{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MLhw2_Q1.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "svPQYZtKXaSA"
      },
      "source": [
        "Neural Network definition"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8yRHqA15eC46"
      },
      "source": [
        "import numpy as np\n",
        "import math\n",
        "import matplotlib.pyplot as plt\n",
        "import copy\n",
        "\n",
        "class MyNeuralNetwork():\n",
        "    \"\"\"\n",
        "    My implementation of a Neural Network Classifier.\n",
        "    \"\"\"\n",
        "\n",
        "    acti_fns = ['relu', 'sigmoid', 'linear', 'tanh', 'softmax']\n",
        "    weight_inits = ['zero', 'random', 'normal']\n",
        "\n",
        "    def __init__(self, n_layers, layer_sizes, activation, learning_rate, weight_init, batch_size, num_epochs):\n",
        "        \"\"\"\n",
        "        Initializing a new MyNeuralNetwork object\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        n_layers : int value specifying the number of layers\n",
        "\n",
        "        layer_sizes : integer array of size n_layers specifying the number of nodes in each layer\n",
        "\n",
        "        activation : string specifying the activation function to be used\n",
        "                     possible inputs: relu, sigmoid, linear, tanh\n",
        "\n",
        "        learning_rate : float value specifying the learning rate to be used\n",
        "\n",
        "        weight_init : string specifying the weight initialization function to be used\n",
        "                      possible inputs: zero, random, normal\n",
        "\n",
        "        batch_size : int value specifying the batch size to be used\n",
        "\n",
        "        num_epochs : int value specifying the number of epochs to be used\n",
        "        \"\"\"\n",
        "\n",
        "        if activation not in self.acti_fns:\n",
        "            raise Exception('Incorrect Activation Function')\n",
        "\n",
        "        if weight_init not in self.weight_inits:\n",
        "            raise Exception('Incorrect Weight Initialization Function')\n",
        "        pass\n",
        "        self.n_layers = n_layers\n",
        "        self.layer_sizes = layer_sizes\n",
        "        self.activation = activation\n",
        "        self.learning_rate = learning_rate\n",
        "        self.weight_init = weight_init\n",
        "        self.batch_size = batch_size\n",
        "        self.num_epochs = num_epochs\n",
        "        # weights and biases\n",
        "        self.Ws = []\n",
        "        self.Bs = []\n",
        "        # storing weights and biases after each epoch (for plotting loss-vs-epochs curves)\n",
        "        self.wts_epochs = [] \n",
        "        self.bias_epochs = []\n",
        "\n",
        "    def relu(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the ReLU activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "    \n",
        "        x_calc = np.maximum(X,0)\n",
        "        return x_calc\n",
        "\n",
        "    def relu_grad(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the gradient of ReLU activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = np.zeros(shape = X.shape)\n",
        "        x_calc[np.where(X>0)]=1\n",
        "        return x_calc\n",
        "\n",
        "    def sigmoid(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the Sigmoid activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = 1 / (1 + np.exp(-X))\n",
        "        return x_calc\n",
        "\n",
        "    def sigmoid_grad(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the gradient of Sigmoid activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        sig = self.sigmoid(X)\n",
        "        x_calc = sig*(1-sig)\n",
        "        return x_calc\n",
        "\n",
        "    def linear(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the Linear activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = np.array(X)\n",
        "        return x_calc\n",
        "\n",
        "    def linear_grad(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the gradient of Linear activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = np.ones(shape = X.shape)\n",
        "        return x_calc\n",
        "\n",
        "    def tanh(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the Tanh activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = ((np.exp(X) - np.exp(-X))/(np.exp(X) + np.exp(-X)))\n",
        "        return x_calc\n",
        "\n",
        "    def tanh_grad(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the gradient of Tanh activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        tnh = self.tanh(X)\n",
        "        x_calc = 1 - np.square(tnh)\n",
        "        return x_calc\n",
        "\n",
        "    def softmax(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the Softmax activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x_calc = np.exp(X) / (np.sum(np.exp(X), axis=-1).reshape(X.shape[0],1))\n",
        "        return x_calc\n",
        "\n",
        "    def softmax_grad(self, X):\n",
        "        \"\"\"\n",
        "        Calculating the gradient of Softmax activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimentional numpy array of shape=(batch_size, layer_size)\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        x_calc : 2-dimensional numpy array after calculating the necessary function over X\n",
        "        \"\"\"\n",
        "        x = np.array(X)\n",
        "        sm = self.softmax(x)\n",
        "        x_calc = sm*(1-sm)\n",
        "        return x_calc\n",
        "\n",
        "    def zero_init(self, shape):\n",
        "        \"\"\"\n",
        "        Calculating the initial weights after Zero Activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        shape : tuple specifying the shape of the layer for which weights have to be generated \n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        weight : 2-dimensional numpy array which contains the initial weights for the requested layer\n",
        "        \"\"\"\n",
        "        weight = np.zeros(shape)\n",
        "        return weight\n",
        "\n",
        "    def random_init(self, shape):\n",
        "        \"\"\"\n",
        "        Calculating the initial weights after Random Activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        shape : tuple specifying the shape of the layer for which weights have to be generated \n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        weight : 2-dimensional numpy array which contains the initial weights for the requested layer\n",
        "        \"\"\"\n",
        "        np.random.seed(70)\n",
        "        weight = 0.01*np.random.rand(shape)\n",
        "        return weight\n",
        "\n",
        "    def normal_init(self, shape):\n",
        "        \"\"\"\n",
        "        Calculating the initial weights after Normal(0,1) Activation for a particular layer\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        shape : tuple specifying the shape of the layer for which weights have to be generated \n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        weight : 2-dimensional numpy array which contains the initial weights for the requested layer\n",
        "        \"\"\"\n",
        "        np.random.seed(70)\n",
        "        weight = 0.01*np.random.normal(0, 1, shape)\n",
        "        return weight\n",
        "\n",
        "    def init_weights(self, shape):\n",
        "      \"\"\"\n",
        "      Applying the initialization according to hyperparameter\n",
        "      \"\"\"\n",
        "      if self.weight_init=='zero':\n",
        "        return self.zero_init(shape)\n",
        "      elif self.weight_init=='random':\n",
        "        return self.random_init(shape)\n",
        "      else:\n",
        "        return self.normal_init(shape)\n",
        "\n",
        "    def cross_entropy(self, true_labels, pred_probs):\n",
        "      \"\"\"\n",
        "      Calculating cross entropy loss\n",
        "      \"\"\"\n",
        "      n = pred_probs.shape[0]\n",
        "      crossentropyloss = -np.sum(true_labels * np.log(pred_probs + (0.00000000001))) / n\n",
        "      return crossentropyloss\n",
        "    \n",
        "    def activ(self,z):\n",
        "      \"\"\"\n",
        "      Apply activation according to hyperparameter\n",
        "      \"\"\"\n",
        "      if self.activation=='relu':\n",
        "        return self.relu(z)\n",
        "      elif self.activation=='sigmoid':\n",
        "        return self.sigmoid(z)\n",
        "      elif self.activation=='linear':\n",
        "        return self.linear(z)\n",
        "      elif self.activation=='tanh':\n",
        "        return self.tanh(z)\n",
        "      else:\n",
        "        return self.softmax(z)\n",
        "    \n",
        "    def activ_grad(self,z):\n",
        "      \"\"\"\n",
        "      find gradient of activation function according to hyperparameter\n",
        "      \"\"\"\n",
        "      if self.activation=='relu':\n",
        "        return self.relu_grad(z)\n",
        "      elif self.activation=='sigmoid':\n",
        "        return self.sigmoid_grad(z)\n",
        "      elif self.activation=='linear':\n",
        "        return self.linear_grad(z)\n",
        "      elif self.activation=='tanh':\n",
        "        return self.tanh_grad(z)\n",
        "      else:\n",
        "        return self.softmax_grad(z)\n",
        "\n",
        "    def one_hot_encode(self, y):\n",
        "        \"\"\"\n",
        "        Function to one-hot-encode y\n",
        "        \"\"\"\n",
        "        ycats = list(np.unique(y))\n",
        "        ncats = len(ycats)\n",
        "        y_hot = np.zeros(shape=(ncats,ncats))\n",
        "        for i in range(ncats):\n",
        "          y_hot[i][i] = 1\n",
        "        yonehot = []\n",
        "        for i in range(len(y)):\n",
        "          yonehot.append(y_hot[ycats.index(y[i])])\n",
        "        \n",
        "        yonehot = np.array(yonehot)\n",
        "        return yonehot\n",
        "\n",
        "    def fit(self, X, y):\n",
        "        \"\"\"\n",
        "        Fitting (training) the linear model.\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimensional numpy array of shape (n_samples, n_features) which acts as training data.\n",
        "\n",
        "        y : 1-dimensional numpy array of shape (n_samples,) which acts as training labels.\n",
        "        \n",
        "        Returns\n",
        "        -------\n",
        "        self : an instance of self\n",
        "        \"\"\"\n",
        "        X = np.array(X)\n",
        "        y = np.array(y)\n",
        "\n",
        "        # one hot encoding y\n",
        "        y_onehot = self.one_hot_encode(y)\n",
        "\n",
        "        # initializing weights and biases\n",
        "        Ns = self.layer_sizes\n",
        "        \n",
        "        self.Ws.append([])\n",
        "        self.Bs.append([])\n",
        "      \n",
        "        for i in range(1, self.n_layers):\n",
        "          w = self.init_weights((Ns[i-1],Ns[i]))\n",
        "          b = np.zeros(shape=(Ns[i]))\n",
        "          self.Ws.append(w)\n",
        "          self.Bs.append(b)\n",
        "          \n",
        "        # calculate number of batches\n",
        "        n_batches = X.shape[0]//self.batch_size\n",
        "\n",
        "        # training for each epoch\n",
        "        for i in range(self.num_epochs):\n",
        "          for j in range(n_batches):\n",
        "            index1 = j*self.batch_size\n",
        "            index2 = index1 + self.batch_size\n",
        "            X_batch = X[index1:index2:]\n",
        "            Y_batch = y_onehot[index1:index2]\n",
        "            y_forward = self.forward_pass(X_batch)\n",
        "            Loss = self.cross_entropy(Y_batch, y_forward)\n",
        "            self.backpropagate(X_batch, Y_batch, y_forward)\n",
        "          self.wts_epochs.append(copy.deepcopy(self.Ws))\n",
        "          self.bias_epochs.append(copy.deepcopy(self.Bs))\n",
        "       \n",
        "        return self\n",
        "    \n",
        "    def forward_pass(self, X):\n",
        "        \"\"\"\n",
        "        Forward pass to predict the probabilities of classes according to the weights\n",
        "        \"\"\"\n",
        "        self.As = []\n",
        "        self.Zs = []\n",
        "        self.Zs.append([])\n",
        "        self.As.append(np.array(X))\n",
        "\n",
        "        for i in range(1, self.n_layers):\n",
        "          z = np.dot(self.As[i-1], self.Ws[i]) + self.Bs[i]\n",
        "          if (i == self.n_layers-1):\n",
        "            a = self.softmax(z)\n",
        "          else:\n",
        "            a = self.activ(z)\n",
        "          self.Zs.append(z)\n",
        "          self.As.append(a)\n",
        "        \n",
        "        y_forward = self.As[-1]\n",
        "        return y_forward\n",
        "\n",
        "    def backpropagate(self, X, y_onehot, y_forward):\n",
        "        \"\"\"\n",
        "        Applying gradient descent to train the weigths and biases using backpropagation\n",
        "        \"\"\"\n",
        "        dz = (y_forward - y_onehot)/y_onehot.shape[0]\n",
        "        dw = np.dot(self.As[self.n_layers-2].T, dz)\n",
        "        db = dz.mean(axis=0)*self.As[self.n_layers-2].shape[0]\n",
        "\n",
        "        dz_next = np.dot(dz, self.Ws[self.n_layers-1].T)\n",
        "\n",
        "        self.Ws[self.n_layers-1] = self.Ws[self.n_layers-1] - (self.learning_rate*dw)\n",
        "        self.Bs[self.n_layers-1] = self.Bs[self.n_layers-1] - (self.learning_rate*db)\n",
        "\n",
        "        for i in range(self.n_layers-2, 0, -1):\n",
        "            act_grad = self.activ_grad(self.Zs[i])\n",
        "            dz = dz_next * act_grad\n",
        "            dw = np.dot(self.As[i-1].T, dz) \n",
        "            db = dz.mean(axis=0)*self.As[i-1].shape[0]\n",
        "            \n",
        "            dz_next = np.dot(dz, self.Ws[i].T)\n",
        "\n",
        "            self.Ws[i] = self.Ws[i] - (self.learning_rate * dw)\n",
        "            self.Bs[i] = self.Bs[i] - (self.learning_rate * db)\n",
        "\n",
        "        \n",
        "    def predict_proba(self, X):\n",
        "        \"\"\"\n",
        "        Predicting probabilities using the trained linear model.\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimensional numpy array of shape (n_samples, n_features) which acts as testing data.\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        y : 2-dimensional numpy array of shape (n_samples, n_classes) which contains the \n",
        "            class wise prediction probabilities.\n",
        "        \"\"\"\n",
        "        y_probs = self.forward_pass(X)\n",
        "        return y_probs\n",
        "\n",
        "    def predict(self, X):\n",
        "        \"\"\"\n",
        "        Predicting values using the trained linear model.\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimensional numpy array of shape (n_samples, n_features) which acts as testing data.\n",
        "\n",
        "        Returns\n",
        "        # return the numpy array y which contains the predicted values\n",
        "        -------\n",
        "        y : 1-dimensional numpy array of shape (n_samples,) which contains the predicted values.\n",
        "        \"\"\"\n",
        "        probs = self.predict_proba(X)\n",
        "        n_samples = probs.shape[0]\n",
        "        y = np.zeros(shape=n_samples)\n",
        "        for i in range(n_samples):\n",
        "          cls = np.argmax(probs[i])\n",
        "          y[i] = cls\n",
        "\n",
        "        y = np.array(y)\n",
        "        return y\n",
        "\n",
        "    def score(self, X, y):\n",
        "        \"\"\"\n",
        "        Predicting values using the trained linear model.\n",
        "\n",
        "        Parameters\n",
        "        ----------\n",
        "        X : 2-dimensional numpy array of shape (n_samples, n_features) which acts as testing data.\n",
        "\n",
        "        y : 1-dimensional numpy array of shape (n_samples,) which acts as testing labels.\n",
        "\n",
        "        Returns\n",
        "        -------\n",
        "        acc : float value specifying the accuracy of the model on the provided testing set\n",
        "        \"\"\"\n",
        "        y = np.array(y)\n",
        "        predicted = self.predict(X)\n",
        "        n_samples = y.shape[0]\n",
        "        acc = np.sum(y==predicted)/n_samples\n",
        "      \n",
        "        return acc\n",
        "\n",
        "    def losses_epochs(self, X, y):\n",
        "        \"\"\"\n",
        "        Calculate loss after each epoch according to the weights at each epoch\n",
        "        \"\"\"\n",
        "        losses = []\n",
        "        for ep in range(self.num_epochs):\n",
        "            A = X\n",
        "            for i in range(1, self.n_layers):\n",
        "                z = np.dot(A, self.wts_epochs[ep][i]) + self.bias_epochs[ep][i]\n",
        "                if (i == self.n_layers-1):\n",
        "                  a = self.softmax(z)\n",
        "                else:\n",
        "                  a = self.activ(z)\n",
        "                A = a\n",
        "            ls = self.cross_entropy(y, A)\n",
        "            losses.append(ls)\n",
        "        return losses\n",
        "    \n",
        "    def plot_loss_epoch(self, X_train, y_train, X_val, y_val):\n",
        "        \"\"\"\n",
        "        Plot the training-loss vs epoch curve and validation-loss vs epoch curve\n",
        "        \"\"\"\n",
        "        Y_train = self.one_hot_encode(y_train)\n",
        "        Y_val = self.one_hot_encode(y_val)\n",
        "        \n",
        "        train_losses = self.losses_epochs(X_train, Y_train)\n",
        "        val_losses = self.losses_epochs(X_val, Y_val)\n",
        "\n",
        "        plt.plot(train_losses,label='Training loss')\n",
        "        plt.plot(val_losses,label='Validation loss')\n",
        "        plt.ylabel('Cross Entropy loss')\n",
        "        plt.xlabel('Epochs')\n",
        "        plt.legend()\n",
        "        plt.title(self.activation+\": Loss vs Epochs\")\n",
        "        plt.show()\n"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0y81KeSLXr0j"
      },
      "source": [
        "Loading Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FtXd-Tdoksdj"
      },
      "source": [
        "from keras.datasets import mnist\n",
        "import pickle\n",
        "\n",
        "(train_X, train_y), (test_X, test_y) = mnist.load_data()\n",
        "\n",
        "train_X = train_X.reshape((train_X.shape[0],784)) # flatten images\n",
        "train_X = train_X/255.0 # normalize data\n",
        "\n",
        "# divide train-set into train and validation\n",
        "ind = int(0.8*(train_X.shape[0]))\n",
        "X_train = train_X[:ind]\n",
        "y_train = train_y[:ind]\n",
        "\n",
        "X_val = train_X[ind:]\n",
        "y_val = train_y[ind:]\n",
        "\n",
        "X_test = test_X.reshape((test_X.shape[0],784)) # flatten images\n",
        "X_test = X_test/255.0 # normalize data"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7u4bSfu2XxAX"
      },
      "source": [
        "ReLU\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xfg5mmAhthzT",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "outputId": "7cad973d-c6b7-4dfe-91be-9c137ba85308"
      },
      "source": [
        "relu_model = MyNeuralNetwork(5,[784,256,128,64,10],'relu',0.1, 'normal', 100, 100)\n",
        "relu_model.fit(X_train, y_train)\n",
        "print(\"ReLU\")\n",
        "print(\"Train accuracy:\", relu_model.score(X_train, y_train))\n",
        "print(\"Validation accuracy:\", relu_model.score(X_val, y_val))\n",
        "print(\"Test accuracy:\", relu_model.score(X_test, test_y))\n",
        "relu_model.plot_loss_epoch(X_train, y_train, X_val, y_val)\n",
        "pickle.dump(relu_model, open(\"relu_model\",\"wb\"))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ReLU\n",
            "Train accuracy: 1.0\n",
            "Validation accuracy: 0.97425\n",
            "Test accuracy: 0.974\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qELtMpCqX0pP"
      },
      "source": [
        "Sigmoid"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kDZ2hVY6hEmv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "outputId": "5c2b9cb9-b3da-4866-dc19-8a754e08e4a3"
      },
      "source": [
        "sigmoid_model = MyNeuralNetwork(5,[784,256,128,64,10],'sigmoid',0.1, 'normal', 100, 100)\n",
        "sigmoid_model.fit(X_train, y_train)\n",
        "print(\"Sigmoid\")\n",
        "print(\"Train accuracy:\", sigmoid_model.score(X_train, y_train))\n",
        "print(\"Validation accuracy:\", sigmoid_model.score(X_val, y_val))\n",
        "print(\"Test accuracy:\", sigmoid_model.score(X_test, test_y))\n",
        "sigmoid_model.plot_loss_epoch(X_train, y_train, X_val, y_val)\n",
        "pickle.dump(sigmoid_model, open(\"sigmoid_model\",\"wb\"))"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sigmoid\n",
            "Train accuracy: 0.9604375\n",
            "Validation accuracy: 0.9506666666666667\n",
            "Test accuracy: 0.9487\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "09Pun1YMX2H7"
      },
      "source": [
        "Linear"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BIRghgqZhRNz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "outputId": "7e181774-e2a6-4cf8-deb2-0874bb6bdf6d"
      },
      "source": [
        "linear_model = MyNeuralNetwork(5,[784,256,128,64,10],'linear',0.1, 'normal', 100, 100)\n",
        "linear_model.fit(X_train, y_train)\n",
        "print(\"Linear\")\n",
        "print(\"Train accuracy:\", linear_model.score(X_train, y_train))\n",
        "print(\"Validation accuracy:\", linear_model.score(X_val, y_val))\n",
        "print(\"Test accuracy:\", linear_model.score(X_test, test_y))\n",
        "linear_model.plot_loss_epoch(X_train, y_train, X_val, y_val)\n",
        "pickle.dump(linear_model, open(\"linear_model\",\"wb\"))"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Linear\n",
            "Train accuracy: 0.924125\n",
            "Validation accuracy: 0.9191666666666667\n",
            "Test accuracy: 0.9143\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deZhU1bX38e+vqqu6emAGURkEFTAOjI2oOKBRI8aIGo1yTRSNGn1NjHpjYgYjMfG+yY2512viEOKcqJg4vSbO8xCvAyDRgBKVYAQHEAQaeqphvX+c003RVHcXTVdX070+z1NPndpnWqcLatXe+9TeMjOcc8655iLFDsA551zX5AnCOedcTp4gnHPO5eQJwjnnXE6eIJxzzuXkCcI551xOniBcwUlaJunwcPkHkm4sdkxu20maLekPxY7DFY4nCNepzOw/zOysYseRiySTtHux42gPSbMkpSVtaPbYudixue1XSbEDcK4zSIqaWbrYcRTY/5rZgcUOwnUfXoNwnSq7WULSiPBb++mS/iXpU0k/zNo2IulSSe9JWi3pj5L6Z63/k6SPJa2T9LykvbLW3SrpekkPS9oIHLoNMfeRdLukVZLel/QjSZFw3e6Sngtj+FTS3WG5JP23pJWS1kt6U9LeOY59sqR5zcoukvRguHy0pMWSqiWtkPSddl7DMknfD4/1maRbJCWy1p8t6V1JayQ9mF3zkLSXpCfCdZ9I+kHWoePh36Za0iJJVVn7fS+MuVrSEkmfb0/srng8Qbiu4EBgDPB54MeSPheWfws4DjgE2Bn4DLg2a79HgFHADsAC4I5mx/034EqgF/CipOskXdeO+H4N9AF2DWM5DTgjXPdT4HGgHzA03BbgSOBgYHS471eA1TmO/WdgjKRRzeK+M1y+CfiGmfUC9gaebkf8jU4FvgDsFsb1IwBJhwH/N4xxJ+B9YG64rhfwJPAowXuwO/BU1jGPDbftCzwI/CbcbwzwTWByGPsXgGXbELsrBjPzhz8K+iD4YDg8XJ4N/CFcHgEYMDRr21eBU8Llt4DPZ63bCUgCJTnO0Tc8Vp/w9a3A7VsZpwG7NyuLAg3Anlll3wCeDZdvB+ZkX0NYfhjwD2A/INLGef8A/DhcHgVUA+Xh63+F5+vdxjFmASlgbdbjvWbvwblZr49uXE+QhP4za11l+HceAcwEXm/hnLOBJ7Ne7wnUhsu7AyuBw4FYsf8N+qN9D69BuK7g46zlGoIPKIBdgPslrZW0liBhpIHBkqKSfh42P61n07fTgVnH+qADYhsIxAi+VTd6HxgSLn8XEPBq2MRyJoCZPU3wbfpaYKWkOZJ6t3COOwk+iCGoPTxgZjXh6y8TfJi/HzZl7d9KrC+bWd+sx27N1mf/Pd4nqBEQPjddn5ltIKjtDAGGAe+1cs7m711CUomZvQtcSJBEVkqa6x3m2x9PEK4r+wCY3uxDL2FmKwg+SGcQfEPtQ/BtF4IP60YdMVTxpwTfpnfJKhsOrAAws4/N7Gwz25ngm/51jXdCmdk1ZjaJ4Jv1aOCSFs7xBDBI0niCRNHYvISZvWZmMwia0R4A/rgN1zKs2TV8GC5/mH19kiqAAeE1fkDQtLbVzOxOCzrNdyF4L37RnuO44vEE4bqyG4ArJe0CIGmQpBnhul5APcE33XLgPzronHFJicZHWPbHMI5eYSwXEzQLIekkSUPD7T4j+CDMSJosaYqkGLARqAMyuU5oZkngT8Avgf4ECQNJcUmnSuoTbrO+pWPk6XxJQ8OO/h8Cd4fldwFnSBovqZTgb/mKmS0D/gLsJOlCSaXh32BKWyeSNEbSYeHx6oDabYzdFYEnCNeV/Q9Bx+fjkqqBl4HGD6fbCZpFVgCLw3WtknSDpBva2GwRwYdZ4+MMgs7yjcBS4EWCb/g3h9tPBl6RtCGM9dtmthToDfyOIGm8T5DIftnKee8kqA39ycxSWeVfA5aFzWjnEnQ0t2R/bfk7iMnNzvF4eB3vAT8DMLMngcuAe4GPCDqxTwnXVQNHAF8iaE56h/zuCCsFfk5QA/uYoAb0/Tz2c12IzHzCIOe6O0nLgLPCZOBcXrwG4ZxzLidPEM4553LyJibnnHM5eQ3COedcTt1qsL6BAwfaiBEjih2Gc85tN+bPn/+pmQ3Kta5bJYgRI0Ywb968tjd0zjkHgKT3W1rnTUzOOedy8gThnHMuJ08QzjnncupWfRDOuc6VTCZZvnw5dXV1xQ7FtSGRSDB06FBisVje+3iCcM612/Lly+nVqxcjRoxAUts7uKIwM1avXs3y5csZOXJk3vt5E5Nzrt3q6uoYMGCAJ4cuThIDBgzY6pqeJwjn3Dbx5LB9aM/75AkCeOOOH/DOS/cXOwznnOtSPEEAu//jRlb/7bFih+Gc20qrV69m/PjxjB8/nh133JEhQ4Y0vW5oaGh133nz5nHBBRe0eY4DDjigQ2J99tlnOeaYYzrkWJ3FO6mBesVRur7YYTjnttKAAQNYuHAhALNnz6ayspLvfOc7TetTqRQlJbk/5qqqqqiqqmrzHC+99FLHBLsdKlgNQtIwSc9IWhxO5v7tHNucKukNSW9KeknSuKx1y8LyhZIKOn5GkpgnCOe6iVmzZnHuuecyZcoUvvvd7/Lqq6+y//77M2HCBA444ACWLFkCbP6Nfvbs2Zx55plMmzaNXXfdlWuuuabpeJWVlU3bT5s2jRNPPJE99tiDU089lcbRsB9++GH22GMPJk2axAUXXNBmTWHNmjUcd9xxjB07lv3224833ngDgOeee66pBjRhwgSqq6v56KOPOPjggxk/fjx77703L7zwQof/zVpSyBpECvh3M1sgqRcwX9ITZrY4a5t/AoeY2WeSpgNz2DSlJMChZvZpAWMEoEFxIp4gnNsmP/nzIhZ/uL5Dj7nnzr25/Et7bfV+y5cv56WXXiIajbJ+/XpeeOEFSkpKePLJJ/nBD37Avffeu8U+b7/9Ns888wzV1dWMGTOG8847b4vfDLz++ussWrSInXfemalTp/LXv/6VqqoqvvGNb/D8888zcuRIZs6c2WZ8l19+ORMmTOCBBx7g6aef5rTTTmPhwoVcddVVXHvttUydOpUNGzaQSCSYM2cOX/jCF/jhD39IOp2mpqZmq/8e7VWwBGFmHxHMb4uZVUt6CxhCMH9w4zbZdbeXgaEUQUpxIunW2yudc9uPk046iWg0CsC6des4/fTTeeedd5BEMpnMuc8Xv/hFSktLKS0tZYcdduCTTz5h6NDNP5L23XffprLx48ezbNkyKisr2XXXXZt+XzBz5kzmzJnTanwvvvhiU5I67LDDWL16NevXr2fq1KlcfPHFnHrqqZxwwgkMHTqUyZMnc+aZZ5JMJjnuuOMYP378Nv1ttkan9EFIGgFMAF5pZbOvA49kvTaCyeoN+K2Z5fyLSzoHOAdg+PDh7YovqTjRjNcgnNsW7fmmXygVFRVNy5dddhmHHnoo999/P8uWLWPatGk59yktLW1ajkajpFKpdm2zLS699FK++MUv8vDDDzN16lQee+wxDj74YJ5//nkeeughZs2axcUXX8xpp53WoedtScHvYpJUCdwLXGhmOeufkg4lSBDfyyo+0MwmAtOB8yUdnGtfM5tjZlVmVjVoUM4hzduUjniCcK67WrduHUOGDAHg1ltv7fDjjxkzhqVLl7Js2TIA7r777jb3Oeigg7jjjjuAoG9j4MCB9O7dm/fee4999tmH733ve0yePJm3336b999/n8GDB3P22Wdz1llnsWDBgg6/hpYUNEFIihEkhzvM7L4WthkL3AjMMLPVjeVmtiJ8XgncD+xbqDhTkTjRjDcxOdcdffe73+X73/8+EyZM6PBv/ABlZWVcd911HHXUUUyaNIlevXrRp0+fVveZPXs28+fPZ+zYsVx66aXcdtttAFx99dXsvffejB07llgsxvTp03n22WcZN24cEyZM4O677+bb397ifp+CKdic1Ap+tncbsMbMLmxhm+HA08Bp2f0RkiqASNh3UQE8AVxhZo+2ds6qqiprz4RBf/vFEZQ3rGHUZfO3el/nerK33nqLz33uc8UOo+g2bNhAZWUlZsb555/PqFGjuOiii4od1hZyvV+S5ptZzvt9C9kHMRX4GvCmpIVh2Q+A4QBmdgPwY2AAcF34M/BUGOhg4P6wrAS4s63ksC0ykVJKLHfHlXPOteV3v/sdt912Gw0NDUyYMIFvfOMbxQ6pQxTyLqYXgVYH/zCzs4CzcpQvBcZtuUdhpKOlxMybmJxz7XPRRRd1yRrDtvKhNgCLeIJwzrnmPEEAmZJS4niCcM65bJ4gAKJx4t4H4Zxzm/EEAVhJgjhJCnVHl3PObY88QQBES4kpTSrltQjntieHHnoojz22+VD9V199Needd16L+0ybNo3G2+GPPvpo1q5du8U2s2fP5qqrrmr13A888ACLF28aWu7HP/4xTz755NaEn1NXGhbcEwSgWPDz+fq62iJH4pzbGjNnzmTu3Lmblc2dOzevAfMgGIW1b9++7Tp38wRxxRVXcPjhh7frWF2VJwiAkgQADZ4gnNuunHjiiTz00ENNkwMtW7aMDz/8kIMOOojzzjuPqqoq9tprLy6//PKc+48YMYJPPw0GjL7yyisZPXo0Bx54YNOQ4BD8xmHy5MmMGzeOL3/5y9TU1PDSSy/x4IMPcskllzB+/Hjee+89Zs2axT333APAU089xYQJE9hnn30488wzqa+vbzrf5ZdfzsSJE9lnn314++23W72+Yg8L7hMGAZFYmCDqPUE4126PXAofv9mxx9xxH5j+8xZX9+/fn3333ZdHHnmEGTNmMHfuXL7yla8giSuvvJL+/fuTTqf5/Oc/zxtvvMHYsWNzHmf+/PnMnTuXhQsXkkqlmDhxIpMmTQLghBNO4OyzzwbgRz/6ETfddBPf+ta3OPbYYznmmGM48cQTNztWXV0ds2bN4qmnnmL06NGcdtppXH/99Vx4YTCgxMCBA1mwYAHXXXcdV111FTfeeGOL11fsYcG9BgEoTBBJTxDObXeym5mym5f++Mc/MnHiRCZMmMCiRYs2aw5q7oUXXuD444+nvLyc3r17c+yxxzat+/vf/85BBx3EPvvswx133MGiRYtajWfJkiWMHDmS0aNHA3D66afz/PPPN60/4YQTAJg0aVLTAH8tefHFF/na174G5B4W/JprrmHt2rWUlJQwefJkbrnlFmbPns2bb75Jr169Wj12PrwGAUTiZQAk6ztvIg7nup1WvukX0owZM7joootYsGABNTU1TJo0iX/+859cddVVvPbaa/Tr149Zs2ZRV1fXruPPmjWLBx54gHHjxnHrrbfy7LPPblO8jUOGb8tw4Z01LLjXIIBo2Emd8hqEc9udyspKDj30UM4888ym2sP69eupqKigT58+fPLJJzzyyCOtHuPggw/mgQceoLa2lurqav785z83rauurmannXYimUw2DdEN0KtXL6qrq7c41pgxY1i2bBnvvvsuAL///e855JBD2nVtxR4W3GsQQDQW1CBSDe37huGcK66ZM2dy/PHHNzU1NQ6PvcceezBs2DCmTp3a6v4TJ07k5JNPZty4ceywww5Mnjy5ad1Pf/pTpkyZwqBBg5gyZUpTUjjllFM4++yzueaaa5o6pwESiQS33HILJ510EqlUismTJ3Puuee267oa58oeO3Ys5eXlmw0L/swzzxCJRNhrr72YPn06c+fO5Ze//CWxWIzKykpuv/32dp0zW8GG+y6G9g73/fbLD7PHozN54/O/Z+xBx7a9g3MO8OG+tzdbO9y3NzEB0bAPIt3gTUzOOdfIEwQQi5cDkG7waUedc65RwRKEpGGSnpG0WNIiSVvMk6fANZLelfSGpIlZ606X9E74OL1QcQLEEsFtrpmk1yCc21rdqZm6O2vP+1TITuoU8O9mtkBSL2C+pCfMLPtm5OnAqPAxBbgemCKpP3A5UAVYuO+DZvZZIQKNlQY1iEzSO6md2xqJRILVq1czYMAAwhkgXRdkZqxevZpE+GU4X4WcUe4j4KNwuVrSW8AQIDtBzAButyC1vSypr6SdgGnAE2a2BkDSE8BRwF2FiDVeGvRBmCcI57bK0KFDWb58OatWrSp2KK4NiUSCoUOHbtU+nXKbq6QRwATglWarhgAfZL1eHpa1VJ7r2OcA5wAMHz68XfHFE2GCSHmCcG5rxGIxRo4cWewwXIEUvJNaUiVwL3Chma3v6OOb2RwzqzKzqkGDBrXrGI01CFLeSe2cc40KmiAkxQiSwx1mdl+OTVYAw7JeDw3LWioviJJ4Y4LwGoRzzjUq5F1MAm4C3jKz/2phsweB08K7mfYD1oV9F48BR0rqJ6kfcGRYVqhgqbeY1yCccy5LIfsgpgJfA96UtDAs+wEwHMDMbgAeBo4G3gVqgDPCdWsk/RR4LdzvisYO60JpkCcI55zLVsi7mF4EWr3vLbx76fwW1t0M3FyA0HJqIE4k7QnCOeca+S+pQ0nFkCcI55xr4gkilFScSKah2GE451yX4QkilFKcqNcgnHOuiSeIUCoSJ5rxBOGcc408QYRSihP1JibnnGviCSKUjsQpMU8QzjnXyBNEKB0tpcRrEM4518QTRCgTKfUahHPOZfEEEcpES4lbsthhOOdcl+EJImTRODG8BuGcc408QYSsxGsQzjmXzRNEo2gppTT4/LrOORfyBBGykgQJJUmmMsUOxTnnugRPECGVBJN51zf4pEHOOQeeIDaJlQLQUFdb5ECcc65r8AQRioQ1iIb6miJH4pxzXUPBJgySdDNwDLDSzPbOsf4S4NSsOD4HDApnk1sGVANpIGVmVYWKsymeWJggvAbhnHNAYWsQtwJHtbTSzH5pZuPNbDzwfeC5ZtOKHhquL3hyAIiGCSJV7wnCOeeggAnCzJ4H8p1HeiZwV6FiyUckHiSIZIMnCOecgy7QByGpnKCmcW9WsQGPS5ov6Zw29j9H0jxJ81atWtXuOKKxMsBrEM4516joCQL4EvDXZs1LB5rZRGA6cL6kg1va2czmmFmVmVUNGjSo3UGUlAY1iLTXIJxzDsgjQUj6tqTeCtwkaYGkIzswhlNo1rxkZivC55XA/cC+HXi+nEriYQ2iwWeVc845yK8GcaaZrQeOBPoBXwN+3hEnl9QHOAT4f1llFZJ6NS6H5/17R5yvNSWlQYJIJ70G4ZxzkN9trgqfjwZ+b2aLJKm1HQAk3QVMAwZKWg5cDsQAzOyGcLPjgcfNbGPWroOB+8NTlAB3mtmjecS5TWLxcgAy/ktq55wD8ksQ8yU9DowEvh9+u29zwCIzm5nHNrcS3A6bXbYUGJdHXB0qFvZBZJKeIJxzDvJLEF8HxgNLzaxGUn/gjMKG1fliiaCJybyJyTnngPz6IPYHlpjZWklfBX4ErCtsWJ0vXho0MVnKO6mdcw7ySxDXAzWSxgH/DrwH3F7QqIog3lSD8CYm55yD/BJEyoJZdGYAvzGza4FehQ2r8zX+UI6UJwjnnIP8+iCqJX2f4PbWgyRFCO9G6laiJaQsAmmfl9o55yC/GsTJQD3B7yE+BoYCvyxoVEXSoDjyGoRzzgF5JIgwKdwB9JF0DFBnZt2uDwKggRhKeye1c85BfkNtfAV4FTgJ+ArwiqQTCx1YMSQVI+IJwjnngPz6IH4ITA7HRULSIOBJ4J5CBlYMScU9QTjnXCifPohIY3IIrc5zv+1OSnEiGe+kds45yK8G8aikx9g04urJwMOFC6l4UooT9RqEc84BeSQIM7tE0peBqWHRHDO7v7BhFUcq4jUI55xrlE8NAjO7l81nfOuW0pFSSnyoDeecA1pJEJKqCab+3GIVYGbWu2BRFUk6EqckU13sMJxzrktoMUGYWbcbTqMtmWgpMfMmJuecgwLejSTpZkkrJeWcDU7SNEnrJC0MHz/OWneUpCWS3pV0aaFibC4TjXuCcM65UCFvV70VOKqNbV4ws/Hh4woASVHgWmA6sCcwU9KeBYyziUVLiZPsjFM551yXV7AEYWbPA2vaseu+wLtmttTMGoC5BCPJFpxFS4l7DcI554D8htr4lqR+BTr//pL+JukRSXuFZUOAD7K2WR6WtRTfOZLmSZq3atWqbQqmsQYRjG7unHM9Wz41iMHAa5L+GPYNqIPOvQDYxczGAb8GHmjPQcxsjplVmVnVoEGDti2ikgRxkjSk25xy2znnur18RnP9ETAKuAmYBbwj6T8k7bYtJzaz9Wa2IVx+GIhJGgisAIZlbTo0LCu8klLiSlPf4P0QzjmXVx9EOKPcx+EjBfQD7pH0n+09saQdG2sjkvYNY1kNvAaMkjRSUhw4BXiwvefZKrEEAPV1tZ1yOuec68ra/CW1pG8DpwGfAjcCl5hZMpxZ7h3guy3sdxcwDRgoaTlwOeFMdGZ2A3AicJ6kFFALnBImopSkbwKPAVHgZjNbtE1XmadISZAgGuprCHKgc871XPkMtdEfOMHM3s8uNLNMOIFQTmY2s7WDmtlvgN+0sO5hijAgoGKlACTrvQbhnHP5DNZ3uaSJkmYQDL3xVzNbEK57q9ABdqZIrAyApDcxOedcXre5XgbcBgwABgK3SPpRoQMrhmg8aGJKeQ3COefyamL6KjDOzOoAJP0cWAj8rJCBFUMkTBDJBk8QzjmXz11MHwKJrNeldNZtp52sJB40MaUb6oociXPOFV8+NYh1wCJJTxD0QRwBvCrpGgAzu6CA8XWqaJggUl6DcM65vBLE/eGj0bOFCaX4YqVBRSnjNQjnnMvrLqbbwh+sjQ6LlphZt/ypcayxiSnpCcI55/L5odw0gruYlhHMJjdM0unhaK3dSqw0SBCZpDcxOedcPk1MvwKONLMlAJJGA3cBkwoZWDHESssBMK9BOOdcXncxxRqTA4CZ/YNwyIzuJp4IahCWrC9yJM45V3z51CDmS7oR+EP4+lRgXuFCKp542MRkKa9BOOdcPgniXOB8oPF21heA6woWURE1DrVhKa9BOOdcqwkinB/6b2a2B/BfnRNSEZUEg/XJE4RzzrXeB2FmaWCJpOGdFE9xSdQTA29ics65vJqY+hH8kvpVYGNjoZkdW7CoiihJDKW9BuGcc/kkiMvac2BJNwPHACvNbO8c608Fvkfw24pq4Dwz+1u4bllYlgZSZlbVnhjaI6kYeBOTc87llSCONrPvZRdI+gXwXBv73UowIdDtLaz/J3CImX0maTowB5iStf5QM/s0j/g6VINKiaS9ick55/L5HcQROcqmt7VT+EvrNa2sf8nMPgtfvgwMzSOWgktGEkRTNcUOwznniq7FBCHpPElvAmMkvZH1+CfwZgfH8XXgkazXBjwuab6kc1rbUdI5kuZJmrdq1aptDqQhWk4s7UNtOOdca01MdxJ8aP9f4NKs8moza7FmsLUkHUqQIA7MKj7QzFZI2gF4QtLbLY39ZGZzCJqnqKqqsm2NJxUtJ+5jMTnnXMs1CDNbZ2bLzGwmsBxIEnyzr+yo214ljQVuBGaY2eqsc68In1cSDDW+b0ecLx+paBmlGe+DcM65fEZz/SYwG/gEyITFBozdlhOHSeY+4Gvh+E6N5RVAxMyqw+UjgSu25VxbIxMrp8K8BuGcc/ncxXQhMCb7G34+JN0FTAMGSloOXE44yJ+Z3QD8GBgAXCcJNt3OOhi4PywrAe40s0e35tzbIhOroIw6MhkjElFnndY557qcfBLEBwTTjm6VsGmqtfVnAWflKF8KjNva83WYWDll1FOXSlMez+fP45xz3VM+n4BLgWclPQQ0/YLMzLrn2EzxCiqoY3VdyhOEc65Hy+cT8F/hIx4+ujXFK4jIqK2pht6JYofjnHNFk8+c1D9pXiap2361jiYqAair2QAMKm4wzjlXRK39UO7FrOXfN1v9asEiKrJIohcA9TXrixyJc84VV2tDbVRkLTcfbK/b3t4TSwSXXV+zociROOdccbWWIKyF5Vyvu41YojcAyVqvQTjnerbW+hL6SjqeIIn0lXRCWC6gT8EjK5J4edDElKr1GoRzrmdrLUE8BxybtfylrHU5x0XqDkobE0S9JwjnXM/WYoIwszM6M5CuIlERJIhMnScI51zPls98ED1KaVmYIBo2trGlc851b54gmomEv4Og3hOEc65n8wTRXCy8uzfpCcI517O1mSAknSSpV7j8I0n3SZpY+NCKJFpCPTGU9GlHnXM9Wz41iMvCuRkOBA4HbgKuL2xYxVWnMiJeg3DO9XD5JIh0+PxFYI6ZPUQ3H7SvQQlK0l6DcM71bPkkiBWSfgucDDwsqTTP/ZB0s6SVkv7ewnpJukbSu5LeyG66knS6pHfCx+n5nK+j1EfKKEn5rHLOuZ4tnw/6rwCPAV8ws7VAf+CSPI9/K3BUK+unA6PCxzmETVeS+hPMQDeFYD7qyyX1y/Oc2ywZLSOW8RqEc65nyydB7AQ8ZGbvSJoGnESeo7ma2fPAmlY2mQHcboGXCYb02An4AvCEma0xs8+AJ2g90XSoVLSMeNprEM65ni2fBHEvkJa0OzAHGAbc2UHnH0IwpWmj5WFZS+WdIl1STqnVddbpnHOuS8onQWTMLAWcAPzazC4hqFV0CZLOkTRP0rxVq1Z1yDHTsQoSniCccz1cPgkiKWkmcBrwl7As1kHnX0FQI2k0NCxrqXwLZjbHzKrMrGrQoI6ZAc5i5ZRRRyqd6ZDjOefc9iifBHEGsD9wpZn9U9JIoPkMc+31IHBaeDfTfsA6M/uIoFP8SEn9ws7pI8OyTmGxCsqppyaZbntj55zrpvKZk3qxpO8AoyXtDSwxs1/kc3BJdwHTgIGSlhPcmRQLj3sD8DBwNPAuUEOQjDCzNZJ+CrwWHuoKM2uts7tDKV5OBXWsrEvRO9FRlSXnnNu+tJkgwjuXbgOWEUwWNEzS6eEdSq0ys5ltrDfg/BbW3Qzc3NY5CiFSWklERk3tBuhbVowQnHOu6NpMEMCvgCPNbAmApNHAXcCkQgZWTJHSYETXug3VQMf0azjn3PYmnz6IWGNyADCzf9BxndRdUkk45HdDjc9L7ZzrufKpQcyXdCPwh/D1qcC8woVUfNFw0qAGn5faOdeD5ZMgziXoJ7ggfP0CcF3BIuoCYmv7rngAABVrSURBVI01iDqvQTjneq5WE4SkKPA3M9sD+K/OCan44uVBgkjV+pDfzrmeq9U+CDNLA0skDe+keLqERHlvANJ11UWOxDnniiefJqZ+wCJJrwJNX6nN7NiCRVVkiYqgDyJT730QzrmeK58EcVnBo+hiYokwQTR4E5NzrudqMUGEo7cONrPnmpUfCHxU6MCKSeHvIKj3BOGc67la64O4Gsh1G8+6cF33FasAQD4vtXOuB2stQQw2szebF4ZlIwoWUVcQLaGBGCR9VjnnXM/VWoLo28q6bj9AUZ3KiHoNwjnXg7WWIOZJOrt5oaSzgPmFC6lrqI8kiPq0o865Hqy1u5guBO6XdCqbEkIVEAeOL3RgxdYQKSPmCcI514O1mCDM7BPgAEmHAnuHxQ+Z2dOdElmRJSNlxFLeB+Gc67nymTDoGeCZToilS0mVlBFv8HmpnXM9Vz7DfbebpKMkLZH0rqRLc6z/b0kLw8c/JK3NWpfOWvdgIePMJV1SQal5E5NzrufK55fU7RIO9HctcASwHHhN0oNmtrhxGzO7KGv7bwETsg5Ra2bjCxVfWzIlZZRZHWaGpGKF4ZxzRVPIGsS+wLtmttTMGoC5wIxWtp9JMFNdl5CJVVBGHfWpTLFDcc65oihkghgCfJD1enlYtgVJuwAjgewO8ISkeZJelnRcSyeRdE643bxVq1Z1RNzBcePllFNPbUO6w47pnHPbk4L2QWyFU4B7wuHFG+1iZlXAvwFXS9ot145mNsfMqsysatCgDpw/Ol5JOXVsrE923DGdc247UsgEsQIYlvV6aFiWyyk0a14ysxXh81LgWTbvnyg4lVYSlVHrkwY553qoQiaI14BRkkZKihMkgS3uRpK0B8GcE/+bVdZPUmm4PBCYCixuvm8hRUuDAfvqNvi0o865nqlgdzGZWUrSN4HHgChws5ktknQFMM/MGpPFKcBcM7Os3T8H/FZShiCJ/Tz77qfOEA3npa6v9UmDnHM9U8ESBICZPQw83Kzsx81ez86x30vAPoWMrS2xsmDSoIYar0E453qmrtJJ3eXEwhpEqs5rEM65nskTRAvi5b0BSHoTk3Ouh/IE0YLSiqAGka6vLnIkzjlXHJ4gWpAoC2oQmTof0dU51zN5gmhBSdhJbQ3exOSc65k8QbQkVh48N/gP5ZxzPZMniJbEgx/K4fNSO+d6KE8QLYnGaCDG6jVrmPvqv9hYnyp2RM4516kK+kO57V68nN7WwAX3vclP/7KYL+y9I4eMHsRBowbRvyJe7Oicc66gPEG0Il7Wmy+N6MOQiftz5ysf8ORbn3DfghVIcOqU4fzsuKL+2Ns55wrKE0RrYuXo478zqfIzJn1lHOmM8cbytfz2uaXc+cq/uOjw0QyoLC12lM45VxDeB9Ga8TPh0yXwmyr40xlE33+RCTsl+OZhu5MxePrtlcWO0DnnCsZrEK058CIYNxNevg5euxkW3QeRGHvtPJ7TKw/gicWDOalqWNvHcc657ZDXINrSa0c44gq4eDHMnAv7n49q1/LDzG957Z0V1CV9SlLnXPfkCSJfid4wZjoc8RM4+pfEM3VMTr/OX9/9tNiROedcQRQ0QUg6StISSe9KujTH+lmSVklaGD7Oylp3uqR3wsfphYxzq404EEv05ZjYfJ5Y/Emxo3HOuYIoWIKQFAWuBaYDewIzJe2ZY9O7zWx8+Lgx3Lc/cDkwBdgXuFxSv0LFutWiMTTmaA6PLuCZxR+SyVjb+zjn3HamkDWIfYF3zWypmTUAc4EZee77BeAJM1tjZp8BTwBHFSjO9vncMZRnNjCqdiGvf7C22NE451yHK2SCGAJ8kPV6eVjW3JclvSHpHkmNtwTluy+SzpE0T9K8VatWdUTc+dntMCxWzvToazz5ljczOee6n2J3Uv8ZGGFmYwlqCbdt7QHMbI6ZVZlZ1aBBgzo8wBbFytCoIzg6toCH/racmgYfq8k5170UMkGsALJ/JDA0LGtiZqvNrD58eSMwKd99u4Q9vkS/zBp2WPcmF85d6H0RzrlupZAJ4jVglKSRkuLAKcCD2RtI2inr5bHAW+HyY8CRkvqFndNHhmVdy+gjIRJj9qilPL74E37x6NvFjsg55zpMwX5JbWYpSd8k+GCPAjeb2SJJVwDzzOxB4AJJxwIpYA0wK9x3jaSfEiQZgCvMbE2hYm23RB/YdRp7ffwY5088jmufX8rOfcs4dcpwSqLFbr1zzrltI7Pu0yxSVVVl8+bN69yTrlgAt8/Ayvrx7+U/476lUXqVlrD/bgM4cNRAJo/oz+jBvYhG1LlxOedcHiTNN7OqnOs8QXSA5fPhD8dj8V48s/8tPPFRghfe+ZTln9UC0Ku0hH1H9uesg3Zl/90GdH58zjnXAk8QneHDhfD74yCdhCGTsJ0nsrrPniyoH8Lzqyp4/K1PWVldzwG7DeDiI0ZTNaJ/ceJ0zrksniA6y6ol8PL18OEC+GQRZMJbX0vKyAzem7/2PppL/zGaFRvFr04ax5cnDS1erM45R+sJwof77kiDxsCXrg6Wk3WwchF8shhWLiay9FkOeusnvFjWn/vLjuKy+9LsMqDcaxLOuS7LE0ShxBIwZFLwADCDZS+gl2/ghCV3MjCxlP9ze5x7v3kww/qXFzdW55zLwe/F7CwSjDwYZt4JR/2Cg9Mvc1nmWs669RXmv/+Z/8jOOdfleA2iGPY7F+qr+dIzP2P92gT/dv2/0adXL47YczB77tyb4f3LGdavnMG9E5TFo8WO1jnXQ3mCKJaDvwP16zj1pV9zSvkzLIuM4rnXd+W9ef1YYBWso4KNJLCSBInyCsoSFZSVJUiUlVOeKKUsUUZZIkFZIkF5Ik55aZzyeJSyeJSKeAnl8SiJWPA6EYtSFov6bzGcc1vFE0SxSHDET2HkIUT/+Ty7ffAKu374OIo0bLltXfhoZVTxjIkUEdJESRMhQ4R0+KghQnVYZkQwbXpABCRMUVAkeEQiEL5W+FqKoEhYFoluekRLULSESKyUaKyMaGk55f13Ij5gJPQbAYP3glhZYf6GzrmC8gRRTBKMOiJ4AMqkoX491H4GtWshWQPJ2uA51QDpBkjXQzoFmSTpVAPJhiQNyQaSyQZSySSpVJJ0KkkqnSaTSpHOpMLnDJZJkUmnyYTLZNJkMoZZGjJpzDKQyUA6jZmBZZClwDJgFqaY4BElQ5SgLEaKOClKlSRGPXHVNl1iQ7QC2+MYSifOhBEHQdT/yTm3vfD/rV1JJApl/YJHHqLhI1HQoAJmRipjNKQyNKQy1Kcy1KfS1KcybEymWVmfZmN9ig31KT5e9SnrPnqPhpXvMnrdXzn67w9Suuhu0pE41n83SgaNgr7DIdE3GM8q0QfiFVBaCfFKKEkEtY5YWbBcUgrRUk8uznUy/x/n8iKJWFTEohEqStvaeggwDoAlH1fzq5eWsHbhX9gj9Q67ffIhY1a/xmAep9TqtjKICETjEIkFySISg2gMIiXNHtHcr6OxYP+SRJCQ4hXQewgMHA0DR0HfXYLmNecc4L+kdp1kY32KN5av480Va3lzxXreX72R1euqSW74jApqKKeeCmopVx0JkpRRT3kkSe9Yml4laSqjacqjGRKRNOWRFPGIURpJE1eGmDKUKE2MNCUySpSmxNJhM1jQExO1NFFLEbEkkXQdkWQNqt8AyY2bgizrD7tOg90OhSFV0HcYlPYq1p9s+2EWPjJAuNzqc6ZZGXnsk+uZ3GVtbpPHdtnraXxq7fw5llu8rlz7sRXHa76tBTXs0UfSHv5Lald0FeEIt80HK0ylM6ypaeCzjUnWbGzgs5oG1tUmWVebZH1tkg31KT6pS7G+LkVNQ4qNDUFTVm1tmtpkmpqGFHXJzFbHI0FlvIQ9ejcwpc8axpV+zF6pxey47K9EFt23acNEX6gcvKnGESvfvNYiAQqfQ7k+FCyz+Qdo44dkY/kWj+b75NqGHOtz7UeOsmYf6E3LzbfN44PfFV/FDnDJOx1+WE8QrqhKohF26JVgh17t70kxM+pTGWob0tSl0tQlM9Qlg/6RumQ6fATLtckgwayvS7G+Nsnyz2r4y8pKrlszkIztjXQSRw1ay0F9VzEiuoad9Cm902tJZGqIJWspqVuPMsnghoJMcvMPy+wk0ZQ0tOnuMGUtN65vKo9uuRyJgko2bR/eRbbFMSFY13TMaAvnVrP91cJy1nE3W5fjufm1IFDz62/+3MYxoY31Wc9tbksr29DGsWh528ZrzN4n53ILf4s291PTrjmP1/y6ojEKoaAJQtJRwP8Q9KXeaGY/b7b+YuAsggmDVgFnmtn74bo08Ga46b/M7NhCxuq2X5JIxILfe7RXXTLNG8vX8fLS1fzve6v51YrBrN6Y45bjUFksSiIWIV4SPqIRYtEIJVFREolQEhHRiCiJimgkQlQQjYiIgvJI47IgIiGJaCTYRhKRsDxYR9N+aioHET5vtt2m15Hw86jx+CJYH8k6R+MxIgo+9LL3Ewrzypbnazou4V3R4XLzbcnarmlds/0bl2HTeSORTeePhB+SzffZFN/m51Xz+KHp7yMU5Ce23Idmx93sXJsl/56jYH0QkqLAP4AjgOUEs8PNNLPFWdscCrxiZjWSzgOmmdnJ4boNZla5Nef0PgjXkTbWp1j+WS0rq+tYV5tkbU2S6roUtck0tWHTVkMqQ0M6eE6mM6QyRjKdIWNGMm2kM5s/MhYum2FG02szNis323ybTMbINC5nlRvNmqVdQeVKOI0ViuZJSWplmcYKQXZ5C/s3T2JZx2g894CKUv547v7tvKbi9EHsC7xrZkvDIOYCM4CmBGFmz2Rt/zLw1QLG49xWqSgtYcyOvRizY9fuqLbsRGJgbEo4Fq7PZLLKs5JQ4/YZo2k8sKbj2Kb9g+OzeXKyLY8JjftuSnJNXRtZSS3YvzGO5slu0/mMTecj+xhNx9w8jux9LOuYLe3TGGt41k37NT9WePLG42WXb36uZuU5Ymm+L2z6O7R23GBLNr+OsLB3WWE+yguZIIYAH2S9Xg5MaWX7rwOPZL1OSJpH0Pz0czN7oONDdG7719RkQ89sBnGF0yU6qSV9FagCDskq3sXMVkjaFXha0ptm9l6Ofc8BzgEYPnx4p8TrnHM9QSF/FbQCGJb1emhYthlJhwM/BI41s/rGcjNbET4vBZ4FJuQ6iZnNMbMqM6saNGhQx0XvnHM9XCETxGvAKEkjJcWBU4AHszeQNAH4LUFyWJlV3k9Sabg8EJhKVt+Fc865witYE5OZpSR9E3iM4DbXm81skaQrgHlm9iDwS6AS+FN4G1nj7ayfA34rKUOQxH6effeTc865wvOhNpxzrgdr7TZXH5nMOedcTp4gnHPO5eQJwjnnXE7dqg9C0irg/XbuPhD4tAPD2R70xGuGnnndPfGaoWde99Ze8y5mlvM3At0qQWwLSfNa6qjprnriNUPPvO6eeM3QM6+7I6/Zm5icc87l5AnCOedcTp4gNplT7ACKoCdeM/TM6+6J1ww987o77Jq9D8I551xOXoNwzjmXkycI55xzOfX4BCHpKElLJL0r6dJix1MokoZJekbSYkmLJH07LO8v6QlJ74TP/Yoda0eTFJX0uqS/hK9HSnolfM/vDkcb7lYk9ZV0j6S3Jb0laf/u/l5Luij8t/13SXdJSnTH91rSzZJWSvp7VlnO91aBa8Lrf0PSxK05V49OEOG82dcC04E9gZmS9ixuVAWTAv7dzPYE9gPOD6/1UuApMxsFPBW+7m6+DbyV9foXwH+b2e7AZwSzGXY3/wM8amZ7AOMIrr/bvteShgAXAFVmtjfBCNKn0D3f61uBo5qVtfTeTgdGhY9zgOu35kQ9OkGQNW+2mTUAjfNmdztm9pGZLQiXqwk+MIYQXO9t4Wa3AccVJ8LCkDQU+CJwY/hawGHAPeEm3fGa+wAHAzcBmFmDma2lm7/XBNMXlEkqAcqBj+iG77WZPQ+saVbc0ns7A7jdAi8DfSXtlO+5enqCyDVv9pAixdJpJI0gmKHvFWCwmX0UrvoYGFyksArlauC7QCZ8PQBYa2ap8HV3fM9HAquAW8KmtRslVdCN3+twBsqrgH8RJIZ1wHy6/3vdqKX3dps+43p6guhxJFUC9wIXmtn67HUW3PPcbe57lnQMsNLM5hc7lk5WAkwErjezCcBGmjUndcP3uh/Bt+WRwM5ABVs2w/QIHfne9vQEkde82d2FpBhBcrjDzO4Liz9prHKGzytb2n87NBU4VtIygubDwwja5vuGzRDQPd/z5cByM3slfH0PQcLozu/14cA/zWyVmSWB+wje/+7+Xjdq6b3dps+4np4g2pw3u7sI295vAt4ys//KWvUgcHq4fDrw/zo7tkIxs++b2VAzG0Hw3j5tZqcCzwAnhpt1q2sGMLOPgQ8kjQmLPk8wp3u3fa8Jmpb2k1Qe/ltvvOZu/V5naem9fRA4LbybaT9gXVZTVJt6/C+pJR1N0E7dOG/2lUUOqSAkHQi8ALzJpvb4HxD0Q/wRGE4wVPpXzKx5B9h2T9I04DtmdoykXQlqFP2B14Gvmll9MePraJLGE3TMx4GlwBkEXwi77Xst6SfAyQR37L0OnEXQ3t6t3mtJdwHTCIb1/gS4HHiAHO9tmCx/Q9DcVgOcYWZ5z8vc4xOEc8653Hp6E5NzzrkWeIJwzjmXkycI55xzOXmCcM45l5MnCOecczl5gnCuDZLSkhZmPTpskDtJI7JH5XSuKylpexPnerxaMxtf7CCc62xeg3CunSQtk/Sfkt6U9Kqk3cPyEZKeDsfff0rS8LB8sKT7Jf0tfBwQHioq6XfhXAaPSyoLt79Awfwdb0iaW6TLdD2YJwjn2lbWrInp5Kx168xsH4Jfq14dlv0auM3MxgJ3ANeE5dcAz5nZOIKxkRaF5aOAa81sL2At8OWw/FJgQniccwt1cc61xH9J7VwbJG0ws8oc5cuAw8xsaTgQ4sdmNkDSp8BOZpYMyz8ys4GSVgFDs4d6CIdefyKc6AVJ3wNiZvYzSY8CGwiGUXjAzDYU+FKd24zXIJzbNtbC8tbIHhsozaa+wS8SzHg4EXgta1RS5zqFJwjnts3JWc//Gy6/RDB6LMCpBIMkQjAV5HnQNE92n5YOKikCDDOzZ4DvAX2ALWoxzhWSfyNxrm1lkhZmvX7UzBpvde0n6Q2CWsDMsOxbBLO5XUIws9sZYfm3gTmSvk5QUziPYPazXKLAH8IkIuCacNpQ5zqN90E4105hH0SVmX1a7FicKwRvYnLOOZeT1yCcc87l5DUI55xzOXmCcM45l5MnCOecczl5gnDOOZeTJwjnnHM5/X9abdRlHH33RQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N2AcDC1MX3Mn"
      },
      "source": [
        "Tanh"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZueRP3d0heUf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "outputId": "c44bd975-5a91-4c3a-da9a-727636a882d8"
      },
      "source": [
        "tanh_model = MyNeuralNetwork(5,[784,256,128,64,10],'tanh',0.1, 'normal', 100, 100)\n",
        "tanh_model.fit(X_train, y_train)\n",
        "print(\"Tanh\")\n",
        "print(\"Train accuracy:\", tanh_model.score(X_train, y_train))\n",
        "print(\"Validation accuracy:\", tanh_model.score(X_val, y_val))\n",
        "print(\"Test accuracy:\", tanh_model.score(X_test, test_y))\n",
        "tanh_model.plot_loss_epoch(X_train, y_train, X_val, y_val)\n",
        "pickle.dump(tanh_model, open(\"tanh_model\",\"wb\"))"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Tanh\n",
            "Train accuracy: 1.0\n",
            "Validation accuracy: 0.9746666666666667\n",
            "Test accuracy: 0.9733\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zaITGXQ_X7BY"
      },
      "source": [
        "tSNE plot "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_IXU6SETSqUA"
      },
      "source": [
        "from sklearn.manifold import TSNE\n",
        "import seaborn as sns\n",
        "\n",
        "x = relu_model.As[3]\n",
        "tsne = TSNE(n_components=2)\n",
        "x1 = tsne.fit_transform(x)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R30ty8GhUu3_",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        },
        "outputId": "17b14ed4-8171-4226-c55f-21dcc01fd7a4"
      },
      "source": [
        "y = test_y\n",
        "plt.figure(figsize=(11,6))\n",
        "sns.scatterplot(x=x1[:,0],y=x1[:,1],  hue=y, palette=sns.color_palette(\"hls\", 10))\n",
        "plt.show()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 792x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Fr2YSulhX_V6"
      },
      "source": [
        "MLP Classifier"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sqpnQNWRLz-o",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dca39f50-46f7-4319-ec19-c0a4aac2d04d"
      },
      "source": [
        "from sklearn.neural_network import MLPClassifier\n",
        "\n",
        "MLP_clf = MLPClassifier(hidden_layer_sizes=(256,128,64,), activation='relu', batch_size=100, max_iter=100)\n",
        "MLP_clf.fit(X_train, y_train)\n",
        "print(\"ReLU: Test accuracy: \", MLP_clf.score(X_test, test_y))\n",
        "\n",
        "MLP_clf = MLPClassifier(hidden_layer_sizes=(256,128,64,), activation='logistic', batch_size=100, max_iter=100)\n",
        "MLP_clf.fit(X_train, y_train)\n",
        "print(\"Sigmoid: Test accuracy: \", MLP_clf.score(X_test, test_y))\n",
        "\n",
        "\n",
        "MLP_clf = MLPClassifier(hidden_layer_sizes=(256,128,64,), activation='identity', batch_size=100, max_iter=100)\n",
        "MLP_clf.fit(X_train, y_train)\n",
        "print(\"Linear: Test accuracy: \", MLP_clf.score(X_test, test_y))\n",
        "\n",
        "\n",
        "MLP_clf = MLPClassifier(hidden_layer_sizes=(256,128,64,), activation='tanh', batch_size=100, max_iter=100)\n",
        "MLP_clf.fit(X_train, y_train)\n",
        "print(\"Tanh: Test accuracy: \", MLP_clf.score(X_test, test_y))"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "ReLU: Test accuracy:  0.9808\n",
            "Sigmoid: Test accuracy:  0.9807\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (100) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Linear: Test accuracy:  0.9198\n",
            "Tanh: Test accuracy:  0.9786\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}