tareen_kinney_mlcb_2019.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from __future__ import division\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from numpy import argmax\n",
    "from scipy import constants\n",
    "import random\n",
    "\n",
    "import os\n",
    "import sys\n",
    "import re\n",
    "import pdb\n",
    "import glob\n",
    "\n",
    "import suftware\n",
    "\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "import keras\n",
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Dense, Dropout, Activation, Input, Lambda, Concatenate\n",
    "from keras.optimizers import Adam\n",
    "from keras import metrics\n",
    "from keras import regularizers\n",
    "from keras import callbacks\n",
    "import keras.backend as K\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow.python.framework import ops\n",
    "from tensorflow.python.ops import math_ops\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.patches as patches\n",
    "from matplotlib import cm\n",
    "from matplotlib.colors import ListedColormap, LinearSegmentedColormap\n",
    "%matplotlib inline\n",
    "plt.ion()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " Load the energy matrices from Kinney et al. 2010. These will be used later for comparison"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "crp_df = pd.read_csv('data/energy_matrices/crp_tau_final_all.26.txt', sep = \"\\t\", usecols = range(1,5))\n",
    "rnap_df = pd.read_csv('data/energy_matrices/rnap_tau_final_all.41.txt', sep = \"\\t\", usecols = range(1,5))\n",
    "\n",
    "# Convert the dataframe to numpy array for downstream processing\n",
    "crp_array = pd.DataFrame.to_numpy(crp_df)\n",
    "rnap_array = pd.DataFrame.to_numpy(rnap_df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Helper function for one-hot encoding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Fit a label encoder and a onehot encoder\n",
    "bases = [\"A\",\"C\",\"G\",\"T\"]\n",
    "label_encoder = LabelEncoder()\n",
    "label_encoder.fit(bases)\n",
    "tmp = label_encoder.transform(bases)\n",
    "tmp = tmp.reshape(len(tmp), 1)\n",
    "onehot_encoder = OneHotEncoder(sparse = False)\n",
    "onehot_encoder.fit(tmp)\n",
    "\n",
    "# Encode sequence into onehot\n",
    "def onehot_sequence(sequence, lab_encoder = label_encoder, one_encoder = onehot_encoder):\n",
    "    \"\"\"Sequence as a string\"\"\"\n",
    "    tmp = lab_encoder.transform(list(sequence))\n",
    "    tmp = tmp.reshape(len(tmp),1)\n",
    "    tmp = one_encoder.transform(tmp)\n",
    "    return tmp"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Input Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_data_full_wt = pd.read_csv('data/full-wt/data.txt',delim_whitespace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Put data into lists and feed them to sklearn's train_test_split method"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sequences = []\n",
    "bin_number = []\n",
    "\n",
    "for _ in range(len(input_data_full_wt)):\n",
    "    sequences.append(input_data_full_wt['seq'][_])\n",
    "    bin_number.append(list(input_data_full_wt.iloc[_,1:len(input_data_full_wt.columns)].values))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(sequences, bin_number, test_size=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "######\n",
    "## Define custom function for keras\n",
    "######\n",
    "\n",
    "def log_poisson_loss(y_true, y_pred):\n",
    "    return tf.nn.log_poisson_loss(y_true,y_pred)\n",
    "\n",
    "def poisson_loss(y_true, y_pred):\n",
    "    y_pred = ops.convert_to_tensor(y_pred)\n",
    "    y_true = math_ops.cast(y_true, y_pred.dtype)\n",
    "    return K.mean(y_pred - y_true * math_ops.log(y_pred + K.epsilon()), axis=-1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "One-hot encode the sequences at the location of CRP and RNAP binding sites"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_seqs_ohe = []\n",
    "for _ in range(len(X_train)):\n",
    "    input_seqs_ohe.append(onehot_sequence(X_train[_][1:27]+X_train[_][34:75]))\n",
    "    \n",
    "test_seqs_ohe = []    \n",
    "for _ in range(len(X_test)):\n",
    "    test_seqs_ohe.append(onehot_sequence(X_test[_][1:27]+X_test[_][34:75]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Shapes of the training and test sets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(40414, 67, 4)\n",
      "(10104, 67, 4)\n"
     ]
    }
   ],
   "source": [
    "print(np.shape(input_seqs_ohe))\n",
    "print(np.shape(test_seqs_ohe))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras.layers import multiply\n",
    "\n",
    "class CustomConnected(Dense):\n",
    "\n",
    "    def __init__(self,units,connections, output_dim,**kwargs):\n",
    "        \n",
    "        self.output_dim = output_dim\n",
    "        #this is matrix of connections. I.e., this should be [0,0,trainable,trainable]\n",
    "        self.connections = connections                        \n",
    "        #initalize the original Dense with all the usual arguments   \n",
    "        super(CustomConnected,self).__init__(units,**kwargs)  \n",
    "\n",
    "    def call(self, inputs):\n",
    "        #print(self.connections)\n",
    "        \n",
    "        kernel_times_connections = multiply([self.kernel, self.connections])\n",
    "        #output = K.dot(inputs, self.kernel * self.connections)\n",
    "        \n",
    "        # for implementing custom constraint, print out kernel. Set last two elements equal to each other.\n",
    "        output = K.dot(inputs,kernel_times_connections)\n",
    "        \n",
    "        if self.use_bias:\n",
    "            output = K.bias_add(output, self.bias)\n",
    "        if self.activation is not None:\n",
    "            output = self.activation(output)\n",
    "        return output\n",
    "    \n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return (input_shape[0], self.output_dim)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model definitions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Variable *= will be deprecated. Use `var.assign(var * other)` if you want assignment to the variable value or `x = x * y` if you want a new python Tensor object.\n"
     ]
    }
   ],
   "source": [
    "number_input_layer_nodes = len(input_seqs_ohe[0])*4 +1   \n",
    "inputTensor = Input((number_input_layer_nodes,),name='Sequence')\n",
    "\n",
    "# define two groups which represent binding sites for the two proteins crp and rnap\n",
    "group1 = Lambda(lambda x: x[:,0:104], output_shape=((104,)))(inputTensor)\n",
    "group2 = Lambda(lambda x: x[:,104:268], output_shape=((164,)))(inputTensor)\n",
    "group3 = Lambda(lambda x: x[:,268:269], output_shape=((1,)))(inputTensor)\n",
    "\n",
    "\n",
    "# From input to first hidden layer. The groups are connected to two nodes\n",
    "# node group1 represents energy of crp and group2 represents rnap energy.\n",
    "\n",
    "# rename group1 to crp pwm\n",
    "\n",
    "group1 = Dense(1,use_bias=True,name='CRP_pwm')(group1)\n",
    "group2 = Dense(1,use_bias=True,name='RNAP_pwm')(group2)\n",
    "group3 = Dense(1,use_bias=False,name='one-feature')(group3)\n",
    "\n",
    "# the groups from the previous layer have to be concatenated:\n",
    "concatenateLayer = Concatenate()([group1,group2,group3])\n",
    "\n",
    "# gammaInput = Input(shape=(1,),name='gamma_Input')   # i.e. weights going into softmax: [0,0,0,1]\n",
    "\n",
    "softmaxLayer = Dense(4,activation='softmax', trainable=False, use_bias=False,name='softmax_layer')(concatenateLayer)\n",
    "\n",
    "tauTensor = CustomConnected(1,tf.constant([0.0,0.0,1.0,1.0]),1,use_bias=False,name='custom_tau_layer')(softmaxLayer)\n",
    "intermediateTensor = Dense(10,activation='softmax', trainable=True, use_bias=False,name='softmax_layer_intermediate')(tauTensor)\n",
    "outputTensor = Dense(np.shape(y_train[1])[0],activation='softmax', trainable=True, use_bias=False,name='softmax_layer_bins')(intermediateTensor)\n",
    "\n",
    "\n",
    "#create the model:\n",
    "model = Model(inputTensor,outputTensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "Sequence (InputLayer)           (None, 269)          0                                            \n",
      "__________________________________________________________________________________________________\n",
      "lambda_1 (Lambda)               (None, 104)          0           Sequence[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "lambda_2 (Lambda)               (None, 164)          0           Sequence[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "lambda_3 (Lambda)               (None, 1)            0           Sequence[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "CRP_pwm (Dense)                 (None, 1)            105         lambda_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "RNAP_pwm (Dense)                (None, 1)            165         lambda_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "one-feature (Dense)             (None, 1)            1           lambda_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 3)            0           CRP_pwm[0][0]                    \n",
      "                                                                 RNAP_pwm[0][0]                   \n",
      "                                                                 one-feature[0][0]                \n",
      "__________________________________________________________________________________________________\n",
      "softmax_layer (Dense)           (None, 4)            12          concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "custom_tau_layer (CustomConnect (None, 1)            4           softmax_layer[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "softmax_layer_intermediate (Den (None, 10)           10          custom_tau_layer[0][0]           \n",
      "__________________________________________________________________________________________________\n",
      "softmax_layer_bins (Dense)      (None, 10)           100         softmax_layer_intermediate[0][0] \n",
      "==================================================================================================\n",
      "Total params: 397\n",
      "Trainable params: 385\n",
      "Non-trainable params: 12\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<keras.engine.input_layer.InputLayer at 0x11871b080>,\n",
       " <keras.layers.core.Lambda at 0x118730ef0>,\n",
       " <keras.layers.core.Lambda at 0x118730b70>,\n",
       " <keras.layers.core.Lambda at 0x11870b208>,\n",
       " <keras.layers.core.Dense at 0x11873bd68>,\n",
       " <keras.layers.core.Dense at 0x11873b978>,\n",
       " <keras.layers.core.Dense at 0x11873b898>,\n",
       " <keras.layers.merge.Concatenate at 0x118743e48>,\n",
       " <keras.layers.core.Dense at 0x11874e5c0>,\n",
       " <__main__.CustomConnected at 0x11875ae80>,\n",
       " <keras.layers.core.Dense at 0x11875a470>,\n",
       " <keras.layers.core.Dense at 0x105780748>]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model layer 6 represents the value of $\\Delta G_I$, check its value before training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[-0.6022438]], dtype=float32)]\n"
     ]
    }
   ],
   "source": [
    "print(model.layers[6].get_weights())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Set weights for $\\tau$ layer...\n",
    "... so that only the two microstates (with bound RNAP) contribute to transcription"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(4, 1)"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tau_layer_weights = model.layers[9].get_weights()[0].copy()\n",
    "tau_layer_weights.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.],\n",
       "       [0.],\n",
       "       [1.],\n",
       "       [1.]], dtype=float32)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tau_layer_initial_weights = np.array([0,0,1,1])\n",
    "tau_layer_initial_weights = tau_layer_initial_weights.reshape(4,1)\n",
    "model.layers[9].set_weights([tau_layer_initial_weights])\n",
    "model.layers[9].get_weights()[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The following are the non-trainable and trainable weights in the softmax layer\n",
    "and their values before training."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[-0.6904056 ,  0.5450921 ,  0.64685476, -0.87900007],\n",
      "       [ 0.78645945,  0.36611283,  0.79684687,  0.9018415 ],\n",
      "       [ 0.8125454 ,  0.10814333, -0.5667167 ,  0.81788135]],\n",
      "      dtype=float32)]\n"
     ]
    }
   ],
   "source": [
    "print(model.layers[8].get_weights())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Set softmax layer weights manually, will be kept frozen during training\n",
    "Note that the bias weights here may be interpreted as gamma"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[0., 1., 0., 1.],\n",
       "        [0., 0., 1., 1.],\n",
       "        [0., 0., 0., 1.]], dtype=float32)]"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# the following are the input weights to the softmax array.\n",
    "# the first two arrays indicate 4 weights each coming out of the e_c and \n",
    "# e_r nodes. The last node indicates the gamma input, connecting only to\n",
    "# microstate 4.\n",
    "\n",
    "softmax_weights = np.array([[0,1,0,1],[0,0,1,1],[0,0,0,1]]) # this works\n",
    "\n",
    "# syntax for set_weights expects weights in a list\n",
    "softmax_weights = [softmax_weights]\n",
    "\n",
    "# set weights\n",
    "model.layers[8].set_weights(softmax_weights)\n",
    "# see set weights\n",
    "model.layers[8].get_weights()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Appenindg a constant input of 1 to the input tensor. The weight of this tensor will correspond to $\\Delta G_I$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_sequences_tensor2 = []\n",
    "test_input_sequences_tensor2 = []\n",
    "\n",
    "for _ in range(len(input_seqs_ohe)):\n",
    "    temp = input_seqs_ohe[_].flatten()\n",
    "    temp = np.append(temp,1)\n",
    "    input_sequences_tensor2.append(temp)\n",
    "    \n",
    "for _ in range(len(test_seqs_ohe)):\n",
    "    \n",
    "    temp = test_seqs_ohe[_].flatten()\n",
    "    temp = np.append(temp,1)\n",
    "    test_input_sequences_tensor2.append(temp)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Change inputs to TF/Keras to numpy arrays because that's the fit function expects."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "input_sequences_tensor2 = np.array(input_sequences_tensor2)\n",
    "test_input_sequences_tensor2 = np.array(test_input_sequences_tensor2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(269,)"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_sequences_tensor2[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "labels = np.array(y_train)\n",
    "test_labels = np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Define custom function as follows\n",
    "\n",
    "def poisson_loss(y_true, y_pred):\n",
    "    y_pred = ops.convert_to_tensor(y_pred)\n",
    "    y_true = math_ops.cast(y_true, y_pred.dtype)\n",
    "    return K.mean(y_pred - y_true * math_ops.log(y_pred + K.epsilon()), axis=-1)\n",
    "\n",
    "def log_poisson_loss(y_true, y_pred):\n",
    "    return tf.nn.log_poisson_loss(y_true,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.compile(loss=log_poisson_loss,optimizer=Adam(lr=0.0005))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 32331 samples, validate on 8083 samples\n",
      "Epoch 1/300\n",
      "32331/32331 [==============================] - 1s 39us/step - loss: 1.0952 - categorical_accuracy: 0.1114 - val_loss: 1.0952 - val_categorical_accuracy: 0.1181\n",
      "Epoch 2/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0952 - categorical_accuracy: 0.1115 - val_loss: 1.0951 - val_categorical_accuracy: 0.1181\n",
      "Epoch 3/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0951 - categorical_accuracy: 0.1115 - val_loss: 1.0951 - val_categorical_accuracy: 0.1181\n",
      "Epoch 4/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0950 - categorical_accuracy: 0.1165 - val_loss: 1.0948 - val_categorical_accuracy: 0.1636\n",
      "Epoch 5/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0944 - categorical_accuracy: 0.1801 - val_loss: 1.0939 - val_categorical_accuracy: 0.1732\n",
      "Epoch 6/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0934 - categorical_accuracy: 0.1906 - val_loss: 1.0931 - val_categorical_accuracy: 0.1931\n",
      "Epoch 7/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0928 - categorical_accuracy: 0.1979 - val_loss: 1.0926 - val_categorical_accuracy: 0.2117\n",
      "Epoch 8/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0924 - categorical_accuracy: 0.2119 - val_loss: 1.0923 - val_categorical_accuracy: 0.2185\n",
      "Epoch 9/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0922 - categorical_accuracy: 0.2265 - val_loss: 1.0921 - val_categorical_accuracy: 0.2343\n",
      "Epoch 10/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0920 - categorical_accuracy: 0.2364 - val_loss: 1.0919 - val_categorical_accuracy: 0.2446\n",
      "Epoch 11/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0918 - categorical_accuracy: 0.2446 - val_loss: 1.0918 - val_categorical_accuracy: 0.2479\n",
      "Epoch 12/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0917 - categorical_accuracy: 0.2524 - val_loss: 1.0916 - val_categorical_accuracy: 0.2472\n",
      "Epoch 13/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0915 - categorical_accuracy: 0.2573 - val_loss: 1.0915 - val_categorical_accuracy: 0.2625\n",
      "Epoch 14/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0914 - categorical_accuracy: 0.2628 - val_loss: 1.0914 - val_categorical_accuracy: 0.2493\n",
      "Epoch 15/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0913 - categorical_accuracy: 0.2652 - val_loss: 1.0913 - val_categorical_accuracy: 0.2619\n",
      "Epoch 16/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0912 - categorical_accuracy: 0.2698 - val_loss: 1.0912 - val_categorical_accuracy: 0.2662\n",
      "Epoch 17/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0912 - categorical_accuracy: 0.2696 - val_loss: 1.0912 - val_categorical_accuracy: 0.2669\n",
      "Epoch 18/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0911 - categorical_accuracy: 0.2731 - val_loss: 1.0911 - val_categorical_accuracy: 0.2758\n",
      "Epoch 19/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0910 - categorical_accuracy: 0.2750 - val_loss: 1.0910 - val_categorical_accuracy: 0.2708\n",
      "Epoch 20/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0910 - categorical_accuracy: 0.2781 - val_loss: 1.0910 - val_categorical_accuracy: 0.2728\n",
      "Epoch 21/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0909 - categorical_accuracy: 0.2788 - val_loss: 1.0909 - val_categorical_accuracy: 0.2821\n",
      "Epoch 22/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0909 - categorical_accuracy: 0.2813 - val_loss: 1.0909 - val_categorical_accuracy: 0.2708\n",
      "Epoch 23/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0908 - categorical_accuracy: 0.2832 - val_loss: 1.0909 - val_categorical_accuracy: 0.2775\n",
      "Epoch 24/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0908 - categorical_accuracy: 0.2841 - val_loss: 1.0908 - val_categorical_accuracy: 0.2785\n",
      "Epoch 25/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0908 - categorical_accuracy: 0.2827 - val_loss: 1.0908 - val_categorical_accuracy: 0.2842\n",
      "Epoch 26/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0907 - categorical_accuracy: 0.2842 - val_loss: 1.0908 - val_categorical_accuracy: 0.2855\n",
      "Epoch 27/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0907 - categorical_accuracy: 0.2867 - val_loss: 1.0907 - val_categorical_accuracy: 0.2863\n",
      "Epoch 28/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0907 - categorical_accuracy: 0.2863 - val_loss: 1.0907 - val_categorical_accuracy: 0.2880\n",
      "Epoch 29/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0907 - categorical_accuracy: 0.2868 - val_loss: 1.0907 - val_categorical_accuracy: 0.2813\n",
      "Epoch 30/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0906 - categorical_accuracy: 0.2897 - val_loss: 1.0907 - val_categorical_accuracy: 0.2889\n",
      "Epoch 31/300\n",
      "32331/32331 [==============================] - 1s 34us/step - loss: 1.0906 - categorical_accuracy: 0.2885 - val_loss: 1.0906 - val_categorical_accuracy: 0.2829\n",
      "Epoch 32/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0906 - categorical_accuracy: 0.2904 - val_loss: 1.0906 - val_categorical_accuracy: 0.2873\n",
      "Epoch 33/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0906 - categorical_accuracy: 0.2913 - val_loss: 1.0906 - val_categorical_accuracy: 0.2839\n",
      "Epoch 34/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0906 - categorical_accuracy: 0.2894 - val_loss: 1.0907 - val_categorical_accuracy: 0.2921\n",
      "Epoch 35/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0906 - categorical_accuracy: 0.2917 - val_loss: 1.0906 - val_categorical_accuracy: 0.2920\n",
      "Epoch 36/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0905 - categorical_accuracy: 0.2912 - val_loss: 1.0906 - val_categorical_accuracy: 0.2858\n",
      "Epoch 37/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2915 - val_loss: 1.0906 - val_categorical_accuracy: 0.2892\n",
      "Epoch 38/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2903 - val_loss: 1.0906 - val_categorical_accuracy: 0.2828\n",
      "Epoch 39/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2917 - val_loss: 1.0905 - val_categorical_accuracy: 0.2900\n",
      "Epoch 40/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2918 - val_loss: 1.0906 - val_categorical_accuracy: 0.2796\n",
      "Epoch 41/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2929 - val_loss: 1.0905 - val_categorical_accuracy: 0.2935\n",
      "Epoch 42/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2935 - val_loss: 1.0906 - val_categorical_accuracy: 0.2967\n",
      "Epoch 43/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2941 - val_loss: 1.0905 - val_categorical_accuracy: 0.2897\n",
      "Epoch 44/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0905 - categorical_accuracy: 0.2948 - val_loss: 1.0905 - val_categorical_accuracy: 0.2880\n",
      "Epoch 45/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2926 - val_loss: 1.0905 - val_categorical_accuracy: 0.2959\n",
      "Epoch 46/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2946 - val_loss: 1.0905 - val_categorical_accuracy: 0.2881\n",
      "Epoch 47/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0904 - categorical_accuracy: 0.2954 - val_loss: 1.0905 - val_categorical_accuracy: 0.2938\n",
      "Epoch 48/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0904 - categorical_accuracy: 0.2927 - val_loss: 1.0905 - val_categorical_accuracy: 0.2970\n",
      "Epoch 49/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2950 - val_loss: 1.0905 - val_categorical_accuracy: 0.2864\n",
      "Epoch 50/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2954 - val_loss: 1.0904 - val_categorical_accuracy: 0.2949\n",
      "Epoch 51/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2953 - val_loss: 1.0905 - val_categorical_accuracy: 0.2941\n",
      "Epoch 52/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2966 - val_loss: 1.0904 - val_categorical_accuracy: 0.2937\n",
      "Epoch 53/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2951 - val_loss: 1.0904 - val_categorical_accuracy: 0.2883\n",
      "Epoch 54/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2958 - val_loss: 1.0904 - val_categorical_accuracy: 0.2912\n",
      "Epoch 55/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2969 - val_loss: 1.0904 - val_categorical_accuracy: 0.2878\n",
      "Epoch 56/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0904 - categorical_accuracy: 0.2947 - val_loss: 1.0904 - val_categorical_accuracy: 0.2944\n",
      "Epoch 57/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2951 - val_loss: 1.0904 - val_categorical_accuracy: 0.2933\n",
      "Epoch 58/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2970 - val_loss: 1.0904 - val_categorical_accuracy: 0.2932\n",
      "Epoch 59/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0904 - categorical_accuracy: 0.2962 - val_loss: 1.0904 - val_categorical_accuracy: 0.2915\n",
      "Epoch 60/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2962 - val_loss: 1.0904 - val_categorical_accuracy: 0.2946\n",
      "Epoch 61/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2973 - val_loss: 1.0904 - val_categorical_accuracy: 0.2923\n",
      "Epoch 62/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2977 - val_loss: 1.0904 - val_categorical_accuracy: 0.2940\n",
      "Epoch 63/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2968 - val_loss: 1.0904 - val_categorical_accuracy: 0.2914\n",
      "Epoch 64/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2976 - val_loss: 1.0904 - val_categorical_accuracy: 0.2926\n",
      "Epoch 65/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2975 - val_loss: 1.0904 - val_categorical_accuracy: 0.2982\n",
      "Epoch 66/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2962 - val_loss: 1.0904 - val_categorical_accuracy: 0.2938\n",
      "Epoch 67/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2969 - val_loss: 1.0904 - val_categorical_accuracy: 0.2962\n",
      "Epoch 68/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2973 - val_loss: 1.0904 - val_categorical_accuracy: 0.2962\n",
      "Epoch 69/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2973 - val_loss: 1.0904 - val_categorical_accuracy: 0.2982\n",
      "Epoch 70/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2978 - val_loss: 1.0904 - val_categorical_accuracy: 0.2979\n",
      "Epoch 71/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2971 - val_loss: 1.0903 - val_categorical_accuracy: 0.2944\n",
      "Epoch 72/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2959\n",
      "Epoch 73/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.3012 - val_loss: 1.0904 - val_categorical_accuracy: 0.2995\n",
      "Epoch 74/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2979 - val_loss: 1.0903 - val_categorical_accuracy: 0.2959\n",
      "Epoch 75/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2985 - val_loss: 1.0903 - val_categorical_accuracy: 0.2959\n",
      "Epoch 76/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2983 - val_loss: 1.0903 - val_categorical_accuracy: 0.2989\n",
      "Epoch 77/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2963 - val_loss: 1.0904 - val_categorical_accuracy: 0.2985\n",
      "Epoch 78/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2983 - val_loss: 1.0903 - val_categorical_accuracy: 0.2946\n",
      "Epoch 79/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2981 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 80/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2994 - val_loss: 1.0903 - val_categorical_accuracy: 0.2957\n",
      "Epoch 81/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0903 - categorical_accuracy: 0.2996 - val_loss: 1.0903 - val_categorical_accuracy: 0.2962\n",
      "Epoch 82/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2988 - val_loss: 1.0903 - val_categorical_accuracy: 0.2961\n",
      "Epoch 83/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2962\n",
      "Epoch 84/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2982 - val_loss: 1.0903 - val_categorical_accuracy: 0.2932\n",
      "Epoch 85/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2987 - val_loss: 1.0903 - val_categorical_accuracy: 0.2949\n",
      "Epoch 86/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2978 - val_loss: 1.0904 - val_categorical_accuracy: 0.2925\n",
      "Epoch 87/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2989 - val_loss: 1.0903 - val_categorical_accuracy: 0.2978\n",
      "Epoch 88/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0903 - categorical_accuracy: 0.2993 - val_loss: 1.0903 - val_categorical_accuracy: 0.2982\n",
      "Epoch 89/300\n",
      "32331/32331 [==============================] - 1s 33us/step - loss: 1.0903 - categorical_accuracy: 0.2986 - val_loss: 1.0903 - val_categorical_accuracy: 0.2941\n",
      "Epoch 90/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2999 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 91/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0903 - val_categorical_accuracy: 0.2972\n",
      "Epoch 92/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0903 - val_categorical_accuracy: 0.2964\n",
      "Epoch 93/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2966 - val_loss: 1.0903 - val_categorical_accuracy: 0.2985\n",
      "Epoch 94/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2983 - val_loss: 1.0903 - val_categorical_accuracy: 0.2961\n",
      "Epoch 95/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2968\n",
      "Epoch 96/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2988 - val_loss: 1.0904 - val_categorical_accuracy: 0.3003\n",
      "Epoch 97/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0903 - val_categorical_accuracy: 0.2957\n",
      "Epoch 98/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2996 - val_loss: 1.0903 - val_categorical_accuracy: 0.2974\n",
      "Epoch 99/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2990 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 100/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2970\n",
      "Epoch 101/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2984 - val_loss: 1.0903 - val_categorical_accuracy: 0.2965\n",
      "Epoch 102/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2986 - val_loss: 1.0903 - val_categorical_accuracy: 0.2972\n",
      "Epoch 103/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2946\n",
      "Epoch 104/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2969\n",
      "Epoch 105/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0903 - val_categorical_accuracy: 0.2980\n",
      "Epoch 106/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2989 - val_loss: 1.0903 - val_categorical_accuracy: 0.2964\n",
      "Epoch 107/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2970\n",
      "Epoch 108/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2984\n",
      "Epoch 109/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2982 - val_loss: 1.0903 - val_categorical_accuracy: 0.2959\n",
      "Epoch 110/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2975\n",
      "Epoch 111/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2986 - val_loss: 1.0903 - val_categorical_accuracy: 0.2978\n",
      "Epoch 112/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3013 - val_loss: 1.0903 - val_categorical_accuracy: 0.2985\n",
      "Epoch 113/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 114/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 115/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2990 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 116/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 117/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2953\n",
      "Epoch 118/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2993 - val_loss: 1.0903 - val_categorical_accuracy: 0.2964\n",
      "Epoch 119/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0903 - val_categorical_accuracy: 0.2952\n",
      "Epoch 120/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2953\n",
      "Epoch 121/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2975\n",
      "Epoch 122/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2998 - val_loss: 1.0903 - val_categorical_accuracy: 0.2978\n",
      "Epoch 123/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2999 - val_loss: 1.0903 - val_categorical_accuracy: 0.2984\n",
      "Epoch 124/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2989 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 125/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 126/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2982\n",
      "Epoch 127/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3002 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 128/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 129/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2999 - val_loss: 1.0903 - val_categorical_accuracy: 0.2927\n",
      "Epoch 130/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3002 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 131/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2975\n",
      "Epoch 132/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2967\n",
      "Epoch 133/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2980\n",
      "Epoch 134/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0904 - val_categorical_accuracy: 0.2931\n",
      "Epoch 135/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3002 - val_loss: 1.0903 - val_categorical_accuracy: 0.2969\n",
      "Epoch 136/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 137/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2996 - val_loss: 1.0903 - val_categorical_accuracy: 0.2990\n",
      "Epoch 138/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2994 - val_loss: 1.0903 - val_categorical_accuracy: 0.2993\n",
      "Epoch 139/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3013 - val_loss: 1.0903 - val_categorical_accuracy: 0.3005\n",
      "Epoch 140/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2974\n",
      "Epoch 141/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2989\n",
      "Epoch 142/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0903 - val_categorical_accuracy: 0.2987\n",
      "Epoch 143/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 144/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 145/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0903 - val_categorical_accuracy: 0.2993\n",
      "Epoch 146/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0903 - val_categorical_accuracy: 0.2984\n",
      "Epoch 147/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2973\n",
      "Epoch 148/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 149/300\n",
      "32331/32331 [==============================] - 1s 33us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2993\n",
      "Epoch 150/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 151/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 152/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2986 - val_loss: 1.0903 - val_categorical_accuracy: 0.2958\n",
      "Epoch 153/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 154/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2999 - val_loss: 1.0903 - val_categorical_accuracy: 0.2972\n",
      "Epoch 155/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2986 - val_loss: 1.0903 - val_categorical_accuracy: 0.2980\n",
      "Epoch 156/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3010 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 157/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2999 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 158/300\n",
      "32331/32331 [==============================] - 1s 34us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 159/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 160/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2985\n",
      "Epoch 161/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2987\n",
      "Epoch 162/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 163/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0903 - val_categorical_accuracy: 0.2967\n",
      "Epoch 164/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3013 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988\n",
      "Epoch 165/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2978\n",
      "Epoch 166/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2998 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 167/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 168/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2992 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 169/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3013 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 170/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0902 - val_categorical_accuracy: 0.3000\n",
      "Epoch 171/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 172/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0903 - val_categorical_accuracy: 0.2998\n",
      "Epoch 173/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2975\n",
      "Epoch 174/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2973\n",
      "Epoch 175/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0903 - val_categorical_accuracy: 0.2988- loss: 1.0903 - categori\n",
      "Epoch 176/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0903 - val_categorical_accuracy: 0.2964\n",
      "Epoch 177/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0902 - val_categorical_accuracy: 0.2979\n",
      "Epoch 178/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 179/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2993\n",
      "Epoch 180/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2990\n",
      "Epoch 181/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 182/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 183/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2987 - val_loss: 1.0903 - val_categorical_accuracy: 0.2969\n",
      "Epoch 184/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0903 - val_categorical_accuracy: 0.3005\n",
      "Epoch 185/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0903 - val_categorical_accuracy: 0.3010\n",
      "Epoch 186/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2975\n",
      "Epoch 187/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 188/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3017 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 189/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2969\n",
      "Epoch 190/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0903 - val_categorical_accuracy: 0.3011\n",
      "Epoch 191/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3014 - val_loss: 1.0903 - val_categorical_accuracy: 0.3000\n",
      "Epoch 192/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0902 - val_categorical_accuracy: 0.2994\n",
      "Epoch 193/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3013 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 194/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0903 - val_categorical_accuracy: 0.2998\n",
      "Epoch 195/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2990\n",
      "Epoch 196/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 197/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3024 - val_loss: 1.0902 - val_categorical_accuracy: 0.2995\n",
      "Epoch 198/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0903 - val_categorical_accuracy: 0.2979\n",
      "Epoch 199/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3023 - val_loss: 1.0902 - val_categorical_accuracy: 0.2979\n",
      "Epoch 200/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0902 - val_categorical_accuracy: 0.2999\n",
      "Epoch 201/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0902 - val_categorical_accuracy: 0.2980\n",
      "Epoch 202/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0902 - val_categorical_accuracy: 0.2990\n",
      "Epoch 203/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 204/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3017 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 205/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2982\n",
      "Epoch 206/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3010 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 207/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 208/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 209/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0902 - val_categorical_accuracy: 0.2990\n",
      "Epoch 210/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0902 - val_categorical_accuracy: 0.2984\n",
      "Epoch 211/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3012 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 212/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2973\n",
      "Epoch 213/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0902 - val_categorical_accuracy: 0.2970\n",
      "Epoch 214/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 215/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0902 - val_categorical_accuracy: 0.2983\n",
      "Epoch 216/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0903 - val_categorical_accuracy: 0.2993\n",
      "Epoch 217/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 218/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3020 - val_loss: 1.0903 - val_categorical_accuracy: 0.2985\n",
      "Epoch 219/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0903 - val_categorical_accuracy: 0.2946\n",
      "Epoch 220/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0902 - val_categorical_accuracy: 0.3001\n",
      "Epoch 221/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0903 - val_categorical_accuracy: 0.2953\n",
      "Epoch 222/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3022 - val_loss: 1.0902 - val_categorical_accuracy: 0.2989\n",
      "Epoch 223/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3022 - val_loss: 1.0902 - val_categorical_accuracy: 0.2980\n",
      "Epoch 224/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3001 - val_loss: 1.0902 - val_categorical_accuracy: 0.2978\n",
      "Epoch 225/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3000 - val_loss: 1.0903 - val_categorical_accuracy: 0.2972: 1.0901 - categorical_ac\n",
      "Epoch 226/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2982\n",
      "Epoch 227/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3025 - val_loss: 1.0902 - val_categorical_accuracy: 0.2995\n",
      "Epoch 228/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 229/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2991\n",
      "Epoch 230/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977ss: 1.0901 - categorical_accuracy: 0.\n",
      "Epoch 231/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3022 - val_loss: 1.0903 - val_categorical_accuracy: 0.2946\n",
      "Epoch 232/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0903 - val_categorical_accuracy: 0.2970\n",
      "Epoch 233/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0902 - val_categorical_accuracy: 0.3001\n",
      "Epoch 234/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2984\n",
      "Epoch 235/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2982\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 236/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2958\n",
      "Epoch 237/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2998 - val_loss: 1.0903 - val_categorical_accuracy: 0.2978\n",
      "Epoch 238/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3017 - val_loss: 1.0902 - val_categorical_accuracy: 0.2980\n",
      "Epoch 239/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0903 - val_categorical_accuracy: 0.2995\n",
      "Epoch 240/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3020 - val_loss: 1.0903 - val_categorical_accuracy: 0.3004\n",
      "Epoch 241/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2995 - val_loss: 1.0902 - val_categorical_accuracy: 0.3003\n",
      "Epoch 242/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3005 - val_loss: 1.0903 - val_categorical_accuracy: 0.2995\n",
      "Epoch 243/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0903 - val_categorical_accuracy: 0.2998\n",
      "Epoch 244/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2975\n",
      "Epoch 245/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2998 - val_loss: 1.0902 - val_categorical_accuracy: 0.2983\n",
      "Epoch 246/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0903 - val_categorical_accuracy: 0.2984\n",
      "Epoch 247/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2995 - val_loss: 1.0903 - val_categorical_accuracy: 0.2947\n",
      "Epoch 248/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3027 - val_loss: 1.0902 - val_categorical_accuracy: 0.2978\n",
      "Epoch 249/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3010 - val_loss: 1.0903 - val_categorical_accuracy: 0.2968\n",
      "Epoch 250/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0903 - val_categorical_accuracy: 0.2947\n",
      "Epoch 251/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0902 - val_categorical_accuracy: 0.2967\n",
      "Epoch 252/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3019 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 253/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.2997 - val_loss: 1.0902 - val_categorical_accuracy: 0.2972\n",
      "Epoch 254/300\n",
      "32331/32331 [==============================] - 1s 33us/step - loss: 1.0902 - categorical_accuracy: 0.3020 - val_loss: 1.0903 - val_categorical_accuracy: 0.2989\n",
      "Epoch 255/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3022 - val_loss: 1.0903 - val_categorical_accuracy: 0.2995\n",
      "Epoch 256/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.2993 - val_loss: 1.0902 - val_categorical_accuracy: 0.2988\n",
      "Epoch 257/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 258/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3025 - val_loss: 1.0903 - val_categorical_accuracy: 0.2964\n",
      "Epoch 259/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3011 - val_loss: 1.0902 - val_categorical_accuracy: 0.2984\n",
      "Epoch 260/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0903 - val_categorical_accuracy: 0.2990\n",
      "Epoch 261/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2984\n",
      "Epoch 262/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3025 - val_loss: 1.0903 - val_categorical_accuracy: 0.2983\n",
      "Epoch 263/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3035 - val_loss: 1.0902 - val_categorical_accuracy: 0.2983\n",
      "Epoch 264/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3024 - val_loss: 1.0903 - val_categorical_accuracy: 0.2962\n",
      "Epoch 265/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0903 - val_categorical_accuracy: 0.2991\n",
      "Epoch 266/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0902 - val_categorical_accuracy: 0.2994\n",
      "Epoch 267/300\n",
      "32331/32331 [==============================] - 1s 33us/step - loss: 1.0902 - categorical_accuracy: 0.3010 - val_loss: 1.0902 - val_categorical_accuracy: 0.2982\n",
      "Epoch 268/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3026 - val_loss: 1.0902 - val_categorical_accuracy: 0.2978\n",
      "Epoch 269/300\n",
      "32331/32331 [==============================] - 1s 35us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2989\n",
      "Epoch 270/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0902 - val_categorical_accuracy: 0.2987\n",
      "Epoch 271/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3012 - val_loss: 1.0903 - val_categorical_accuracy: 0.2985\n",
      "Epoch 272/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3012 - val_loss: 1.0902 - val_categorical_accuracy: 0.2970\n",
      "Epoch 273/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3019 - val_loss: 1.0903 - val_categorical_accuracy: 0.2977\n",
      "Epoch 274/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3016 - val_loss: 1.0902 - val_categorical_accuracy: 0.2991\n",
      "Epoch 275/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3014 - val_loss: 1.0902 - val_categorical_accuracy: 0.2990\n",
      "Epoch 276/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3024 - val_loss: 1.0902 - val_categorical_accuracy: 0.2996\n",
      "Epoch 277/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3014 - val_loss: 1.0902 - val_categorical_accuracy: 0.2989\n",
      "Epoch 278/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 279/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0903 - val_categorical_accuracy: 0.2944\n",
      "Epoch 280/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3020 - val_loss: 1.0902 - val_categorical_accuracy: 0.2987\n",
      "Epoch 281/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0902 - val_categorical_accuracy: 0.2970\n",
      "Epoch 282/300\n",
      "32331/32331 [==============================] - 1s 33us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0902 - val_categorical_accuracy: 0.2991\n",
      "Epoch 283/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3018 - val_loss: 1.0902 - val_categorical_accuracy: 0.2989\n",
      "Epoch 284/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3010 - val_loss: 1.0902 - val_categorical_accuracy: 0.2985\n",
      "Epoch 285/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3020 - val_loss: 1.0903 - val_categorical_accuracy: 0.3014\n",
      "Epoch 286/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3024 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 287/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0902 - val_categorical_accuracy: 0.2979\n",
      "Epoch 288/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0902 - val_categorical_accuracy: 0.2969\n",
      "Epoch 289/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3004 - val_loss: 1.0902 - val_categorical_accuracy: 0.2984\n",
      "Epoch 290/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3015 - val_loss: 1.0902 - val_categorical_accuracy: 0.2994\n",
      "Epoch 291/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3007 - val_loss: 1.0903 - val_categorical_accuracy: 0.2963\n",
      "Epoch 292/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3021 - val_loss: 1.0903 - val_categorical_accuracy: 0.2998\n",
      "Epoch 293/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3019 - val_loss: 1.0902 - val_categorical_accuracy: 0.2977\n",
      "Epoch 294/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3006 - val_loss: 1.0902 - val_categorical_accuracy: 0.2959\n",
      "Epoch 295/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3008 - val_loss: 1.0902 - val_categorical_accuracy: 0.2983\n",
      "Epoch 296/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3024 - val_loss: 1.0902 - val_categorical_accuracy: 0.2973\n",
      "Epoch 297/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3019 - val_loss: 1.0902 - val_categorical_accuracy: 0.3010\n",
      "Epoch 298/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3009 - val_loss: 1.0902 - val_categorical_accuracy: 0.2989\n",
      "Epoch 299/300\n",
      "32331/32331 [==============================] - 1s 31us/step - loss: 1.0902 - categorical_accuracy: 0.3022 - val_loss: 1.0903 - val_categorical_accuracy: 0.2984\n",
      "Epoch 300/300\n",
      "32331/32331 [==============================] - 1s 32us/step - loss: 1.0902 - categorical_accuracy: 0.3003 - val_loss: 1.0902 - val_categorical_accuracy: 0.2979\n"
     ]
    }
   ],
   "source": [
    "# fit model with early stopping call back\n",
    "# esCallBack = keras.callbacks.EarlyStopping(monitor='val_loss',mode='auto')\n",
    "# history = model.fit(input_sequences_tensor2, labels, \n",
    "#                     validation_split=0.2, epochs=300,verbose=1,callbacks=[esCallBack])  \n",
    "\n",
    "history = model.fit(input_sequences_tensor2, labels, validation_split=0.2, epochs=300,verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "plot losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "\n",
    "plt.figure()\n",
    "plt.plot(history.history['loss'],color='blue')\n",
    "plt.plot(history.history['val_loss'],color='orange')\n",
    "plt.title('Model loss',fontsize=12)\n",
    "plt.ylabel('loss',fontsize=12)\n",
    "plt.xlabel('epoch',fontsize=12)\n",
    "plt.legend(['train', 'validation'])\n",
    "#plt.savefig('Write_up/model_loss.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "predictions_on_test = model.predict(test_input_sequences_tensor2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10104/10104 [==============================] - 0s 12us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[1.0902692891262111, 0.2964172604908947]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(test_input_sequences_tensor2,test_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualize inferred parameters (PWMs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "crp_weights = model.layers[4].get_weights()[0].copy()\n",
    "rnap_weights = model.layers[5].get_weights()[0].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "crp_weights = crp_weights.reshape(26,4)\n",
    "rnap_weights = rnap_weights.reshape(41,4)\n",
    "\n",
    "\n",
    "plt.figure(figsize=(12,3))\n",
    "sns.heatmap(crp_weights.T,cmap='RdBu_r')\n",
    "plt.show()\n",
    "plt.figure(figsize=(16,3))\n",
    "sns.heatmap(rnap_weights.T,cmap='RdBu_r')\n",
    "#sns.heatmap(rnap_weights.T,cmap='RdBu_r')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### visualize parameters as sequences logos using logomaker"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import logomaker\n",
    "crp_logo_df = pd.DataFrame(crp_weights,columns=['A','C','G','T'])\n",
    "\n",
    "crp_df.columns = ['A','C','G','T']\n",
    "logomaker.Logo(-crp_df,center_values=True, font_name='Arial Rounded MT Bold')\n",
    "plt.tick_params(labelsize=14)\n",
    "plt.title('CRP PSAM Reported in Kinney et al. 2010',fontsize=14)\n",
    "plt.ylabel(\"$-\\Delta \\Delta G$ ($k_{B}T$)\", labelpad=-1)\n",
    "logomaker.Logo(-crp_logo_df,center_values=True, font_name='Arial Rounded MT Bold')\n",
    "plt.title('CRP PSAM learned by neural network (full-wt data).',fontsize=14)\n",
    "plt.tick_params(labelsize=14)\n",
    "plt.ylabel(\"$-\\Delta \\Delta G$ ($k_{B}T$)\", labelpad=-1)\n",
    "\n",
    "plt.show()\n",
    "#logomaker.Logo(rnap_logo_df,center_values=True, font_name='Arial Rounded MT Bold')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rnap_logo_df = pd.DataFrame(rnap_weights,columns=['A','C','G','T'])\n",
    "rnap_df.columns = ['A','C','G','T']\n",
    "logomaker.Logo(-rnap_df,center_values=True, font_name='Arial Rounded MT Bold')\n",
    "plt.tick_params(labelsize=14)\n",
    "plt.title('RNAP PSAM Reported in Kinney et al. 2010',fontsize=14)\n",
    "plt.ylabel(\"$-\\Delta \\Delta G$ ($k_{B}T$)\", labelpad=-1)\n",
    "logomaker.Logo(rnap_logo_df,center_values=True, font_name='Arial Rounded MT Bold')\n",
    "plt.tick_params(labelsize=14)\n",
    "plt.title('RNAP PSAM learned by neural network (full-wt data).',fontsize=14)\n",
    "plt.ylabel(\"$-\\Delta \\Delta G$ ($k_{B}T$)\", labelpad=-1)\n",
    "plt.tick_params(labelsize=14)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inferred $\\Delta G_I$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[-4.6800294]], dtype=float32)]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[6].get_weights()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "value in kcal/mol"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-2.901618228"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "-4.6800294*0.62"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ensure that weights connecting $\\Delta G$ nodes to microstate nodes stay fixed after training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[0., 1., 0., 1.],\n",
       "        [0., 0., 1., 1.],\n",
       "        [0., 0., 0., 1.]], dtype=float32)]"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[8].get_weights()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Weights of the $\\tau$ layer, with the last two elements corresponding to $\\tau_{max}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[0.       ],\n",
       "        [0.       ],\n",
       "        [3.9697642],\n",
       "        [3.904942 ]], dtype=float32)]"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[9].get_weights()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}