ] }, { "cell_type": "code", "execution_count": 29, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iteration: 0\n", "Validation error: 0.0991\n", "Iteration: 1\n", "Validation error: 0.0957\n", "Iteration: 2\n", "Validation error: 0.093\n", "Iteration: 3\n", "Validation error: 0.0967\n", "Iteration: 4\n", "Validation error: 0.0918\n", "Iteration: 5\n", "Validation error: 0.0938\n", "Iteration: 6\n", "Validation error: 0.092\n", "Iteration: 7\n", "Validation error: 0.0937\n", "Iteration: 8\n", "Validation error: 0.0933\n", "Iteration: 9\n", "Validation error: 0.0926\n", "The code ran for 19.849208 seconds\n" ] } ], "source": [ "# Step 4. Perform computation\n", "# reset W and b to zero\n", "W.set_value(np.zeros((28*28, 10),\n", " dtype=theano.config.floatX)) \n", "b.set_value(np.zeros((10,),\n", " dtype=theano.config.floatX))\n", "\n", "val_monitor_batch = []\n", "n_epochs = 10\n", "\n", "start_time = time.clock()\n", "for epoch in range(n_epochs):\n", " for batch_index in range(n_train_batches):\n", " loss = train_batch(batch_index) \n", " val_monitor_batch.append(validate())\n", " if epoch%1 == 0:\n", " print \"Iteration: \", epoch\n", " print \"Validation error: \",val_monitor_batch[-1]\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time)) " ] }, { "cell_type": "code", "execution_count": 30, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEPCAYAAABV6CMBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlZna6DkAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plt.plot(val_monitor_batch, c='b')\n", "plt.xlabel('Number of Epochs')\n", "plt.ylabel('Missclassification')\n", "plt.show()\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Overall our logistic regression seems to be able to reach a missclassification rate of around 7.5%. That is in the range of the result of the (deep learning tutorial)[http://deeplearning.net/tutorial/logreg.html]. But, we have barely started yet! After all this workshop is about deep learning, so let's add some layers to our model! But first it is time for another exercise." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Exercise: Learning Rate Decay\n", "There are a lot of tips and tricks to influence the training of a neural network. One option is to have the learning rate slowly decay over the iterations. Modify our training function such that the learning rate is multiplied with a factor of $0.995$ after each gradient step. Hint: the idea is to define an update for the learning rate." ] }, { "cell_type": "code", "execution_count": 31, "metadata": { "collapsed": false }, "outputs": [], "source": [ "## What we had so far\n", "learning_rate = 0.1\n", "\n", "updates = [(W, W - learning_rate * g_W),\n", " (b, b - learning_rate * g_b)]\n", "\n", "train_batch_lr = theano.function(inputs=[index],\n", " outputs=cost, \n", " updates=updates,\n", " givens={\n", " x: train_set_x[index * batch_size: (index + 1) * batch_size],\n", " y: train_set_y[index * batch_size: (index + 1) * batch_size]\n", " }\n", " )" ] }, { "cell_type": "code", "execution_count": 33, "metadata": { "collapsed": false }, "outputs": [], "source": [ "# your solution\n" ] }, { "cell_type": "code", "execution_count": 34, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.10000000149\n", "0.0995000004768\n", "0.0990025028586\n", "0.0985074937344\n", "0.0980149582028\n", "0.0975248813629\n", "0.097037255764\n", "0.096552066505\n", "0.0960693061352\n", "0.0955889597535\n" ] } ], "source": [ "# test your solution by calling the training \n", "#function and printing the learning rate\n", "\n", "for i in xrange(10):\n", " print shared_learning_rate.get_value()\n", " train_batch_lr(1)\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Making the model deep" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now that we understand the code for logistic regression, we are going to just use the pre-defined class `LogisticRegression` from the [Theano deep learning tutorial](http://deeplearning.net/tutorial/logreg.html) as our last classification layer. To make our model deeper we need to define a hidden layer. Now that we know about _shared_variables_, _Theano_functions_, and how to do _stochastic_gradient_decent_, this is actually pretty easy, and we can just read and understand the definition of the `HiddenLayer` class from the [Theano deep learning tutorial](http://deeplearning.net/tutorial/mlp.html).\n", "\n", "Short recap: \n", "\n", "![MLP](http://deeplearning.net/tutorial/_images/mlp.png)\n", "\n", "The hidden layer computes the function $s(b+Wx)$, where $s$ is our activation function, $b$ are the biases and $W$ is the weight matrix. " ] }, { "cell_type": "code", "execution_count": 35, "metadata": { "collapsed": false }, "outputs": [], "source": [ "from logistic_sgd import LogisticRegression\n", "\n", "class HiddenLayer(object):\n", " def __init__(self, rng, input, n_in, n_out, W=None, b=None,\n", " activation=T.tanh):\n", "\n", " self.input = input\n", "\n", " if W is None:\n", " W_values = np.asarray(\n", " rng.uniform(\n", " low=-np.sqrt(6. / (n_in + n_out)),\n", " high=np.sqrt(6. / (n_in + n_out)),\n", " size=(n_in, n_out)\n", " ),\n", " dtype=theano.config.floatX\n", " )\n", " if activation == theano.tensor.nnet.sigmoid:\n", " W_values *= 4\n", "\n", " W = theano.shared(value=W_values, name='W', borrow=True)\n", "\n", " if b is None:\n", " b_values = np.zeros((n_out,), dtype=theano.config.floatX)\n", " b = theano.shared(value=b_values, name='b', borrow=True)\n", "\n", " self.W = W\n", " self.b = b\n", "\n", " lin_output = T.dot(input, self.W) + self.b\n", " self.output = (\n", " lin_output if activation is None\n", " else activation(lin_output)\n", " )\n", " # parameters of the model\n", " self.params = [self.W, self.b]\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "With the definition of this class we also start to being introduced to the `black magic` of deep networks. The intialization for $W$ looks quite strange. The reason is a paper from [Glorot et al.](http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf) which looked at gradient propagation in deep networks and the effect of initialization, and proposed this normalized intialization as a good way to intialize networks with tanh activation. \n", "\n", "So we have now a class for the logistic regression layer and a class for a hidden layer. All that is missing is putting it all together. The class name `MLP` stands for Multi Layer Perceptron." ] }, { "cell_type": "code", "execution_count": 36, "metadata": { "collapsed": false }, "outputs": [], "source": [ "class MLP(object):\n", "\n", " def __init__(self, rng, input, n_in, n_hidden, n_out):\n", "\n", " self.hiddenLayer = HiddenLayer(\n", " rng=rng,\n", " input=input,\n", " n_in=n_in,\n", " n_out=n_hidden,\n", " activation=T.tanh\n", " )\n", "\n", " self.logRegressionLayer = LogisticRegression(\n", " input=self.hiddenLayer.output,\n", " n_in=n_hidden,\n", " n_out=n_out\n", " )\n", "\n", " self.negative_log_likelihood = (\n", " self.logRegressionLayer.negative_log_likelihood\n", " )\n", "\n", " self.errors = self.logRegressionLayer.errors\n", "\n", " self.params = self.hiddenLayer.params + self.logRegressionLayer.params" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Note that the input for the logistic regression layer is the output of the hidden layer. This way the two layers are connected in the model. This means we can just use the error and loss functions from the logistic regression layer as the error and loss function of our whole model. This might seem strange at first, but the error of the logistic regression layer is defined in terms of the layer's input. This input now is connected to the output of the hidden layer, thus computing the gradient of the loss function will unravel the whole network all the way down to the first layer's input, which is our image data. \n", "\n", "\n", "Now we need to define a new training function along with new functions for testing and validation. We use two separate data sets for validation and testing. The reason is that deep network classifiers have a lot of parameters to tune. We use the training data for the actual gradient decent optimization, but we also have to tune the hyper-parameter, like how many layers to choose, how many units in each layer, which learning rate, activation funtion, etc. In a way this is like a whole second level of optimizing our network. This means our validation data can be seen as part of the training data, just for training different parameters than the training data. The test data then is there to test the generalization performance our our trained network. If you are strict about it, you are only allowed to compute the test error once all parameters in the network are fixed. " ] }, { "cell_type": "code", "execution_count": 38, "metadata": { "collapsed": false }, "outputs": [], "source": [ "n_hidden = 100\n", "classifier = MLP(\n", " rng=rng,\n", " input=x,\n", " n_in=28 * 28,\n", " n_hidden=n_hidden,\n", " n_out=10\n", " )\n", "\n", "cost = classifier.negative_log_likelihood(y)\n", "\n", "n_validation_batches= valid_set_x.shape[0].eval() / batch_size\n", "\n", "validate_model = theano.function(\n", " inputs=[index],\n", " outputs=classifier.errors(y),\n", " givens={\n", " x: valid_set_x[index * batch_size:(index + 1) * batch_size],\n", " y: valid_set_y[index * batch_size:(index + 1) * batch_size]\n", " }\n", " )\n", "\n", "test_model = theano.function(\n", " inputs=[],\n", " outputs=classifier.errors(y),\n", " givens={\n", " x: test_set_x,\n", " y: test_set_y\n", " }\n", " )\n", "\n", "\n", "gparams = [T.grad(cost, param) for param in classifier.params]\n", "\n", "learning_rate = 0.1\n", "updates = [(param, param - learning_rate * gparam)\n", " for param, gparam in zip(classifier.params, gparams)\n", " ]\n", "\n", "train_model = theano.function(\n", " inputs=[index],\n", " outputs=cost,\n", " updates=updates,\n", " givens={\n", " x: train_set_x[index * batch_size: (index + 1) * batch_size],\n", " y: train_set_y[index * batch_size: (index + 1) * batch_size]\n", " }\n", " )\n", " " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Let's give it a try!**" ] }, { "cell_type": "code", "execution_count": 39, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "epoch 0, minibatches 2500, validation error 5.330000 %\n", "epoch 1, minibatches 2500, validation error 4.370000 %\n", "epoch 2, minibatches 2500, validation error 4.140000 %\n", "epoch 3, minibatches 2500, validation error 3.870000 %\n", "epoch 4, minibatches 2500, validation error 3.930000 %\n", "epoch 5, minibatches 2500, validation error 3.810000 %\n", "epoch 6, minibatches 2500, validation error 3.660000 %\n", "epoch 7, minibatches 2500, validation error 3.600000 %\n", "epoch 8, minibatches 2500, validation error 3.510000 %\n", "epoch 9, minibatches 2500, validation error 3.460000 %\n", "The code ran for 32.392306 seconds\n", "Test error is 3.570000 %\n" ] } ], "source": [ "start_time = time.clock()\n", "\n", "n_epochs = 10\n", "start_time = time.clock()\n", "\n", "for epoch in xrange(n_epochs):\n", " for minibatch_index in xrange(n_train_batches):\n", " minibatch_avg_cost = train_model(minibatch_index)\n", " \n", " validation_losses = [validate_model(i) for i\n", " in xrange(n_validation_batches)]\n", " \n", " this_validation_loss = np.mean(validation_losses)\n", " print('epoch %i, minibatches %i, validation error %f %%' %\n", " (\n", " epoch,\n", " n_train_batches,\n", " this_validation_loss * 100.\n", " )\n", " )\n", "\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time)) \n", "\n", "print \"Test error is %f %%\" % (test_model() * 100)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Visualizing the Weights\n", "\n", "It is very hard to understand what a DNN is actually learning. One way to get a bit of insight is to look at the weight matrix $W$ of the first layer. The weights conveniently have the same size as our input and can be understood as the filters that the DNN learned at the first stage. \n", "\n", "We can again use the `tile_raster_images` function to visualize the weights." ] }, { "cell_type": "code", "execution_count": 40, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQMAAAEACAYAAAC3RRNlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\neDFwHCfixcBxHMCLgeM4ES8GjuMA8AvC0PXlCfiqrQAAAABJRU5ErkJggg==\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "filters = tile_raster_images(classifier.hiddenLayer.W.T.eval(), img_shape=(28,28), tile_shape=(10,10), tile_spacing=(0, 0),\n", " scale_rows_to_unit_interval=True,\n", " output_pixel_vals=True)\n", "\n", "plt.imshow(filters)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Even though our network was not very optimized, we can see that the filters are starting to form patterns. Some of these patterns can be interpreted as looking for line strokes, which makes sense given that our training data consists of handwritten digits. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Exercise: Hidden Unit Activations\n", "It can be useful to not only look at the learned filters, but also at the activation of the hidden units. Write a function that computes the activation of the hidden units for the validation data. The result should be a nr_of_samples $\\times$ nr_of_hidden_units matrix. You can visualize the result for the first 100 validation samples. If all is working well the visualization should show a random pattern without obvious horizontal or vertical lines.\n" ] }, { "cell_type": "code", "execution_count": 41, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQwAAAEOCAYAAAB8RCmYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Adding even more Layers\n", "According to the universal approximation theorem a single hidden layer neural network with a linear output unit can approximate any continuous function arbitrarily well, given enough hidden units. The result applies for sigmoid, tanh and many other hidden layer activation functions. While this is a great result and sounds like we don't really need to bother about adding more layers, it can be very hard in practice to find the right local minimum of such a network. It also doesn't specify how many hidden units are enough. It is computationally more feasible to stack moderate size hidden layers on top of each other, than to try and optimize a gigantic single hidden layer. So finally here is the generalization of our `MLP` class for an arbitrary number of hidden layers:\n", "\n" ] }, { "cell_type": "code", "execution_count": 42, "metadata": { "collapsed": false }, "outputs": [], "source": [ "class MLP_deep(object):\n", " def __init__(self, rng, input, n_in, n_hidden, n_out):\n", " \n", " self.hiddenLayers = []\n", " \n", " self.params = []\n", " \n", " next_input = input\n", " next_n_in = n_in\n", " \n", " for n_h in n_hidden:\n", " hl = HiddenLayer(\n", " rng=rng,\n", " input=next_input,\n", " n_in=next_n_in,\n", " n_out=n_h,\n", " activation=T.tanh\n", " )\n", " next_input = hl.output\n", " next_n_in = n_h\n", " self.hiddenLayers.append(hl)\n", " self.params += hl.params \n", "\n", "\n", " self.logRegressionLayer = LogisticRegression(\n", " input=self.hiddenLayers[-1].output,\n", " n_in=n_hidden[-1],\n", " n_out=n_out\n", " )\n", "\n", " self.negative_log_likelihood = self.logRegressionLayer.negative_log_likelihood\n", " self.errors = self.logRegressionLayer.errors\n", "\n", " self.params += self.logRegressionLayer.params" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Note that again we need to take care to get the layers connected by specifying the output of the lower layer as input for the new layer on top. All the rest should look fairly familiar by now:" ] }, { "cell_type": "code", "execution_count": 43, "metadata": { "collapsed": false }, "outputs": [], "source": [ "n_hidden = [200,100]\n", "classifier = MLP_deep(\n", " rng=rng,\n", " input=x,\n", " n_in=28 * 28,\n", " n_hidden=n_hidden,\n", " n_out=10\n", " )\n", "\n", "cost = classifier.negative_log_likelihood(y)\n", "\n", "n_validation_batches= valid_set_x.shape[0].eval() / batch_size\n", "\n", "validate_model = theano.function(\n", " inputs=[index],\n", " outputs=classifier.errors(y),\n", " givens={\n", " x: valid_set_x[index * batch_size:(index + 1) * batch_size],\n", " y: valid_set_y[index * batch_size:(index + 1) * batch_size]\n", " }\n", " )\n", "\n", "test_model = theano.function(\n", " inputs=[],\n", " outputs=classifier.errors(y),\n", " givens={\n", " x: test_set_x,\n", " y: test_set_y\n", " }\n", " )\n", "\n", "\n", "gparams = [T.grad(cost, param) for param in classifier.params]\n", "\n", "learning_rate = 0.1\n", "updates = [(param, param - learning_rate * gparam)\n", " for param, gparam in zip(classifier.params, gparams)\n", " ]\n", "\n", "train_model = theano.function(\n", " inputs=[index],\n", " outputs=cost,\n", " updates=updates,\n", " givens={\n", " x: train_set_x[index * batch_size: (index + 1) * batch_size],\n", " y: train_set_y[index * batch_size: (index + 1) * batch_size]\n", " }\n", " )\n", " " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**And here we go!**" ] }, { "cell_type": "code", "execution_count": 44, "metadata": { "collapsed": false }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "epoch 0, minibatches 2500, validation error 4.730000 %\n", "epoch 1, minibatches 2500, validation error 3.790000 %\n", "epoch 2, minibatches 2500, validation error 3.320000 %\n", "epoch 3, minibatches 2500, validation error 3.080000 %\n", "epoch 4, minibatches 2500, validation error 2.930000 %\n", "epoch 5, minibatches 2500, validation error 2.840000 %\n", "epoch 6, minibatches 2500, validation error 2.710000 %\n", "epoch 7, minibatches 2500, validation error 2.680000 %\n", "epoch 8, minibatches 2500, validation error 2.690000 %\n", "epoch 9, minibatches 2500, validation error 2.640000 %\n", "The code ran for 41.740476 seconds\n", "Test error is 3.220000 %\n" ] } ], "source": [ "start_time = time.clock()\n", "\n", "n_epochs = 10\n", "start_time = time.clock()\n", "\n", "for epoch in xrange(n_epochs):\n", " for minibatch_index in xrange(n_train_batches):\n", " minibatch_avg_cost = train_model(minibatch_index)\n", " \n", " validation_losses = [validate_model(i) for i\n", " in xrange(n_validation_batches)]\n", " this_validation_loss = np.mean(validation_losses)\n", " print('epoch %i, minibatches %i, validation error %f %%' %\n", " (\n", " epoch,\n", " n_train_batches,\n", " this_validation_loss * 100.\n", " )\n", " )\n", "\n", "end_time = time.clock() \n", "print 'The code ran for %f seconds' % ((end_time - start_time)) \n", "\n", "print \"Test error is %f %%\" % (test_model() * 100)" ] }, { "cell_type": "code", "execution_count": 45, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQMAAAEACAYAAAC3RRNlAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n03E6jsfpYnA6TsfpAAD8PwwD0OqFrNJWAAAAAElFTkSuQmCC\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "filters = tile_raster_images(classifier.hiddenLayers[0].W.T.eval(), img_shape=(28,28), tile_shape=(10,10), tile_spacing=(0, 0),\n", " scale_rows_to_unit_interval=True,\n", " output_pixel_vals=True)\n", "\n", "plt.imshow(filters)\n", "plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Exercise: Playing around\n", "\n", "Possible things to try:\n", "\n", "- Play around with the batch size\n", "- Play with different learning rates\n", "- Switch activation function from tanh to sigmoid\n", "- Train a lot of narrow layers\n", "- Train just one very large layer\n", "- add L1 or L2 regularization to the loss function and see the effect (this is also covered in the Theano Deep learning tutorial on MLP)\n" ] }, { "cell_type": "code", "execution_count": null, "metadata": { "collapsed": false }, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 2", "language": "python", "name": "python2" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 2 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython2", "version": "2.7.9" } }, "nbformat": 4, "nbformat_minor": 0 }