BayesionNeuralNetwork.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6aa6431a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\Eam Kadry\\anaconda3\\Lib\\site-packages\\keras\\src\\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import re\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "47660444",
   "metadata": {},
   "outputs": [],
   "source": [
    "#    text_to_clean = re.sub(r'[-a-zA-Z0-9@:%._\\+~#=]{1,256}\\.[a-zA-Z0-9()]{1,6}\\b([-a-zA-Z0-9()@:%_\\+.~#?&//=]*)', '', text_to_fuck)\n",
    "def clean_input(text_to_clean):\n",
    "    text_to_clean = text_to_clean.lower()\n",
    "    text_to_clean = re.sub(r'https?:\\/\\/(www\\.)?[-a-zA-Z0-9@:%._\\+~#=]{1,256}\\.[a-zA-Z0-9()]{1,6}\\b([-a-zA-Z0-9()@:%_\\+.~#?&//=]*)', '', text_to_clean)\n",
    "    text_to_clean = re.sub(r'[0-9]+', 'number', text_to_clean)\n",
    "    text_to_clean = re.sub(r'[^a-zA-Z\\u0621-\\u064A\\s]', '', text_to_clean)\n",
    "    text_to_clean = re.sub(r'numbernumber+', 'number', text_to_clean)\n",
    "    return text_to_clean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "2e1a284d",
   "metadata": {},
   "outputs": [],
   "source": [
    "train=pd.read_csv('Dataset/train.csv')\n",
    "test=pd.read_csv('Dataset/test.csv')\n",
    "y_test=pd.read_csv('Dataset/submit.csv')\n",
    "train=train.fillna(' ')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "52924ca3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(20800,)\n",
      "(20800, 1)\n"
     ]
    }
   ],
   "source": [
    "#Initializ y_train and x_train\n",
    "train['text_merge'] = train['title'].astype(str) + ' ' + train['author'].astype(str) + ' ' + train['text'].astype(str)\n",
    "x_train = train['text_merge']\n",
    "y_train = train.drop(['id','title','author','text','text_merge'],axis=1)\n",
    "y_train = np.array(y_train)\n",
    "print(x_train.shape)\n",
    "print(y_train.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d2ce26c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0       Specter of Trump Loosens Tongues, if Not Purse...\n",
       "1       Russian warships ready to strike terrorists ne...\n",
       "2       #NoDAPL: Native American Leaders Vow to Stay A...\n",
       "3       Tim Tebow Will Attempt Another Comeback, This ...\n",
       "4       Keiser Report: Meme Wars (E995) Truth Broadcas...\n",
       "                              ...                        \n",
       "5195    The Bangladeshi Traffic Jam That Never Ends - ...\n",
       "5196    John Kasich Signs One Abortion Bill in Ohio bu...\n",
       "5197    California Today: What, Exactly, Is in Your Su...\n",
       "5198    300 US Marines To Be Deployed To Russian Borde...\n",
       "5199    Awkward Sex, Onscreen and Off - The New York T...\n",
       "Name: text_merge, Length: 5200, dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Initializ y_test and x_test\n",
    "test['text_merge'] = test['title'].astype(str) + ' ' + test['author'].astype(str) + ' ' + test['text'].astype(str)\n",
    "x_test = test['text_merge']\n",
    "y_test = y_test.drop(['id'],axis=1)\n",
    "y_test = np.array(y_test)\n",
    "x_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "5b0d4407",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i in range(x_train.shape[0]):\n",
    "    x_train.values[i] = clean_input(x_train[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e9fc7273",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "512\n",
      "(5200, 512)\n"
     ]
    }
   ],
   "source": [
    "# x_train_list = x_train.tolist()\n",
    "tokenizer=Tokenizer(oov_token=\"<OOV>\")\n",
    "tokenizer.fit_on_texts(x_train)\n",
    "word_index=tokenizer.word_index\n",
    "sequences_train = tokenizer.texts_to_sequences(x_train)\n",
    "X_train = pad_sequences(sequences_train,maxlen=512,truncating='post',padding='post')\n",
    "sequences_test = tokenizer.texts_to_sequences(x_test)\n",
    "X_test = pad_sequences(sequences_test,maxlen=512,padding='post',truncating='post')\n",
    "# print(sequences_test)\n",
    "# print(word_index)\n",
    "\n",
    "print(X_train.shape[1])\n",
    "print(X_test.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "8bc5c325",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigmoid(x):\n",
    "    return 1/(1+np.exp(-x))\n",
    "def relu(x):\n",
    "    return np.maximum(0,x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "77df82ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "#the derivate of te activation function\n",
    "def sigmoid_derivative(x):\n",
    "    return sigmoid(x) * (1 - sigmoid(x))\n",
    "def relu_derivative(x):\n",
    "    return np.where(x > 0, 1, 0)\n",
    "def tanh_derivative(x):\n",
    "    return 1 - np.tanh(x)**2\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "2e27e97a",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Eam Kadry\\AppData\\Local\\Temp\\ipykernel_19744\\3032088681.py:4: RuntimeWarning: overflow encountered in exp\n",
      "  return 1 / (1 + np.exp(-x))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 2, Loss: 0.5034448317303285\n",
      "Accuracy: 49.64%\n",
      "Iteration 0:\n",
      "Training Loss: 0.5034448317303285, Training Accuracy: 49.65%\n",
      "Test Loss: 0.5463461538460332, Test Accuracy: 45.37%\n",
      "----------------------------------------\n",
      "Epoch 4, Loss: 0.5038298978723973\n",
      "Accuracy: 49.60%\n",
      "Iteration 1:\n",
      "Training Loss: 0.5038298978723973, Training Accuracy: 49.62%\n",
      "Test Loss: 0.546730769230551, Test Accuracy: 45.33%\n",
      "----------------------------------------\n",
      "Epoch 6, Loss: 0.5033489770208758\n",
      "Accuracy: 49.64%\n",
      "Iteration 2:\n",
      "Training Loss: 0.5033489770208758, Training Accuracy: 49.66%\n",
      "Test Loss: 0.5473076923073277, Test Accuracy: 45.27%\n",
      "----------------------------------------\n",
      "Epoch 8, Loss: 0.5033487106771792\n",
      "Accuracy: 49.65%\n",
      "Iteration 3:\n",
      "Training Loss: 0.5033487106771792, Training Accuracy: 49.66%\n",
      "Test Loss: 0.5488461538409092, Test Accuracy: 45.12%\n",
      "----------------------------------------\n",
      "Epoch 10, Loss: 0.5035415575160936\n",
      "Accuracy: 49.63%\n",
      "Iteration 4:\n",
      "Training Loss: 0.5035415575160936, Training Accuracy: 49.64%\n",
      "Test Loss: 0.5484615384607272, Test Accuracy: 45.15%\n",
      "----------------------------------------\n",
      "Epoch 12, Loss: 0.5038301588309201\n",
      "Accuracy: 49.59%\n",
      "Iteration 5:\n",
      "Training Loss: 0.5038301588309201, Training Accuracy: 49.62%\n",
      "Test Loss: 0.5490384615372073, Test Accuracy: 45.10%\n",
      "----------------------------------------\n",
      "Epoch 14, Loss: 0.5042149168933844\n",
      "Accuracy: 49.56%\n",
      "Iteration 6:\n",
      "Training Loss: 0.5042149168933844, Training Accuracy: 49.58%\n",
      "Test Loss: 0.5490384615364664, Test Accuracy: 45.10%\n",
      "----------------------------------------\n",
      "Epoch 16, Loss: 0.5042150625072609\n",
      "Accuracy: 49.55%\n",
      "Iteration 7:\n",
      "Training Loss: 0.5042150625072609, Training Accuracy: 49.58%\n",
      "Test Loss: 0.5478846153805194, Test Accuracy: 45.21%\n",
      "----------------------------------------\n",
      "Epoch 18, Loss: 0.5046480406697763\n",
      "Accuracy: 49.51%\n",
      "Iteration 8:\n",
      "Training Loss: 0.5046480406697763, Training Accuracy: 49.53%\n",
      "Test Loss: 0.5469230769108713, Test Accuracy: 45.31%\n",
      "----------------------------------------\n",
      "Epoch 20, Loss: 0.504551900362999\n",
      "Accuracy: 49.50%\n",
      "Iteration 9:\n",
      "Training Loss: 0.504551900362999, Training Accuracy: 49.54%\n",
      "Test Loss: 0.5471153845727295, Test Accuracy: 45.29%\n",
      "----------------------------------------\n",
      "Epoch 22, Loss: 0.5058982075350831\n",
      "Accuracy: 49.39%\n",
      "Iteration 10:\n",
      "Training Loss: 0.5058982075350831, Training Accuracy: 49.41%\n",
      "Test Loss: 0.5488461536371294, Test Accuracy: 45.12%\n",
      "----------------------------------------\n",
      "Epoch 24, Loss: 0.5058022090238374\n",
      "Accuracy: 49.38%\n",
      "Iteration 11:\n",
      "Training Loss: 0.5058022090238374, Training Accuracy: 49.42%\n",
      "Test Loss: 0.5503846137928259, Test Accuracy: 44.96%\n",
      "----------------------------------------\n",
      "Epoch 26, Loss: 0.505994614777864\n",
      "Accuracy: 49.37%\n",
      "Iteration 12:\n",
      "Training Loss: 0.505994614777864, Training Accuracy: 49.40%\n",
      "Test Loss: 0.5498076749094502, Test Accuracy: 45.02%\n",
      "----------------------------------------\n",
      "Epoch 28, Loss: 0.5057544475139261\n",
      "Accuracy: 49.40%\n",
      "Iteration 13:\n",
      "Training Loss: 0.5057544475139261, Training Accuracy: 49.42%\n",
      "Test Loss: 0.5499997732081561, Test Accuracy: 45.00%\n",
      "----------------------------------------\n",
      "Epoch 30, Loss: 0.5057065299843351\n",
      "Accuracy: 49.39%\n",
      "Iteration 14:\n",
      "Training Loss: 0.5057065299843351, Training Accuracy: 49.43%\n",
      "Test Loss: 0.5496121984662169, Test Accuracy: 45.04%\n",
      "----------------------------------------\n",
      "Epoch 32, Loss: 0.5057547670320036\n",
      "Accuracy: 49.38%\n",
      "Iteration 15:\n",
      "Training Loss: 0.5057547670320036, Training Accuracy: 49.42%\n",
      "Test Loss: 0.550344118955916, Test Accuracy: 44.96%\n",
      "----------------------------------------\n",
      "Epoch 34, Loss: 0.5059953121473763\n",
      "Accuracy: 49.36%\n",
      "Iteration 16:\n",
      "Training Loss: 0.5059953121473763, Training Accuracy: 49.40%\n",
      "Test Loss: 0.5511758463104024, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 36, Loss: 0.5068608573510102\n",
      "Accuracy: 49.26%\n",
      "Iteration 17:\n",
      "Training Loss: 0.5068608573510102, Training Accuracy: 49.31%\n",
      "Test Loss: 0.5513450323110048, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 38, Loss: 0.5073417869495723\n",
      "Accuracy: 49.23%\n",
      "Iteration 18:\n",
      "Training Loss: 0.5073417869495723, Training Accuracy: 49.26%\n",
      "Test Loss: 0.5511486018003928, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 40, Loss: 0.5081111775887728\n",
      "Accuracy: 49.16%\n",
      "Iteration 19:\n",
      "Training Loss: 0.5081111775887728, Training Accuracy: 49.19%\n",
      "Test Loss: 0.5490197170067656, Test Accuracy: 45.10%\n",
      "----------------------------------------\n",
      "Epoch 42, Loss: 0.5083997982508933\n",
      "Accuracy: 49.12%\n",
      "Iteration 20:\n",
      "Training Loss: 0.5083997982508933, Training Accuracy: 49.16%\n",
      "Test Loss: 0.5489751773659947, Test Accuracy: 45.10%\n",
      "----------------------------------------\n",
      "Epoch 44, Loss: 0.5086122360198793\n",
      "Accuracy: 49.11%\n",
      "Iteration 21:\n",
      "Training Loss: 0.5086122360198793, Training Accuracy: 49.13%\n",
      "Test Loss: 0.5494344947863053, Test Accuracy: 45.04%\n",
      "----------------------------------------\n",
      "Epoch 46, Loss: 0.5087085468564777\n",
      "Accuracy: 49.12%\n",
      "Iteration 22:\n",
      "Training Loss: 0.5087085468564777, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5476495644922685, Test Accuracy: 45.21%\n",
      "----------------------------------------\n",
      "Epoch 48, Loss: 0.5087554326796419\n",
      "Accuracy: 49.08%\n",
      "Iteration 23:\n",
      "Training Loss: 0.5087554326796419, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5475111832208273, Test Accuracy: 45.23%\n",
      "----------------------------------------\n",
      "Epoch 50, Loss: 0.5101030871737174\n",
      "Accuracy: 48.98%\n",
      "Iteration 24:\n",
      "Training Loss: 0.5101030871737174, Training Accuracy: 48.99%\n",
      "Test Loss: 0.5490725260590091, Test Accuracy: 45.10%\n",
      "----------------------------------------\n",
      "Epoch 52, Loss: 0.5095263163549557\n",
      "Accuracy: 49.01%\n",
      "Iteration 25:\n",
      "Training Loss: 0.5095263163549557, Training Accuracy: 49.04%\n",
      "Test Loss: 0.5487966895351754, Test Accuracy: 45.12%\n",
      "----------------------------------------\n",
      "Epoch 54, Loss: 0.5100553128781272\n",
      "Accuracy: 48.96%\n",
      "Iteration 26:\n",
      "Training Loss: 0.5100553128781272, Training Accuracy: 48.99%\n",
      "Test Loss: 0.5481456049100416, Test Accuracy: 45.17%\n",
      "----------------------------------------\n",
      "Epoch 56, Loss: 0.5102955282606625\n",
      "Accuracy: 48.93%\n",
      "Iteration 27:\n",
      "Training Loss: 0.5102955282606625, Training Accuracy: 48.97%\n",
      "Test Loss: 0.5471332291021008, Test Accuracy: 45.29%\n",
      "----------------------------------------\n",
      "Epoch 58, Loss: 0.5107734444544664\n",
      "Accuracy: 48.89%\n",
      "Iteration 28:\n",
      "Training Loss: 0.5107734444544664, Training Accuracy: 48.92%\n",
      "Test Loss: 0.5480796735366715, Test Accuracy: 45.19%\n",
      "----------------------------------------\n",
      "Epoch 60, Loss: 0.5113092622512944\n",
      "Accuracy: 48.82%\n",
      "Iteration 29:\n",
      "Training Loss: 0.5113092622512944, Training Accuracy: 48.87%\n",
      "Test Loss: 0.5503849513790119, Test Accuracy: 44.96%\n",
      "----------------------------------------\n",
      "Epoch 62, Loss: 0.5116904042543463\n",
      "Accuracy: 48.79%\n",
      "Iteration 30:\n",
      "Training Loss: 0.5116904042543463, Training Accuracy: 48.83%\n",
      "Test Loss: 0.5501923810912256, Test Accuracy: 44.98%\n",
      "----------------------------------------\n",
      "Epoch 64, Loss: 0.512459878391471\n",
      "Accuracy: 48.71%\n",
      "Iteration 31:\n",
      "Training Loss: 0.512459878391471, Training Accuracy: 48.75%\n",
      "Test Loss: 0.5513461586891067, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 66, Loss: 0.5128927075500561\n",
      "Accuracy: 48.69%\n",
      "Iteration 32:\n",
      "Training Loss: 0.5128927075500561, Training Accuracy: 48.71%\n",
      "Test Loss: 0.5519230775659242, Test Accuracy: 44.81%\n",
      "----------------------------------------\n",
      "Epoch 68, Loss: 0.5138063034812616\n",
      "Accuracy: 48.56%\n",
      "Iteration 33:\n",
      "Training Loss: 0.5138063034812616, Training Accuracy: 48.62%\n",
      "Test Loss: 0.5513461539405607, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 70, Loss: 0.5144541945917831\n",
      "Accuracy: 48.52%\n",
      "Iteration 34:\n",
      "Training Loss: 0.5144541945917831, Training Accuracy: 48.55%\n",
      "Test Loss: 0.5538461538618822, Test Accuracy: 44.62%\n",
      "----------------------------------------\n",
      "Epoch 72, Loss: 0.5150084880068878\n",
      "Accuracy: 48.47%\n",
      "Iteration 35:\n",
      "Training Loss: 0.5150084880068878, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5540384615414392, Test Accuracy: 44.60%\n",
      "----------------------------------------\n",
      "Epoch 74, Loss: 0.5153451534914041\n",
      "Accuracy: 48.44%\n",
      "Iteration 36:\n",
      "Training Loss: 0.5153451534914041, Training Accuracy: 48.46%\n",
      "Test Loss: 0.5540384615390631, Test Accuracy: 44.60%\n",
      "----------------------------------------\n",
      "Epoch 76, Loss: 0.5150568127809028\n",
      "Accuracy: 48.46%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration 37:\n",
      "Training Loss: 0.5150568127809028, Training Accuracy: 48.49%\n",
      "Test Loss: 0.5557692307693688, Test Accuracy: 44.42%\n",
      "----------------------------------------\n",
      "Epoch 78, Loss: 0.5153454000238067\n",
      "Accuracy: 48.42%\n",
      "Iteration 38:\n",
      "Training Loss: 0.5153454000238067, Training Accuracy: 48.46%\n",
      "Test Loss: 0.5565353821562773, Test Accuracy: 44.35%\n",
      "----------------------------------------\n",
      "Epoch 80, Loss: 0.5164512887607943\n",
      "Accuracy: 48.31%\n",
      "Iteration 39:\n",
      "Training Loss: 0.5164512887607943, Training Accuracy: 48.35%\n",
      "Test Loss: 0.5582692307692418, Test Accuracy: 44.17%\n",
      "----------------------------------------\n",
      "Epoch 82, Loss: 0.5166131089708154\n",
      "Accuracy: 48.27%\n",
      "Iteration 40:\n",
      "Training Loss: 0.5166131089708154, Training Accuracy: 48.33%\n",
      "Test Loss: 0.5596153846153866, Test Accuracy: 44.04%\n",
      "----------------------------------------\n",
      "Epoch 84, Loss: 0.5171046253850762\n",
      "Accuracy: 48.25%\n",
      "Iteration 41:\n",
      "Training Loss: 0.5171046253850762, Training Accuracy: 48.29%\n",
      "Test Loss: 0.5598076923076927, Test Accuracy: 44.02%\n",
      "----------------------------------------\n",
      "Epoch 86, Loss: 0.5176716102994108\n",
      "Accuracy: 48.18%\n",
      "Iteration 42:\n",
      "Training Loss: 0.5176716102994108, Training Accuracy: 48.23%\n",
      "Test Loss: 0.5605769230769232, Test Accuracy: 43.94%\n",
      "----------------------------------------\n",
      "Epoch 88, Loss: 0.5184510011213245\n",
      "Accuracy: 48.13%\n",
      "Iteration 43:\n",
      "Training Loss: 0.5184510011213245, Training Accuracy: 48.15%\n",
      "Test Loss: 0.5603846153846154, Test Accuracy: 43.96%\n",
      "----------------------------------------\n",
      "Epoch 90, Loss: 0.5185953392725439\n",
      "Accuracy: 48.11%\n",
      "Iteration 44:\n",
      "Training Loss: 0.5185953392725439, Training Accuracy: 48.14%\n",
      "Test Loss: 0.5607692307692308, Test Accuracy: 43.92%\n",
      "----------------------------------------\n",
      "Epoch 92, Loss: 0.5189800597083223\n",
      "Accuracy: 48.01%\n",
      "Iteration 45:\n",
      "Training Loss: 0.5189800597083223, Training Accuracy: 48.10%\n",
      "Test Loss: 0.5613461538461535, Test Accuracy: 43.87%\n",
      "----------------------------------------\n",
      "Epoch 94, Loss: 0.5202782394660362\n",
      "Accuracy: 47.90%\n",
      "Iteration 46:\n",
      "Training Loss: 0.5202782394660362, Training Accuracy: 47.97%\n",
      "Test Loss: 0.5617307692307693, Test Accuracy: 43.83%\n",
      "----------------------------------------\n",
      "Epoch 96, Loss: 0.5206148783728158\n",
      "Accuracy: 47.91%\n",
      "Iteration 47:\n",
      "Training Loss: 0.5206148783728158, Training Accuracy: 47.94%\n",
      "Test Loss: 0.5615384615384615, Test Accuracy: 43.85%\n",
      "----------------------------------------\n",
      "Epoch 98, Loss: 0.5208072843721405\n",
      "Accuracy: 47.85%\n",
      "Iteration 48:\n",
      "Training Loss: 0.5208072843721405, Training Accuracy: 47.92%\n",
      "Test Loss: 0.5617307692307693, Test Accuracy: 43.83%\n",
      "----------------------------------------\n",
      "Epoch 100, Loss: 0.5215285343535867\n",
      "Accuracy: 47.77%\n",
      "Iteration 49:\n",
      "Training Loss: 0.5215285343535867, Training Accuracy: 47.85%\n",
      "Test Loss: 0.5628846153846154, Test Accuracy: 43.71%\n",
      "----------------------------------------\n",
      "Epoch 102, Loss: 0.5223459360649123\n",
      "Accuracy: 47.70%\n",
      "Iteration 50:\n",
      "Training Loss: 0.5223459360649123, Training Accuracy: 47.76%\n",
      "Test Loss: 0.563076923076923, Test Accuracy: 43.69%\n",
      "----------------------------------------\n",
      "Epoch 104, Loss: 0.523307566457452\n",
      "Accuracy: 47.62%\n",
      "Iteration 51:\n",
      "Training Loss: 0.523307566457452, Training Accuracy: 47.67%\n",
      "Test Loss: 0.5632692307692307, Test Accuracy: 43.67%\n",
      "----------------------------------------\n",
      "Epoch 106, Loss: 0.5242210556049907\n",
      "Accuracy: 47.56%\n",
      "Iteration 52:\n",
      "Training Loss: 0.5242210556049907, Training Accuracy: 47.58%\n",
      "Test Loss: 0.5628846153846154, Test Accuracy: 43.71%\n",
      "----------------------------------------\n",
      "Epoch 108, Loss: 0.5238365903780896\n",
      "Accuracy: 47.60%\n",
      "Iteration 53:\n",
      "Training Loss: 0.5238365903780896, Training Accuracy: 47.62%\n",
      "Test Loss: 0.5626923076923077, Test Accuracy: 43.73%\n",
      "----------------------------------------\n",
      "Epoch 110, Loss: 0.5238847532392719\n",
      "Accuracy: 47.57%\n",
      "Iteration 54:\n",
      "Training Loss: 0.5238847532392719, Training Accuracy: 47.61%\n",
      "Test Loss: 0.5634615384615385, Test Accuracy: 43.65%\n",
      "----------------------------------------\n",
      "Epoch 112, Loss: 0.5243171374575465\n",
      "Accuracy: 47.49%\n",
      "Iteration 55:\n",
      "Training Loss: 0.5243171374575465, Training Accuracy: 47.57%\n",
      "Test Loss: 0.5632199114508026, Test Accuracy: 43.67%\n",
      "----------------------------------------\n",
      "Epoch 114, Loss: 0.5246541502268346\n",
      "Accuracy: 47.50%\n",
      "Iteration 56:\n",
      "Training Loss: 0.5246541502268346, Training Accuracy: 47.53%\n",
      "Test Loss: 0.5632692307692307, Test Accuracy: 43.67%\n",
      "----------------------------------------\n",
      "Epoch 116, Loss: 0.524558076735647\n",
      "Accuracy: 47.48%\n",
      "Iteration 57:\n",
      "Training Loss: 0.524558076735647, Training Accuracy: 47.54%\n",
      "Test Loss: 0.5630769230769228, Test Accuracy: 43.69%\n",
      "----------------------------------------\n",
      "Epoch 118, Loss: 0.52551969375823\n",
      "Accuracy: 47.42%\n",
      "Iteration 58:\n",
      "Training Loss: 0.52551969375823, Training Accuracy: 47.45%\n",
      "Test Loss: 0.5621153846153846, Test Accuracy: 43.79%\n",
      "----------------------------------------\n",
      "Epoch 120, Loss: 0.5255678474727827\n",
      "Accuracy: 47.41%\n",
      "Iteration 59:\n",
      "Training Loss: 0.5255678474727827, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5628846153845933, Test Accuracy: 43.71%\n",
      "----------------------------------------\n",
      "Epoch 122, Loss: 0.5256640763981832\n",
      "Accuracy: 47.38%\n",
      "Iteration 60:\n",
      "Training Loss: 0.5256640763981832, Training Accuracy: 47.43%\n",
      "Test Loss: 0.563076923076923, Test Accuracy: 43.69%\n",
      "----------------------------------------\n",
      "Epoch 124, Loss: 0.5257122266787612\n",
      "Accuracy: 47.40%\n",
      "Iteration 61:\n",
      "Training Loss: 0.5257122266787612, Training Accuracy: 47.43%\n",
      "Test Loss: 0.5634615384614666, Test Accuracy: 43.65%\n",
      "----------------------------------------\n",
      "Epoch 126, Loss: 0.5258565291312753\n",
      "Accuracy: 47.40%\n",
      "Iteration 62:\n",
      "Training Loss: 0.5258565291312753, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5642307692307692, Test Accuracy: 43.58%\n",
      "----------------------------------------\n",
      "Epoch 128, Loss: 0.5257604294394994\n",
      "Accuracy: 47.37%\n",
      "Iteration 63:\n",
      "Training Loss: 0.5257604294394994, Training Accuracy: 47.42%\n",
      "Test Loss: 0.5643555629545256, Test Accuracy: 43.56%\n",
      "----------------------------------------\n",
      "Epoch 130, Loss: 0.5260008982701563\n",
      "Accuracy: 47.37%\n",
      "Iteration 64:\n",
      "Training Loss: 0.5260008982701563, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5640384615384615, Test Accuracy: 43.60%\n",
      "----------------------------------------\n",
      "Epoch 132, Loss: 0.5255201961874916\n",
      "Accuracy: 47.42%\n",
      "Iteration 65:\n",
      "Training Loss: 0.5255201961874916, Training Accuracy: 47.45%\n",
      "Test Loss: 0.5625, Test Accuracy: 43.75%\n",
      "----------------------------------------\n",
      "Epoch 134, Loss: 0.5259048770340415\n",
      "Accuracy: 47.36%\n",
      "Iteration 66:\n",
      "Training Loss: 0.5259048770340415, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5630769230769235, Test Accuracy: 43.69%\n",
      "----------------------------------------\n",
      "Epoch 136, Loss: 0.5260010970040055\n",
      "Accuracy: 47.33%\n",
      "Iteration 67:\n",
      "Training Loss: 0.5260010970040055, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5638461394239114, Test Accuracy: 43.62%\n",
      "----------------------------------------\n",
      "Epoch 138, Loss: 0.5265300034892564\n",
      "Accuracy: 47.32%\n",
      "Iteration 68:\n",
      "Training Loss: 0.5265300034892564, Training Accuracy: 47.35%\n",
      "Test Loss: 0.5638461538461539, Test Accuracy: 43.62%\n",
      "----------------------------------------\n",
      "Epoch 140, Loss: 0.5263858338608272\n",
      "Accuracy: 47.27%\n",
      "Iteration 69:\n",
      "Training Loss: 0.5263858338608272, Training Accuracy: 47.36%\n",
      "Test Loss: 0.5638461538461539, Test Accuracy: 43.62%\n",
      "----------------------------------------\n",
      "Epoch 142, Loss: 0.5270442984302448\n",
      "Accuracy: 47.23%\n",
      "Iteration 70:\n",
      "Training Loss: 0.5270442984302448, Training Accuracy: 47.29%\n",
      "Test Loss: 0.563076923076923, Test Accuracy: 43.69%\n",
      "----------------------------------------\n",
      "Epoch 144, Loss: 0.5277321058223958\n",
      "Accuracy: 47.20%\n",
      "Iteration 71:\n",
      "Training Loss: 0.5277321058223958, Training Accuracy: 47.23%\n",
      "Test Loss: 0.561923076923077, Test Accuracy: 43.81%\n",
      "----------------------------------------\n",
      "Epoch 146, Loss: 0.5277802397796237\n",
      "Accuracy: 47.19%\n",
      "Iteration 72:\n",
      "Training Loss: 0.5277802397796237, Training Accuracy: 47.22%\n",
      "Test Loss: 0.5609615384615385, Test Accuracy: 43.90%\n",
      "----------------------------------------\n",
      "Epoch 148, Loss: 0.5278764493534246\n",
      "Accuracy: 47.18%\n",
      "Iteration 73:\n",
      "Training Loss: 0.5278764493534246, Training Accuracy: 47.21%\n",
      "Test Loss: 0.5605769232517901, Test Accuracy: 43.94%\n",
      "----------------------------------------\n",
      "Epoch 150, Loss: 0.5278651512980213\n",
      "Accuracy: 47.16%\n",
      "Iteration 74:\n",
      "Training Loss: 0.5278651512980213, Training Accuracy: 47.21%\n",
      "Test Loss: 0.5603846153841318, Test Accuracy: 43.96%\n",
      "----------------------------------------\n",
      "Epoch 152, Loss: 0.5280207881737269\n",
      "Accuracy: 47.14%\n",
      "Iteration 75:\n",
      "Training Loss: 0.5280207881737269, Training Accuracy: 47.20%\n",
      "Test Loss: 0.5601923076923077, Test Accuracy: 43.98%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 154, Loss: 0.5286939174072571\n",
      "Accuracy: 47.09%\n",
      "Iteration 76:\n",
      "Training Loss: 0.5286939174072571, Training Accuracy: 47.13%\n",
      "Test Loss: 0.5598076923076923, Test Accuracy: 44.02%\n",
      "----------------------------------------\n",
      "Epoch 156, Loss: 0.5293189686268454\n",
      "Accuracy: 47.00%\n",
      "Iteration 77:\n",
      "Training Loss: 0.5293189686268454, Training Accuracy: 47.07%\n",
      "Test Loss: 0.5601923076923077, Test Accuracy: 43.98%\n",
      "----------------------------------------\n",
      "Epoch 158, Loss: 0.5297915731226788\n",
      "Accuracy: 46.96%\n",
      "Iteration 78:\n",
      "Training Loss: 0.5297915731226788, Training Accuracy: 47.02%\n",
      "Test Loss: 0.5595029645442884, Test Accuracy: 44.04%\n",
      "----------------------------------------\n",
      "Epoch 160, Loss: 0.5298959910002892\n",
      "Accuracy: 46.94%\n",
      "Iteration 79:\n",
      "Training Loss: 0.5298959910002892, Training Accuracy: 47.01%\n",
      "Test Loss: 0.559695272236596, Test Accuracy: 44.02%\n",
      "----------------------------------------\n",
      "Epoch 162, Loss: 0.5299440629577664\n",
      "Accuracy: 46.94%\n",
      "Iteration 80:\n",
      "Training Loss: 0.5299440629577664, Training Accuracy: 47.00%\n",
      "Test Loss: 0.5600798876212114, Test Accuracy: 43.98%\n",
      "----------------------------------------\n",
      "Epoch 164, Loss: 0.5302810276792174\n",
      "Accuracy: 46.90%\n",
      "Iteration 81:\n",
      "Training Loss: 0.5302810276792174, Training Accuracy: 46.97%\n",
      "Test Loss: 0.5604620247690386, Test Accuracy: 43.94%\n",
      "----------------------------------------\n",
      "Epoch 166, Loss: 0.5305209262015483\n",
      "Accuracy: 46.91%\n",
      "Iteration 82:\n",
      "Training Loss: 0.5305209262015483, Training Accuracy: 46.95%\n",
      "Test Loss: 0.5610414260827499, Test Accuracy: 43.88%\n",
      "----------------------------------------\n",
      "Epoch 168, Loss: 0.5306853627033807\n",
      "Accuracy: 46.88%\n",
      "Iteration 83:\n",
      "Training Loss: 0.5306853627033807, Training Accuracy: 46.93%\n",
      "Test Loss: 0.5608491183904422, Test Accuracy: 43.90%\n",
      "----------------------------------------\n",
      "Epoch 170, Loss: 0.5309258153516384\n",
      "Accuracy: 46.87%\n",
      "Iteration 84:\n",
      "Training Loss: 0.5309258153516384, Training Accuracy: 46.90%\n",
      "Test Loss: 0.5608491180880272, Test Accuracy: 43.90%\n",
      "----------------------------------------\n",
      "Epoch 172, Loss: 0.5307816228927816\n",
      "Accuracy: 46.88%\n",
      "Iteration 85:\n",
      "Training Loss: 0.5307816228927816, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5604645030059838, Test Accuracy: 43.94%\n",
      "----------------------------------------\n",
      "Epoch 174, Loss: 0.5305469047869629\n",
      "Accuracy: 46.88%\n",
      "Iteration 86:\n",
      "Training Loss: 0.5305469047869629, Training Accuracy: 46.94%\n",
      "Test Loss: 0.5600798876212114, Test Accuracy: 43.98%\n",
      "----------------------------------------\n",
      "Epoch 176, Loss: 0.5306015506747985\n",
      "Accuracy: 46.87%\n",
      "Iteration 87:\n",
      "Training Loss: 0.5306015506747985, Training Accuracy: 46.94%\n",
      "Test Loss: 0.5596952722331681, Test Accuracy: 44.02%\n",
      "----------------------------------------\n",
      "Epoch 178, Loss: 0.5306015923241123\n",
      "Accuracy: 46.88%\n",
      "Iteration 88:\n",
      "Training Loss: 0.5306015923241123, Training Accuracy: 46.94%\n",
      "Test Loss: 0.5598875799289037, Test Accuracy: 44.00%\n",
      "----------------------------------------\n",
      "Epoch 180, Loss: 0.5303674127827356\n",
      "Accuracy: 46.92%\n",
      "Iteration 89:\n",
      "Training Loss: 0.5303674127827356, Training Accuracy: 46.96%\n",
      "Test Loss: 0.5593106568519807, Test Accuracy: 44.06%\n",
      "----------------------------------------\n",
      "Epoch 182, Loss: 0.5305072214335195\n",
      "Accuracy: 46.87%\n",
      "Iteration 90:\n",
      "Training Loss: 0.5305072214335195, Training Accuracy: 46.95%\n",
      "Test Loss: 0.5591183698983663, Test Accuracy: 44.08%\n",
      "----------------------------------------\n",
      "Epoch 184, Loss: 0.5302231730994941\n",
      "Accuracy: 46.92%\n",
      "Iteration 91:\n",
      "Training Loss: 0.5302231730994941, Training Accuracy: 46.98%\n",
      "Test Loss: 0.5589260414673652, Test Accuracy: 44.10%\n",
      "----------------------------------------\n",
      "Epoch 186, Loss: 0.5300789815328435\n",
      "Accuracy: 46.94%\n",
      "Iteration 92:\n",
      "Training Loss: 0.5300789815328435, Training Accuracy: 46.99%\n",
      "Test Loss: 0.5583491183904398, Test Accuracy: 44.15%\n",
      "----------------------------------------\n",
      "Epoch 188, Loss: 0.5302715470910166\n",
      "Accuracy: 46.90%\n",
      "Iteration 93:\n",
      "Training Loss: 0.5302715470910166, Training Accuracy: 46.97%\n",
      "Test Loss: 0.5579645030058268, Test Accuracy: 44.19%\n",
      "----------------------------------------\n",
      "Epoch 190, Loss: 0.5305598275346259\n",
      "Accuracy: 46.89%\n",
      "Iteration 94:\n",
      "Training Loss: 0.5305598275346259, Training Accuracy: 46.94%\n",
      "Test Loss: 0.5583491183904422, Test Accuracy: 44.15%\n",
      "----------------------------------------\n",
      "Epoch 192, Loss: 0.530607942088987\n",
      "Accuracy: 46.87%\n",
      "Iteration 95:\n",
      "Training Loss: 0.530607942088987, Training Accuracy: 46.94%\n",
      "Test Loss: 0.5579888031473145, Test Accuracy: 44.19%\n",
      "----------------------------------------\n",
      "Epoch 194, Loss: 0.5307041076912037\n",
      "Accuracy: 46.87%\n",
      "Iteration 96:\n",
      "Training Loss: 0.5307041076912037, Training Accuracy: 46.93%\n",
      "Test Loss: 0.557604187762699, Test Accuracy: 44.23%\n",
      "----------------------------------------\n",
      "Epoch 196, Loss: 0.5304637850273402\n",
      "Accuracy: 46.89%\n",
      "Iteration 97:\n",
      "Training Loss: 0.5304637850273402, Training Accuracy: 46.95%\n",
      "Test Loss: 0.557604187763216, Test Accuracy: 44.23%\n",
      "----------------------------------------\n",
      "Epoch 198, Loss: 0.5308003596169066\n",
      "Accuracy: 46.87%\n",
      "Iteration 98:\n",
      "Training Loss: 0.5308003596169066, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5574118800703914, Test Accuracy: 44.25%\n",
      "----------------------------------------\n",
      "Epoch 200, Loss: 0.5306081889263066\n",
      "Accuracy: 46.88%\n",
      "Iteration 99:\n",
      "Training Loss: 0.5306081889263066, Training Accuracy: 46.94%\n",
      "Test Loss: 0.557604187762699, Test Accuracy: 44.23%\n",
      "----------------------------------------\n",
      "Epoch 202, Loss: 0.5307523540243301\n",
      "Accuracy: 46.83%\n",
      "Iteration 100:\n",
      "Training Loss: 0.5307523540243301, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5574118382155638, Test Accuracy: 44.25%\n",
      "----------------------------------------\n",
      "Epoch 204, Loss: 0.5310408509357835\n",
      "Accuracy: 46.81%\n",
      "Iteration 101:\n",
      "Training Loss: 0.5310408509357835, Training Accuracy: 46.89%\n",
      "Test Loss: 0.5566426493011606, Test Accuracy: 44.33%\n",
      "----------------------------------------\n",
      "Epoch 206, Loss: 0.5309897381538303\n",
      "Accuracy: 46.83%\n",
      "Iteration 102:\n",
      "Training Loss: 0.5309897381538303, Training Accuracy: 46.90%\n",
      "Test Loss: 0.557027264685776, Test Accuracy: 44.29%\n",
      "----------------------------------------\n",
      "Epoch 208, Loss: 0.5308486127553759\n",
      "Accuracy: 46.86%\n",
      "Iteration 103:\n",
      "Training Loss: 0.5308486127553759, Training Accuracy: 46.91%\n",
      "Test Loss: 0.5562580339165453, Test Accuracy: 44.37%\n",
      "----------------------------------------\n",
      "Epoch 210, Loss: 0.5308005699795416\n",
      "Accuracy: 46.85%\n",
      "Iteration 104:\n",
      "Training Loss: 0.5308005699795416, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5558734185319298, Test Accuracy: 44.40%\n",
      "----------------------------------------\n",
      "Epoch 212, Loss: 0.5306563729692797\n",
      "Accuracy: 46.87%\n",
      "Iteration 105:\n",
      "Training Loss: 0.5306563729692797, Training Accuracy: 46.93%\n",
      "Test Loss: 0.5552963275794739, Test Accuracy: 44.46%\n",
      "----------------------------------------\n",
      "Epoch 214, Loss: 0.5309929358265266\n",
      "Accuracy: 46.83%\n",
      "Iteration 106:\n",
      "Training Loss: 0.5309929358265266, Training Accuracy: 46.90%\n",
      "Test Loss: 0.5547195723780837, Test Accuracy: 44.52%\n",
      "----------------------------------------\n",
      "Epoch 216, Loss: 0.5312814353383121\n",
      "Accuracy: 46.80%\n",
      "Iteration 107:\n",
      "Training Loss: 0.5312814353383121, Training Accuracy: 46.87%\n",
      "Test Loss: 0.5543349569934684, Test Accuracy: 44.56%\n",
      "----------------------------------------\n",
      "Epoch 218, Loss: 0.5310873888520153\n",
      "Accuracy: 46.85%\n",
      "Iteration 108:\n",
      "Training Loss: 0.5310873888520153, Training Accuracy: 46.89%\n",
      "Test Loss: 0.5543349569934684, Test Accuracy: 44.56%\n",
      "----------------------------------------\n",
      "Epoch 220, Loss: 0.5307518475390717\n",
      "Accuracy: 46.86%\n",
      "Iteration 109:\n",
      "Training Loss: 0.5307518475390717, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5531811108396222, Test Accuracy: 44.67%\n",
      "----------------------------------------\n",
      "Epoch 222, Loss: 0.5307046138090359\n",
      "Accuracy: 46.88%\n",
      "Iteration 110:\n",
      "Training Loss: 0.5307046138090359, Training Accuracy: 46.93%\n",
      "Test Loss: 0.553101223218409, Test Accuracy: 44.69%\n",
      "----------------------------------------\n",
      "Epoch 224, Loss: 0.5308969534155072\n",
      "Accuracy: 46.84%\n",
      "Iteration 111:\n",
      "Training Loss: 0.5308969534155072, Training Accuracy: 46.91%\n",
      "Test Loss: 0.5527166078337954, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 226, Loss: 0.5309444765723861\n",
      "Accuracy: 46.84%\n",
      "Iteration 112:\n",
      "Training Loss: 0.5309444765723861, Training Accuracy: 46.90%\n",
      "Test Loss: 0.5519473750713066, Test Accuracy: 44.81%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 228, Loss: 0.5310888826041636\n",
      "Accuracy: 46.80%\n",
      "Iteration 113:\n",
      "Training Loss: 0.5310888826041636, Training Accuracy: 46.89%\n",
      "Test Loss: 0.5509858386182822, Test Accuracy: 44.90%\n",
      "----------------------------------------\n",
      "Epoch 230, Loss: 0.5309931882723855\n",
      "Accuracy: 46.82%\n",
      "Iteration 114:\n",
      "Training Loss: 0.5309931882723855, Training Accuracy: 46.90%\n",
      "Test Loss: 0.5509726691194688, Test Accuracy: 44.90%\n",
      "----------------------------------------\n",
      "Epoch 232, Loss: 0.5310406776332458\n",
      "Accuracy: 46.84%\n",
      "Iteration 115:\n",
      "Training Loss: 0.5310406776332458, Training Accuracy: 46.89%\n",
      "Test Loss: 0.5514574501199138, Test Accuracy: 44.85%\n",
      "----------------------------------------\n",
      "Epoch 234, Loss: 0.5310894171894923\n",
      "Accuracy: 46.83%\n",
      "Iteration 116:\n",
      "Training Loss: 0.5310894171894923, Training Accuracy: 46.89%\n",
      "Test Loss: 0.551450341608853, Test Accuracy: 44.85%\n",
      "----------------------------------------\n",
      "Epoch 236, Loss: 0.5309452169981986\n",
      "Accuracy: 46.83%\n",
      "Iteration 117:\n",
      "Training Loss: 0.5309452169981986, Training Accuracy: 46.90%\n",
      "Test Loss: 0.55106572622424, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 238, Loss: 0.5307633751187933\n",
      "Accuracy: 46.83%\n",
      "Iteration 118:\n",
      "Training Loss: 0.5307633751187933, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5510657262242376, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 240, Loss: 0.5307529647707249\n",
      "Accuracy: 46.83%\n",
      "Iteration 119:\n",
      "Training Loss: 0.5307529647707249, Training Accuracy: 46.92%\n",
      "Test Loss: 0.5516426493011606, Test Accuracy: 44.83%\n",
      "----------------------------------------\n",
      "Epoch 242, Loss: 0.5304164613410651\n",
      "Accuracy: 46.87%\n",
      "Iteration 120:\n",
      "Training Loss: 0.5304164613410651, Training Accuracy: 46.96%\n",
      "Test Loss: 0.5518349569934683, Test Accuracy: 44.81%\n",
      "----------------------------------------\n",
      "Epoch 244, Loss: 0.5306439666812357\n",
      "Accuracy: 46.88%\n",
      "Iteration 121:\n",
      "Training Loss: 0.5306439666812357, Training Accuracy: 46.93%\n",
      "Test Loss: 0.5522195723780837, Test Accuracy: 44.77%\n",
      "----------------------------------------\n",
      "Epoch 246, Loss: 0.530274701127002\n",
      "Accuracy: 46.90%\n",
      "Iteration 122:\n",
      "Training Loss: 0.530274701127002, Training Accuracy: 46.97%\n",
      "Test Loss: 0.5529888031473145, Test Accuracy: 44.69%\n",
      "----------------------------------------\n",
      "Epoch 248, Loss: 0.5303656375109559\n",
      "Accuracy: 46.92%\n",
      "Iteration 123:\n",
      "Training Loss: 0.5303656375109559, Training Accuracy: 46.96%\n",
      "Test Loss: 0.5531811108396222, Test Accuracy: 44.67%\n",
      "----------------------------------------\n",
      "Epoch 250, Loss: 0.5302273051693687\n",
      "Accuracy: 46.94%\n",
      "Iteration 124:\n",
      "Training Loss: 0.5302273051693687, Training Accuracy: 46.98%\n",
      "Test Loss: 0.552604187762699, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 252, Loss: 0.5302784488620801\n",
      "Accuracy: 46.92%\n",
      "Iteration 125:\n",
      "Training Loss: 0.5302784488620801, Training Accuracy: 46.97%\n",
      "Test Loss: 0.552027264685776, Test Accuracy: 44.79%\n",
      "----------------------------------------\n",
      "Epoch 254, Loss: 0.5301823278282027\n",
      "Accuracy: 46.93%\n",
      "Iteration 126:\n",
      "Training Loss: 0.5301823278282027, Training Accuracy: 46.98%\n",
      "Test Loss: 0.5520272646857851, Test Accuracy: 44.79%\n",
      "----------------------------------------\n",
      "Epoch 256, Loss: 0.5299900379522082\n",
      "Accuracy: 46.96%\n",
      "Iteration 127:\n",
      "Training Loss: 0.5299900379522082, Training Accuracy: 47.00%\n",
      "Test Loss: 0.5527964954551495, Test Accuracy: 44.71%\n",
      "----------------------------------------\n",
      "Epoch 258, Loss: 0.5299420005316899\n",
      "Accuracy: 46.96%\n",
      "Iteration 128:\n",
      "Training Loss: 0.5299420005316899, Training Accuracy: 47.00%\n",
      "Test Loss: 0.5526040421900023, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 260, Loss: 0.5299901058448123\n",
      "Accuracy: 46.94%\n",
      "Iteration 129:\n",
      "Training Loss: 0.5299901058448123, Training Accuracy: 47.00%\n",
      "Test Loss: 0.5527964954550068, Test Accuracy: 44.71%\n",
      "----------------------------------------\n",
      "Epoch 262, Loss: 0.5297475854567212\n",
      "Accuracy: 46.95%\n",
      "Iteration 130:\n",
      "Training Loss: 0.5297475854567212, Training Accuracy: 47.02%\n",
      "Test Loss: 0.552604187762699, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 264, Loss: 0.5298976851914196\n",
      "Accuracy: 46.93%\n",
      "Iteration 131:\n",
      "Training Loss: 0.5298976851914196, Training Accuracy: 47.01%\n",
      "Test Loss: 0.5524118800703914, Test Accuracy: 44.75%\n",
      "----------------------------------------\n",
      "Epoch 266, Loss: 0.5301344207176173\n",
      "Accuracy: 46.93%\n",
      "Iteration 132:\n",
      "Training Loss: 0.5301344207176173, Training Accuracy: 46.99%\n",
      "Test Loss: 0.552604187762699, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 268, Loss: 0.5297031523160233\n",
      "Accuracy: 46.95%\n",
      "Iteration 133:\n",
      "Training Loss: 0.5297031523160233, Training Accuracy: 47.03%\n",
      "Test Loss: 0.5527964954550068, Test Accuracy: 44.71%\n",
      "----------------------------------------\n",
      "Epoch 270, Loss: 0.5294133220362793\n",
      "Accuracy: 46.99%\n",
      "Iteration 134:\n",
      "Training Loss: 0.5294133220362793, Training Accuracy: 47.06%\n",
      "Test Loss: 0.5522195723780837, Test Accuracy: 44.77%\n",
      "----------------------------------------\n",
      "Epoch 272, Loss: 0.5294333213110899\n",
      "Accuracy: 47.01%\n",
      "Iteration 135:\n",
      "Training Loss: 0.5294333213110899, Training Accuracy: 47.05%\n",
      "Test Loss: 0.5522253760934986, Test Accuracy: 44.75%\n",
      "----------------------------------------\n",
      "Epoch 274, Loss: 0.5290487330946161\n",
      "Accuracy: 47.04%\n",
      "Iteration 136:\n",
      "Training Loss: 0.5290487330946161, Training Accuracy: 47.09%\n",
      "Test Loss: 0.551722536922372, Test Accuracy: 44.81%\n",
      "----------------------------------------\n",
      "Epoch 276, Loss: 0.5285245787971227\n",
      "Accuracy: 47.07%\n",
      "Iteration 137:\n",
      "Training Loss: 0.5285245787971227, Training Accuracy: 47.14%\n",
      "Test Loss: 0.5522994599992951, Test Accuracy: 44.75%\n",
      "----------------------------------------\n",
      "Epoch 278, Loss: 0.5285199407976285\n",
      "Accuracy: 47.08%\n",
      "Iteration 138:\n",
      "Training Loss: 0.5285199407976285, Training Accuracy: 47.14%\n",
      "Test Loss: 0.5522994599992951, Test Accuracy: 44.75%\n",
      "----------------------------------------\n",
      "Epoch 280, Loss: 0.528664318519258\n",
      "Accuracy: 47.05%\n",
      "Iteration 139:\n",
      "Training Loss: 0.528664318519258, Training Accuracy: 47.13%\n",
      "Test Loss: 0.553068690730102, Test Accuracy: 44.67%\n",
      "----------------------------------------\n",
      "Epoch 282, Loss: 0.5286642248267516\n",
      "Accuracy: 47.09%\n",
      "Iteration 140:\n",
      "Training Loss: 0.5286642248267516, Training Accuracy: 47.13%\n",
      "Test Loss: 0.5524917677995158, Test Accuracy: 44.73%\n",
      "----------------------------------------\n",
      "Epoch 284, Loss: 0.5281354051049987\n",
      "Accuracy: 47.11%\n",
      "Iteration 141:\n",
      "Training Loss: 0.5281354051049987, Training Accuracy: 47.18%\n",
      "Test Loss: 0.5513379215377566, Test Accuracy: 44.85%\n",
      "----------------------------------------\n",
      "Epoch 286, Loss: 0.5281342881278354\n",
      "Accuracy: 47.11%\n",
      "Iteration 142:\n",
      "Training Loss: 0.5281342881278354, Training Accuracy: 47.18%\n",
      "Test Loss: 0.550977606294629, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 288, Loss: 0.5283277652965234\n",
      "Accuracy: 47.08%\n",
      "Iteration 143:\n",
      "Training Loss: 0.5283277652965234, Training Accuracy: 47.16%\n",
      "Test Loss: 0.5505929909100137, Test Accuracy: 44.92%\n",
      "----------------------------------------\n",
      "Epoch 290, Loss: 0.5280393298191624\n",
      "Accuracy: 47.09%\n",
      "Iteration 144:\n",
      "Training Loss: 0.5280393298191624, Training Accuracy: 47.19%\n",
      "Test Loss: 0.550400683217706, Test Accuracy: 44.94%\n",
      "----------------------------------------\n",
      "Epoch 292, Loss: 0.5283278172878466\n",
      "Accuracy: 47.08%\n",
      "Iteration 145:\n",
      "Training Loss: 0.5283278172878466, Training Accuracy: 47.16%\n",
      "Test Loss: 0.5511698570588519, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 294, Loss: 0.5281614713882735\n",
      "Accuracy: 47.09%\n",
      "Iteration 146:\n",
      "Training Loss: 0.5281614713882735, Training Accuracy: 47.18%\n",
      "Test Loss: 0.5511699139869367, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 296, Loss: 0.5278731464782984\n",
      "Accuracy: 47.12%\n",
      "Iteration 147:\n",
      "Training Loss: 0.5278731464782984, Training Accuracy: 47.21%\n",
      "Test Loss: 0.5511699139869367, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 298, Loss: 0.5280654796942211\n",
      "Accuracy: 47.11%\n",
      "Iteration 148:\n",
      "Training Loss: 0.5280654796942211, Training Accuracy: 47.19%\n",
      "Test Loss: 0.5517468370638599, Test Accuracy: 44.81%\n",
      "----------------------------------------\n",
      "Epoch 300, Loss: 0.5281135817527931\n",
      "Accuracy: 47.12%\n",
      "Iteration 149:\n",
      "Training Loss: 0.5281135817527931, Training Accuracy: 47.18%\n",
      "Test Loss: 0.5515545293715521, Test Accuracy: 44.83%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 302, Loss: 0.5280174535464822\n",
      "Accuracy: 47.12%\n",
      "Iteration 150:\n",
      "Training Loss: 0.5280174535464822, Training Accuracy: 47.19%\n",
      "Test Loss: 0.551362221681078, Test Accuracy: 44.85%\n",
      "----------------------------------------\n",
      "Epoch 304, Loss: 0.5276807090586149\n",
      "Accuracy: 47.16%\n",
      "Iteration 151:\n",
      "Training Loss: 0.5276807090586149, Training Accuracy: 47.23%\n",
      "Test Loss: 0.551746836928536, Test Accuracy: 44.81%\n",
      "----------------------------------------\n",
      "Epoch 306, Loss: 0.5276328883441515\n",
      "Accuracy: 47.16%\n",
      "Iteration 152:\n",
      "Training Loss: 0.5276328883441515, Training Accuracy: 47.23%\n",
      "Test Loss: 0.5511699139869367, Test Accuracy: 44.87%\n",
      "----------------------------------------\n",
      "Epoch 308, Loss: 0.5277282979140481\n",
      "Accuracy: 47.16%\n",
      "Iteration 153:\n",
      "Training Loss: 0.5277282979140481, Training Accuracy: 47.22%\n",
      "Test Loss: 0.5515545269285441, Test Accuracy: 44.83%\n",
      "----------------------------------------\n",
      "Epoch 310, Loss: 0.5278252456743144\n",
      "Accuracy: 47.16%\n",
      "Iteration 154:\n",
      "Training Loss: 0.5278252456743144, Training Accuracy: 47.21%\n",
      "Test Loss: 0.550977606294629, Test Accuracy: 44.88%\n",
      "----------------------------------------\n",
      "Epoch 312, Loss: 0.527632962611966\n",
      "Accuracy: 47.17%\n",
      "Iteration 155:\n",
      "Training Loss: 0.527632962611966, Training Accuracy: 47.23%\n",
      "Test Loss: 0.5505929909100137, Test Accuracy: 44.92%\n",
      "----------------------------------------\n",
      "Epoch 314, Loss: 0.5273448050738376\n",
      "Accuracy: 47.19%\n",
      "Iteration 156:\n",
      "Training Loss: 0.5273448050738376, Training Accuracy: 47.26%\n",
      "Test Loss: 0.5507852986023916, Test Accuracy: 44.90%\n",
      "----------------------------------------\n",
      "Epoch 316, Loss: 0.5276311275048161\n",
      "Accuracy: 47.18%\n",
      "Iteration 157:\n",
      "Training Loss: 0.5276311275048161, Training Accuracy: 47.23%\n",
      "Test Loss: 0.5507852986023214, Test Accuracy: 44.90%\n",
      "----------------------------------------\n",
      "Epoch 318, Loss: 0.5271531021236714\n",
      "Accuracy: 47.22%\n",
      "Iteration 158:\n",
      "Training Loss: 0.5271531021236714, Training Accuracy: 47.28%\n",
      "Test Loss: 0.5507852986023214, Test Accuracy: 44.90%\n",
      "----------------------------------------\n",
      "Epoch 320, Loss: 0.5270080598229021\n",
      "Accuracy: 47.24%\n",
      "Iteration 159:\n",
      "Training Loss: 0.5270080598229021, Training Accuracy: 47.29%\n",
      "Test Loss: 0.5500147238132825, Test Accuracy: 44.98%\n",
      "----------------------------------------\n",
      "Epoch 322, Loss: 0.5266767202105205\n",
      "Accuracy: 47.25%\n",
      "Iteration 160:\n",
      "Training Loss: 0.5266767202105205, Training Accuracy: 47.33%\n",
      "Test Loss: 0.5490388871357295, Test Accuracy: 45.08%\n",
      "----------------------------------------\n",
      "Epoch 324, Loss: 0.526341105688082\n",
      "Accuracy: 47.28%\n",
      "Iteration 161:\n",
      "Training Loss: 0.526341105688082, Training Accuracy: 47.36%\n",
      "Test Loss: 0.5492279868998323, Test Accuracy: 45.06%\n",
      "----------------------------------------\n",
      "Epoch 326, Loss: 0.5263411293275412\n",
      "Accuracy: 47.27%\n",
      "Iteration 162:\n",
      "Training Loss: 0.5263411293275412, Training Accuracy: 47.36%\n",
      "Test Loss: 0.5490545293715521, Test Accuracy: 45.08%\n",
      "----------------------------------------\n",
      "Epoch 328, Loss: 0.526293075879518\n",
      "Accuracy: 47.28%\n",
      "Iteration 163:\n",
      "Training Loss: 0.526293075879518, Training Accuracy: 47.37%\n",
      "Test Loss: 0.5496314524484752, Test Accuracy: 45.02%\n",
      "----------------------------------------\n",
      "Epoch 330, Loss: 0.5262649941691846\n",
      "Accuracy: 47.27%\n",
      "Iteration 164:\n",
      "Training Loss: 0.5262649941691846, Training Accuracy: 47.37%\n",
      "Test Loss: 0.5492468370638598, Test Accuracy: 45.06%\n",
      "----------------------------------------\n",
      "Epoch 332, Loss: 0.5261280841085542\n",
      "Accuracy: 47.27%\n",
      "Iteration 165:\n",
      "Training Loss: 0.5261280841085542, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5482852986023213, Test Accuracy: 45.15%\n",
      "----------------------------------------\n",
      "Epoch 334, Loss: 0.5260666575948578\n",
      "Accuracy: 47.29%\n",
      "Iteration 166:\n",
      "Training Loss: 0.5260666575948578, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5476900131631367, Test Accuracy: 45.21%\n",
      "----------------------------------------\n",
      "Epoch 336, Loss: 0.5262589802920834\n",
      "Accuracy: 47.29%\n",
      "Iteration 167:\n",
      "Training Loss: 0.5262589802920834, Training Accuracy: 47.37%\n",
      "Test Loss: 0.5475160678087592, Test Accuracy: 45.23%\n",
      "----------------------------------------\n",
      "Epoch 338, Loss: 0.5261077203466524\n",
      "Accuracy: 47.28%\n",
      "Iteration 168:\n",
      "Training Loss: 0.5261077203466524, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5479006832177059, Test Accuracy: 45.19%\n",
      "----------------------------------------\n",
      "Epoch 340, Loss: 0.5259705716644687\n",
      "Accuracy: 47.28%\n",
      "Iteration 169:\n",
      "Training Loss: 0.5259705716644687, Training Accuracy: 47.39%\n",
      "Test Loss: 0.5477083755253983, Test Accuracy: 45.21%\n",
      "----------------------------------------\n",
      "Epoch 342, Loss: 0.5261687953624706\n",
      "Accuracy: 47.28%\n",
      "Iteration 170:\n",
      "Training Loss: 0.5261687953624706, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5480929909100136, Test Accuracy: 45.17%\n",
      "----------------------------------------\n",
      "Epoch 344, Loss: 0.5261684424739695\n",
      "Accuracy: 47.29%\n",
      "Iteration 171:\n",
      "Training Loss: 0.5261684424739695, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5479006827662011, Test Accuracy: 45.19%\n",
      "----------------------------------------\n",
      "Epoch 346, Loss: 0.5254733073855867\n",
      "Accuracy: 47.35%\n",
      "Iteration 172:\n",
      "Training Loss: 0.5254733073855867, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5475160678330906, Test Accuracy: 45.23%\n",
      "----------------------------------------\n",
      "Epoch 348, Loss: 0.5255440427670379\n",
      "Accuracy: 47.35%\n",
      "Iteration 173:\n",
      "Training Loss: 0.5255440427670379, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5473237601407829, Test Accuracy: 45.25%\n",
      "----------------------------------------\n",
      "Epoch 350, Loss: 0.5258325259743826\n",
      "Accuracy: 47.32%\n",
      "Iteration 174:\n",
      "Training Loss: 0.5258325259743826, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5467468370638598, Test Accuracy: 45.31%\n",
      "----------------------------------------\n",
      "Epoch 352, Loss: 0.5259287013352453\n",
      "Accuracy: 47.32%\n",
      "Iteration 175:\n",
      "Training Loss: 0.5259287013352453, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5475160647851639, Test Accuracy: 45.23%\n",
      "----------------------------------------\n",
      "Epoch 354, Loss: 0.5259767995940383\n",
      "Accuracy: 47.31%\n",
      "Iteration 176:\n",
      "Training Loss: 0.5259767995940383, Training Accuracy: 47.39%\n",
      "Test Loss: 0.5473237601402595, Test Accuracy: 45.25%\n",
      "----------------------------------------\n",
      "Epoch 356, Loss: 0.5258325900028425\n",
      "Accuracy: 47.29%\n",
      "Iteration 177:\n",
      "Training Loss: 0.5258325900028425, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5471314524484752, Test Accuracy: 45.27%\n",
      "----------------------------------------\n",
      "Epoch 358, Loss: 0.525688380252768\n",
      "Accuracy: 47.30%\n",
      "Iteration 178:\n",
      "Training Loss: 0.525688380252768, Training Accuracy: 47.42%\n",
      "Test Loss: 0.5469391447561676, Test Accuracy: 45.29%\n",
      "----------------------------------------\n",
      "Epoch 360, Loss: 0.5259803497484414\n",
      "Accuracy: 47.29%\n",
      "Iteration 179:\n",
      "Training Loss: 0.5259803497484414, Training Accuracy: 47.39%\n",
      "Test Loss: 0.5467468370638598, Test Accuracy: 45.31%\n",
      "----------------------------------------\n",
      "Epoch 362, Loss: 0.5258405537060713\n",
      "Accuracy: 47.31%\n",
      "Iteration 180:\n",
      "Training Loss: 0.5258405537060713, Training Accuracy: 47.41%\n",
      "Test Loss: 0.547323760140783, Test Accuracy: 45.25%\n",
      "----------------------------------------\n",
      "Epoch 364, Loss: 0.5257906712645841\n",
      "Accuracy: 47.32%\n",
      "Iteration 181:\n",
      "Training Loss: 0.5257906712645841, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5463622196068293, Test Accuracy: 45.35%\n",
      "----------------------------------------\n",
      "Epoch 366, Loss: 0.5257386499112645\n",
      "Accuracy: 47.32%\n",
      "Iteration 182:\n",
      "Training Loss: 0.5257386499112645, Training Accuracy: 47.42%\n",
      "Test Loss: 0.5465545293715522, Test Accuracy: 45.33%\n",
      "----------------------------------------\n",
      "Epoch 368, Loss: 0.5256464811365071\n",
      "Accuracy: 47.32%\n",
      "Iteration 183:\n",
      "Training Loss: 0.5256464811365071, Training Accuracy: 47.43%\n",
      "Test Loss: 0.5457852986023214, Test Accuracy: 45.40%\n",
      "----------------------------------------\n",
      "Epoch 370, Loss: 0.5257481482883476\n",
      "Accuracy: 47.29%\n",
      "Iteration 184:\n",
      "Training Loss: 0.5257481482883476, Training Accuracy: 47.42%\n",
      "Test Loss: 0.5460160542772643, Test Accuracy: 45.38%\n",
      "----------------------------------------\n",
      "Epoch 372, Loss: 0.5259349827293613\n",
      "Accuracy: 47.31%\n",
      "Iteration 185:\n",
      "Training Loss: 0.5259349827293613, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5452083755253982, Test Accuracy: 45.46%\n",
      "----------------------------------------\n",
      "Epoch 374, Loss: 0.5256663931793234\n",
      "Accuracy: 47.33%\n",
      "Iteration 186:\n",
      "Training Loss: 0.5256663931793234, Training Accuracy: 47.42%\n",
      "Test Loss: 0.5461699139869367, Test Accuracy: 45.37%\n",
      "----------------------------------------\n",
      "Epoch 376, Loss: 0.525762686397883\n",
      "Accuracy: 47.32%\n",
      "Iteration 187:\n",
      "Training Loss: 0.525762686397883, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5463622216792444, Test Accuracy: 45.35%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 378, Loss: 0.5258551956558863\n",
      "Accuracy: 47.33%\n",
      "Iteration 188:\n",
      "Training Loss: 0.5258551956558863, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5463622216792444, Test Accuracy: 45.35%\n",
      "----------------------------------------\n",
      "Epoch 380, Loss: 0.5257826971415822\n",
      "Accuracy: 47.32%\n",
      "Iteration 189:\n",
      "Training Loss: 0.5257826971415822, Training Accuracy: 47.41%\n",
      "Test Loss: 0.5462498016081482, Test Accuracy: 45.35%\n",
      "----------------------------------------\n",
      "Epoch 382, Loss: 0.5258307932597397\n",
      "Accuracy: 47.29%\n",
      "Iteration 190:\n",
      "Training Loss: 0.5258307932597397, Training Accuracy: 47.40%\n",
      "Test Loss: 0.5463296918424979, Test Accuracy: 45.33%\n",
      "----------------------------------------\n",
      "Epoch 384, Loss: 0.5260231200052676\n",
      "Accuracy: 47.25%\n",
      "Iteration 191:\n",
      "Training Loss: 0.5260231200052676, Training Accuracy: 47.38%\n",
      "Test Loss: 0.5455604584601288, Test Accuracy: 45.40%\n",
      "----------------------------------------\n",
      "Epoch 386, Loss: 0.5258308312325126\n",
      "Accuracy: 47.27%\n",
      "Iteration 192:\n",
      "Training Loss: 0.5258308312325126, Training Accuracy: 47.40%\n",
      "Test Loss: 0.545368150776654, Test Accuracy: 45.42%\n",
      "----------------------------------------\n",
      "Epoch 388, Loss: 0.5254462338221613\n",
      "Accuracy: 47.29%\n",
      "Iteration 193:\n",
      "Training Loss: 0.5254462338221613, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5457519566987536, Test Accuracy: 45.38%\n",
      "----------------------------------------\n",
      "Epoch 390, Loss: 0.52515779162264\n",
      "Accuracy: 47.35%\n",
      "Iteration 194:\n",
      "Training Loss: 0.52515779162264, Training Accuracy: 47.47%\n",
      "Test Loss: 0.5451758430755154, Test Accuracy: 45.44%\n",
      "----------------------------------------\n",
      "Epoch 392, Loss: 0.5248212718093671\n",
      "Accuracy: 47.38%\n",
      "Iteration 195:\n",
      "Training Loss: 0.5248212718093671, Training Accuracy: 47.50%\n",
      "Test Loss: 0.5455604584601288, Test Accuracy: 45.40%\n",
      "----------------------------------------\n",
      "Epoch 394, Loss: 0.5249655209849502\n",
      "Accuracy: 47.39%\n",
      "Iteration 196:\n",
      "Training Loss: 0.5249655209849502, Training Accuracy: 47.49%\n",
      "Test Loss: 0.5447840477958047, Test Accuracy: 45.48%\n",
      "----------------------------------------\n",
      "Epoch 396, Loss: 0.5247967551738248\n",
      "Accuracy: 47.40%\n",
      "Iteration 197:\n",
      "Training Loss: 0.5247967551738248, Training Accuracy: 47.50%\n",
      "Test Loss: 0.5444066123062704, Test Accuracy: 45.52%\n",
      "----------------------------------------\n",
      "Epoch 398, Loss: 0.5247868092090675\n",
      "Accuracy: 47.41%\n",
      "Iteration 198:\n",
      "Training Loss: 0.5247868092090675, Training Accuracy: 47.50%\n",
      "Test Loss: 0.5447896565927466, Test Accuracy: 45.48%\n",
      "----------------------------------------\n",
      "Epoch 400, Loss: 0.5249672239581012\n",
      "Accuracy: 47.38%\n",
      "Iteration 199:\n",
      "Training Loss: 0.5249672239581012, Training Accuracy: 47.48%\n",
      "Test Loss: 0.5442143044651385, Test Accuracy: 45.54%\n",
      "----------------------------------------\n",
      "Epoch 402, Loss: 0.5247570217159027\n",
      "Accuracy: 47.38%\n",
      "Iteration 200:\n",
      "Training Loss: 0.5247570217159027, Training Accuracy: 47.50%\n",
      "Test Loss: 0.5440117755745563, Test Accuracy: 45.56%\n",
      "----------------------------------------\n",
      "Epoch 404, Loss: 0.524708960626855\n",
      "Accuracy: 47.39%\n",
      "Iteration 201:\n",
      "Training Loss: 0.524708960626855, Training Accuracy: 47.50%\n",
      "Test Loss: 0.5438296892293596, Test Accuracy: 45.58%\n",
      "----------------------------------------\n",
      "Epoch 406, Loss: 0.525139054173849\n",
      "Accuracy: 47.37%\n",
      "Iteration 202:\n",
      "Training Loss: 0.525139054173849, Training Accuracy: 47.46%\n",
      "Test Loss: 0.5434450738447442, Test Accuracy: 45.62%\n",
      "----------------------------------------\n",
      "Epoch 408, Loss: 0.5250454994595849\n",
      "Accuracy: 47.38%\n",
      "Iteration 203:\n",
      "Training Loss: 0.5250454994595849, Training Accuracy: 47.47%\n",
      "Test Loss: 0.5437172691580644, Test Accuracy: 45.58%\n",
      "----------------------------------------\n",
      "Epoch 410, Loss: 0.5250936290781665\n",
      "Accuracy: 47.37%\n",
      "Iteration 204:\n",
      "Training Loss: 0.5250936290781665, Training Accuracy: 47.47%\n",
      "Test Loss: 0.543909576850571, Test Accuracy: 45.56%\n",
      "----------------------------------------\n",
      "Epoch 412, Loss: 0.5253185285311618\n",
      "Accuracy: 47.33%\n",
      "Iteration 205:\n",
      "Training Loss: 0.5253185285311618, Training Accuracy: 47.44%\n",
      "Test Loss: 0.543909576850571, Test Accuracy: 45.56%\n",
      "----------------------------------------\n",
      "Epoch 414, Loss: 0.5253821267681028\n",
      "Accuracy: 47.34%\n",
      "Iteration 206:\n",
      "Training Loss: 0.5253821267681028, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5437172691582632, Test Accuracy: 45.58%\n",
      "----------------------------------------\n",
      "Epoch 416, Loss: 0.5253827609361502\n",
      "Accuracy: 47.34%\n",
      "Iteration 207:\n",
      "Training Loss: 0.5253827609361502, Training Accuracy: 47.44%\n",
      "Test Loss: 0.5423711153203075, Test Accuracy: 45.71%\n",
      "----------------------------------------\n",
      "Epoch 418, Loss: 0.5251715898965694\n",
      "Accuracy: 47.34%\n",
      "Iteration 208:\n",
      "Training Loss: 0.5251715898965694, Training Accuracy: 47.46%\n",
      "Test Loss: 0.5429480383412341, Test Accuracy: 45.65%\n",
      "----------------------------------------\n",
      "Epoch 420, Loss: 0.5249384893042122\n",
      "Accuracy: 47.36%\n",
      "Iteration 209:\n",
      "Training Loss: 0.5249384893042122, Training Accuracy: 47.48%\n",
      "Test Loss: 0.5426433432401803, Test Accuracy: 45.67%\n",
      "----------------------------------------\n",
      "Epoch 422, Loss: 0.5248312309546677\n",
      "Accuracy: 47.39%\n",
      "Iteration 210:\n",
      "Training Loss: 0.5248312309546677, Training Accuracy: 47.49%\n",
      "Test Loss: 0.5426433106256285, Test Accuracy: 45.67%\n",
      "----------------------------------------\n",
      "Epoch 424, Loss: 0.5249274883983668\n",
      "Accuracy: 47.39%\n",
      "Iteration 211:\n",
      "Training Loss: 0.5249274883983668, Training Accuracy: 47.48%\n",
      "Test Loss: 0.5426237651962029, Test Accuracy: 45.67%\n",
      "----------------------------------------\n",
      "Epoch 426, Loss: 0.5251678899010044\n",
      "Accuracy: 47.37%\n",
      "Iteration 212:\n",
      "Training Loss: 0.5251678899010044, Training Accuracy: 47.46%\n",
      "Test Loss: 0.5430279260102439, Test Accuracy: 45.63%\n",
      "----------------------------------------\n",
      "Epoch 428, Loss: 0.5254563680944012\n",
      "Accuracy: 47.36%\n",
      "Iteration 213:\n",
      "Training Loss: 0.5254563680944012, Training Accuracy: 47.43%\n",
      "Test Loss: 0.5429155059391476, Test Accuracy: 45.63%\n",
      "----------------------------------------\n",
      "Epoch 430, Loss: 0.5251198456144165\n",
      "Accuracy: 47.35%\n",
      "Iteration 214:\n",
      "Training Loss: 0.5251198456144165, Training Accuracy: 47.46%\n",
      "Test Loss: 0.5427232415500519, Test Accuracy: 45.65%\n",
      "----------------------------------------\n",
      "Epoch 432, Loss: 0.5250717851767752\n",
      "Accuracy: 47.38%\n",
      "Iteration 215:\n",
      "Training Loss: 0.5250717851767752, Training Accuracy: 47.47%\n",
      "Test Loss: 0.5434924256187667, Test Accuracy: 45.58%\n",
      "----------------------------------------\n",
      "Epoch 434, Loss: 0.5245429026569414\n",
      "Accuracy: 47.41%\n",
      "Iteration 216:\n",
      "Training Loss: 0.5245429026569414, Training Accuracy: 47.52%\n",
      "Test Loss: 0.5433001213138527, Test Accuracy: 45.60%\n",
      "----------------------------------------\n",
      "Epoch 436, Loss: 0.524082174867296\n",
      "Accuracy: 47.44%\n",
      "Iteration 217:\n",
      "Training Loss: 0.524082174867296, Training Accuracy: 47.56%\n",
      "Test Loss: 0.5431078136314353, Test Accuracy: 45.62%\n",
      "----------------------------------------\n",
      "Epoch 438, Loss: 0.5239860376085556\n",
      "Accuracy: 47.47%\n",
      "Iteration 218:\n",
      "Training Loss: 0.5239860376085556, Training Accuracy: 47.57%\n",
      "Test Loss: 0.5427231982463653, Test Accuracy: 45.65%\n",
      "----------------------------------------\n",
      "Epoch 440, Loss: 0.5237937429277927\n",
      "Accuracy: 47.48%\n",
      "Iteration 219:\n",
      "Training Loss: 0.5237937429277927, Training Accuracy: 47.59%\n",
      "Test Loss: 0.5420538947323241, Test Accuracy: 45.71%\n",
      "----------------------------------------\n",
      "Epoch 442, Loss: 0.5232649154146521\n",
      "Accuracy: 47.53%\n",
      "Iteration 220:\n",
      "Training Loss: 0.5232649154146521, Training Accuracy: 47.64%\n",
      "Test Loss: 0.5415981515805989, Test Accuracy: 45.77%\n",
      "----------------------------------------\n",
      "Epoch 444, Loss: 0.5234572390135269\n",
      "Accuracy: 47.52%\n",
      "Iteration 221:\n",
      "Training Loss: 0.5234572390135269, Training Accuracy: 47.62%\n",
      "Test Loss: 0.5413770444006857, Test Accuracy: 45.79%\n",
      "----------------------------------------\n",
      "Epoch 446, Loss: 0.523264947142625\n",
      "Accuracy: 47.54%\n",
      "Iteration 222:\n",
      "Training Loss: 0.523264947142625, Training Accuracy: 47.64%\n",
      "Test Loss: 0.5411847367083784, Test Accuracy: 45.81%\n",
      "----------------------------------------\n",
      "Epoch 448, Loss: 0.5232459416945657\n",
      "Accuracy: 47.55%\n",
      "Iteration 223:\n",
      "Training Loss: 0.5232459416945657, Training Accuracy: 47.65%\n",
      "Test Loss: 0.5401107781757436, Test Accuracy: 45.90%\n",
      "----------------------------------------\n",
      "Epoch 450, Loss: 0.5232719673289354\n",
      "Accuracy: 47.54%\n",
      "Iteration 224:\n",
      "Training Loss: 0.5232719673289354, Training Accuracy: 47.64%\n",
      "Test Loss: 0.540495393560359, Test Accuracy: 45.87%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 452, Loss: 0.5231688403098333\n",
      "Accuracy: 47.54%\n",
      "Iteration 225:\n",
      "Training Loss: 0.5231688403098333, Training Accuracy: 47.65%\n",
      "Test Loss: 0.539918470483436, Test Accuracy: 45.92%\n",
      "----------------------------------------\n",
      "Epoch 454, Loss: 0.5228323173727951\n",
      "Accuracy: 47.60%\n",
      "Iteration 226:\n",
      "Training Loss: 0.5228323173727951, Training Accuracy: 47.69%\n",
      "Test Loss: 0.5401107781757436, Test Accuracy: 45.90%\n",
      "----------------------------------------\n",
      "Epoch 456, Loss: 0.5229064566856763\n",
      "Accuracy: 47.60%\n",
      "Iteration 227:\n",
      "Training Loss: 0.5229064566856763, Training Accuracy: 47.68%\n",
      "Test Loss: 0.540110778175852, Test Accuracy: 45.90%\n",
      "----------------------------------------\n",
      "Epoch 458, Loss: 0.5227596441402061\n",
      "Accuracy: 47.56%\n",
      "Iteration 228:\n",
      "Training Loss: 0.5227596441402061, Training Accuracy: 47.69%\n",
      "Test Loss: 0.540495393560359, Test Accuracy: 45.87%\n",
      "----------------------------------------\n",
      "Epoch 460, Loss: 0.5226860738609511\n",
      "Accuracy: 47.60%\n",
      "Iteration 229:\n",
      "Training Loss: 0.5226860738609511, Training Accuracy: 47.70%\n",
      "Test Loss: 0.5403030858680513, Test Accuracy: 45.88%\n",
      "----------------------------------------\n",
      "Epoch 462, Loss: 0.5224998572351918\n",
      "Accuracy: 47.59%\n",
      "Iteration 230:\n",
      "Training Loss: 0.5224998572351918, Training Accuracy: 47.72%\n",
      "Test Loss: 0.5391741588288367, Test Accuracy: 46.00%\n",
      "----------------------------------------\n",
      "Epoch 464, Loss: 0.5223196689587398\n",
      "Accuracy: 47.60%\n",
      "Iteration 231:\n",
      "Training Loss: 0.5223196689587398, Training Accuracy: 47.74%\n",
      "Test Loss: 0.5397556555569134, Test Accuracy: 45.94%\n",
      "----------------------------------------\n",
      "Epoch 466, Loss: 0.5217427940200596\n",
      "Accuracy: 47.63%\n",
      "Iteration 232:\n",
      "Training Loss: 0.5217427940200596, Training Accuracy: 47.79%\n",
      "Test Loss: 0.5393658715721004, Test Accuracy: 45.98%\n",
      "----------------------------------------\n",
      "Epoch 468, Loss: 0.5215980046799947\n",
      "Accuracy: 47.67%\n",
      "Iteration 233:\n",
      "Training Loss: 0.5215980046799947, Training Accuracy: 47.81%\n",
      "Test Loss: 0.5395581552403067, Test Accuracy: 45.96%\n",
      "----------------------------------------\n",
      "Epoch 470, Loss: 0.5214462420642597\n",
      "Accuracy: 47.65%\n",
      "Iteration 234:\n",
      "Training Loss: 0.5214462420642597, Training Accuracy: 47.82%\n",
      "Test Loss: 0.5405196937018468, Test Accuracy: 45.87%\n",
      "----------------------------------------\n",
      "Epoch 472, Loss: 0.5217066134272019\n",
      "Accuracy: 47.65%\n",
      "Iteration 235:\n",
      "Training Loss: 0.5217066134272019, Training Accuracy: 47.79%\n",
      "Test Loss: 0.5403273860092697, Test Accuracy: 45.88%\n",
      "----------------------------------------\n",
      "Epoch 474, Loss: 0.5213271383400595\n",
      "Accuracy: 47.67%\n",
      "Iteration 236:\n",
      "Training Loss: 0.5213271383400595, Training Accuracy: 47.83%\n",
      "Test Loss: 0.540712001264007, Test Accuracy: 45.85%\n",
      "----------------------------------------\n",
      "Epoch 476, Loss: 0.5211899467988884\n",
      "Accuracy: 47.68%\n",
      "Iteration 237:\n",
      "Training Loss: 0.5211899467988884, Training Accuracy: 47.84%\n",
      "Test Loss: 0.5401350783172314, Test Accuracy: 45.90%\n",
      "----------------------------------------\n",
      "Epoch 478, Loss: 0.521065702532252\n",
      "Accuracy: 47.70%\n",
      "Iteration 238:\n",
      "Training Loss: 0.521065702532252, Training Accuracy: 47.85%\n",
      "Test Loss: 0.5398303505538274, Test Accuracy: 45.92%\n",
      "----------------------------------------\n",
      "Epoch 480, Loss: 0.5205829625772759\n",
      "Accuracy: 47.73%\n",
      "Iteration 239:\n",
      "Training Loss: 0.5205829625772759, Training Accuracy: 47.90%\n",
      "Test Loss: 0.5390611197845967, Test Accuracy: 46.00%\n",
      "----------------------------------------\n",
      "Epoch 482, Loss: 0.5206871896439688\n",
      "Accuracy: 47.75%\n",
      "Iteration 240:\n",
      "Training Loss: 0.5206871896439688, Training Accuracy: 47.89%\n",
      "Test Loss: 0.5391410074058076, Test Accuracy: 45.98%\n",
      "----------------------------------------\n",
      "Epoch 484, Loss: 0.5205490500426423\n",
      "Accuracy: 47.75%\n",
      "Iteration 241:\n",
      "Training Loss: 0.5205490500426423, Training Accuracy: 47.90%\n",
      "Test Loss: 0.5387474706417406, Test Accuracy: 46.02%\n",
      "----------------------------------------\n",
      "Epoch 486, Loss: 0.5206017513561526\n",
      "Accuracy: 47.76%\n",
      "Iteration 242:\n",
      "Training Loss: 0.5206017513561526, Training Accuracy: 47.90%\n",
      "Test Loss: 0.538564084328885, Test Accuracy: 46.04%\n",
      "----------------------------------------\n",
      "Epoch 488, Loss: 0.5202524841061916\n",
      "Accuracy: 47.78%\n",
      "Iteration 243:\n",
      "Training Loss: 0.5202524841061916, Training Accuracy: 47.93%\n",
      "Test Loss: 0.5383717766365773, Test Accuracy: 46.06%\n",
      "----------------------------------------\n",
      "Epoch 490, Loss: 0.5201563445050238\n",
      "Accuracy: 47.81%\n",
      "Iteration 244:\n",
      "Training Loss: 0.5201563445050238, Training Accuracy: 47.94%\n",
      "Test Loss: 0.5377948535596542, Test Accuracy: 46.12%\n",
      "----------------------------------------\n",
      "Epoch 492, Loss: 0.520211212471371\n",
      "Accuracy: 47.79%\n",
      "Iteration 245:\n",
      "Training Loss: 0.520211212471371, Training Accuracy: 47.93%\n",
      "Test Loss: 0.5376025458673466, Test Accuracy: 46.13%\n",
      "----------------------------------------\n",
      "Epoch 494, Loss: 0.5201343427783122\n",
      "Accuracy: 47.77%\n",
      "Iteration 246:\n",
      "Training Loss: 0.5201343427783122, Training Accuracy: 47.94%\n",
      "Test Loss: 0.5366410074058081, Test Accuracy: 46.23%\n",
      "----------------------------------------\n",
      "Epoch 496, Loss: 0.5201543733243589\n",
      "Accuracy: 47.77%\n",
      "Iteration 247:\n",
      "Training Loss: 0.5201543733243589, Training Accuracy: 47.93%\n",
      "Test Loss: 0.5368333150981157, Test Accuracy: 46.21%\n",
      "----------------------------------------\n",
      "Epoch 498, Loss: 0.5202024067274124\n",
      "Accuracy: 47.79%\n",
      "Iteration 248:\n",
      "Training Loss: 0.5202024067274124, Training Accuracy: 47.93%\n",
      "Test Loss: 0.5372179301045512, Test Accuracy: 46.17%\n",
      "----------------------------------------\n",
      "Epoch 500, Loss: 0.5201262514933854\n",
      "Accuracy: 47.77%\n",
      "Iteration 249:\n",
      "Training Loss: 0.5201262514933854, Training Accuracy: 47.93%\n",
      "Test Loss: 0.5366409915044793, Test Accuracy: 46.23%\n",
      "----------------------------------------\n",
      "Epoch 502, Loss: 0.5199882157779456\n",
      "Accuracy: 47.76%\n",
      "Iteration 250:\n",
      "Training Loss: 0.5199882157779456, Training Accuracy: 47.95%\n",
      "Test Loss: 0.5360012123037526, Test Accuracy: 46.29%\n",
      "----------------------------------------\n",
      "Epoch 504, Loss: 0.5198560425628834\n",
      "Accuracy: 47.76%\n",
      "Iteration 251:\n",
      "Training Loss: 0.5198560425628834, Training Accuracy: 47.96%\n",
      "Test Loss: 0.5358717766365773, Test Accuracy: 46.31%\n",
      "----------------------------------------\n",
      "Epoch 506, Loss: 0.5197799399876605\n",
      "Accuracy: 47.76%\n",
      "Iteration 252:\n",
      "Training Loss: 0.5197799399876605, Training Accuracy: 47.97%\n",
      "Test Loss: 0.5360880056183501, Test Accuracy: 46.29%\n",
      "----------------------------------------\n",
      "Epoch 508, Loss: 0.5196473969811113\n",
      "Accuracy: 47.77%\n",
      "Iteration 253:\n",
      "Training Loss: 0.5196473969811113, Training Accuracy: 47.98%\n",
      "Test Loss: 0.5360883844687439, Test Accuracy: 46.29%\n",
      "----------------------------------------\n",
      "Epoch 510, Loss: 0.5195516667465077\n",
      "Accuracy: 47.77%\n",
      "Iteration 254:\n",
      "Training Loss: 0.5195516667465077, Training Accuracy: 47.99%\n",
      "Test Loss: 0.5359203743019346, Test Accuracy: 46.31%\n",
      "----------------------------------------\n",
      "Epoch 512, Loss: 0.5197439878238488\n",
      "Accuracy: 47.75%\n",
      "Iteration 255:\n",
      "Training Loss: 0.5197439878238488, Training Accuracy: 47.97%\n",
      "Test Loss: 0.5361126846118605, Test Accuracy: 46.29%\n",
      "----------------------------------------\n",
      "Epoch 514, Loss: 0.519840154965654\n",
      "Accuracy: 47.77%\n",
      "Iteration 256:\n",
      "Training Loss: 0.519840154965654, Training Accuracy: 47.96%\n",
      "Test Loss: 0.5359203753256522, Test Accuracy: 46.31%\n",
      "----------------------------------------\n",
      "Epoch 516, Loss: 0.519551706963475\n",
      "Accuracy: 47.79%\n",
      "Iteration 257:\n",
      "Training Loss: 0.519551706963475, Training Accuracy: 47.99%\n",
      "Test Loss: 0.535728069227245, Test Accuracy: 46.33%\n",
      "----------------------------------------\n",
      "Epoch 518, Loss: 0.5193313070494815\n",
      "Accuracy: 47.81%\n",
      "Iteration 258:\n",
      "Training Loss: 0.5193313070494815, Training Accuracy: 48.01%\n",
      "Test Loss: 0.535728069227245, Test Accuracy: 46.33%\n",
      "----------------------------------------\n",
      "Epoch 520, Loss: 0.5192551317662251\n",
      "Accuracy: 47.82%\n",
      "Iteration 259:\n",
      "Training Loss: 0.5192551317662251, Training Accuracy: 48.01%\n",
      "Test Loss: 0.535560061676425, Test Accuracy: 46.35%\n",
      "----------------------------------------\n",
      "Epoch 522, Loss: 0.5190628433868236\n",
      "Accuracy: 47.85%\n",
      "Iteration 260:\n",
      "Training Loss: 0.5190628433868236, Training Accuracy: 48.03%\n",
      "Test Loss: 0.5351754462918097, Test Accuracy: 46.38%\n",
      "----------------------------------------\n",
      "Epoch 524, Loss: 0.5186782408685598\n",
      "Accuracy: 47.88%\n",
      "Iteration 261:\n",
      "Training Loss: 0.5186782408685598, Training Accuracy: 48.07%\n",
      "Test Loss: 0.534983138599502, Test Accuracy: 46.40%\n",
      "----------------------------------------\n",
      "Epoch 526, Loss: 0.5185059165743342\n",
      "Accuracy: 47.89%\n",
      "Iteration 262:\n",
      "Training Loss: 0.5185059165743342, Training Accuracy: 48.09%\n",
      "Test Loss: 0.534983138599502, Test Accuracy: 46.40%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 528, Loss: 0.5185734758643791\n",
      "Accuracy: 47.87%\n",
      "Iteration 263:\n",
      "Training Loss: 0.5185734758643791, Training Accuracy: 48.08%\n",
      "Test Loss: 0.5342145021531157, Test Accuracy: 46.48%\n",
      "----------------------------------------\n",
      "Epoch 530, Loss: 0.518738194056817\n",
      "Accuracy: 47.84%\n",
      "Iteration 264:\n",
      "Training Loss: 0.518738194056817, Training Accuracy: 48.06%\n",
      "Test Loss: 0.5342937954514804, Test Accuracy: 46.46%\n",
      "----------------------------------------\n",
      "Epoch 532, Loss: 0.518179091850393\n",
      "Accuracy: 47.87%\n",
      "Iteration 265:\n",
      "Training Loss: 0.518179091850393, Training Accuracy: 48.11%\n",
      "Test Loss: 0.5339091800646447, Test Accuracy: 46.50%\n",
      "----------------------------------------\n",
      "Epoch 534, Loss: 0.5181650673034139\n",
      "Accuracy: 47.90%\n",
      "Iteration 266:\n",
      "Training Loss: 0.5181650673034139, Training Accuracy: 48.11%\n",
      "Test Loss: 0.5339334802083548, Test Accuracy: 46.50%\n",
      "----------------------------------------\n",
      "Epoch 536, Loss: 0.5178070974886251\n",
      "Accuracy: 47.93%\n",
      "Iteration 267:\n",
      "Training Loss: 0.5178070974886251, Training Accuracy: 48.15%\n",
      "Test Loss: 0.5341257879006539, Test Accuracy: 46.48%\n",
      "----------------------------------------\n",
      "Epoch 538, Loss: 0.517692390817855\n",
      "Accuracy: 47.97%\n",
      "Iteration 268:\n",
      "Training Loss: 0.517692390817855, Training Accuracy: 48.15%\n",
      "Test Loss: 0.5335734209503702, Test Accuracy: 46.54%\n",
      "----------------------------------------\n",
      "Epoch 540, Loss: 0.5174641813605918\n",
      "Accuracy: 47.97%\n",
      "Iteration 269:\n",
      "Training Loss: 0.5174641813605918, Training Accuracy: 48.18%\n",
      "Test Loss: 0.5335731649652273, Test Accuracy: 46.54%\n",
      "----------------------------------------\n",
      "Epoch 542, Loss: 0.517057535262328\n",
      "Accuracy: 47.98%\n",
      "Iteration 270:\n",
      "Training Loss: 0.517057535262328, Training Accuracy: 48.22%\n",
      "Test Loss: 0.5335974596203453, Test Accuracy: 46.54%\n",
      "----------------------------------------\n",
      "Epoch 544, Loss: 0.5169332884808343\n",
      "Accuracy: 48.01%\n",
      "Iteration 271:\n",
      "Training Loss: 0.5169332884808343, Training Accuracy: 48.23%\n",
      "Test Loss: 0.5330205420297919, Test Accuracy: 46.60%\n",
      "----------------------------------------\n",
      "Epoch 546, Loss: 0.5165608351852184\n",
      "Accuracy: 48.03%\n",
      "Iteration 272:\n",
      "Training Loss: 0.5165608351852184, Training Accuracy: 48.26%\n",
      "Test Loss: 0.5327158142663879, Test Accuracy: 46.62%\n",
      "----------------------------------------\n",
      "Epoch 548, Loss: 0.5165188452563276\n",
      "Accuracy: 48.03%\n",
      "Iteration 273:\n",
      "Training Loss: 0.5165188452563276, Training Accuracy: 48.27%\n",
      "Test Loss: 0.532411086502984, Test Accuracy: 46.63%\n",
      "----------------------------------------\n",
      "Epoch 550, Loss: 0.5165249244301786\n",
      "Accuracy: 48.03%\n",
      "Iteration 274:\n",
      "Training Loss: 0.5165249244301786, Training Accuracy: 48.27%\n",
      "Test Loss: 0.53210635873958, Test Accuracy: 46.65%\n",
      "----------------------------------------\n",
      "Epoch 552, Loss: 0.516675249854247\n",
      "Accuracy: 48.05%\n",
      "Iteration 275:\n",
      "Training Loss: 0.516675249854247, Training Accuracy: 48.25%\n",
      "Test Loss: 0.5315294356616801, Test Accuracy: 46.71%\n",
      "----------------------------------------\n",
      "Epoch 554, Loss: 0.516528972810521\n",
      "Accuracy: 48.07%\n",
      "Iteration 276:\n",
      "Training Loss: 0.516528972810521, Training Accuracy: 48.26%\n",
      "Test Loss: 0.5315294356626569, Test Accuracy: 46.71%\n",
      "----------------------------------------\n",
      "Epoch 556, Loss: 0.5165026682860396\n",
      "Accuracy: 48.05%\n",
      "Iteration 277:\n",
      "Training Loss: 0.5165026682860396, Training Accuracy: 48.26%\n",
      "Test Loss: 0.5309768127272215, Test Accuracy: 46.77%\n",
      "----------------------------------------\n",
      "Epoch 558, Loss: 0.516520842522606\n",
      "Accuracy: 48.04%\n",
      "Iteration 278:\n",
      "Training Loss: 0.516520842522606, Training Accuracy: 48.26%\n",
      "Test Loss: 0.5305921973426061, Test Accuracy: 46.81%\n",
      "----------------------------------------\n",
      "Epoch 560, Loss: 0.5164494175843025\n",
      "Accuracy: 48.03%\n",
      "Iteration 279:\n",
      "Training Loss: 0.5164494175843025, Training Accuracy: 48.26%\n",
      "Test Loss: 0.5303930685271154, Test Accuracy: 46.83%\n",
      "----------------------------------------\n",
      "Epoch 562, Loss: 0.5162062787621656\n",
      "Accuracy: 48.03%\n",
      "Iteration 280:\n",
      "Training Loss: 0.5162062787621656, Training Accuracy: 48.28%\n",
      "Test Loss: 0.5303998896502984, Test Accuracy: 46.83%\n",
      "----------------------------------------\n",
      "Epoch 564, Loss: 0.5163145937261255\n",
      "Accuracy: 48.03%\n",
      "Iteration 281:\n",
      "Training Loss: 0.5163145937261255, Training Accuracy: 48.27%\n",
      "Test Loss: 0.5305922041169834, Test Accuracy: 46.81%\n",
      "----------------------------------------\n",
      "Epoch 566, Loss: 0.5159560373072762\n",
      "Accuracy: 48.04%\n",
      "Iteration 282:\n",
      "Training Loss: 0.5159560373072762, Training Accuracy: 48.30%\n",
      "Test Loss: 0.5308088051808869, Test Accuracy: 46.79%\n",
      "----------------------------------------\n",
      "Epoch 568, Loss: 0.5156394822061582\n",
      "Accuracy: 48.05%\n",
      "Iteration 283:\n",
      "Training Loss: 0.5156394822061582, Training Accuracy: 48.33%\n",
      "Test Loss: 0.5308088051764016, Test Accuracy: 46.79%\n",
      "----------------------------------------\n",
      "Epoch 570, Loss: 0.515967898959421\n",
      "Accuracy: 48.05%\n",
      "Iteration 284:\n",
      "Training Loss: 0.515967898959421, Training Accuracy: 48.29%\n",
      "Test Loss: 0.530616497484094, Test Accuracy: 46.81%\n",
      "----------------------------------------\n",
      "Epoch 572, Loss: 0.5156962535697259\n",
      "Accuracy: 48.07%\n",
      "Iteration 285:\n",
      "Training Loss: 0.5156962535697259, Training Accuracy: 48.32%\n",
      "Test Loss: 0.5305040774129977, Test Accuracy: 46.81%\n",
      "----------------------------------------\n",
      "Epoch 574, Loss: 0.5156994319539107\n",
      "Accuracy: 48.07%\n",
      "Iteration 286:\n",
      "Training Loss: 0.5156994319539107, Training Accuracy: 48.32%\n",
      "Test Loss: 0.5311053006246493, Test Accuracy: 46.75%\n",
      "----------------------------------------\n",
      "Epoch 576, Loss: 0.5153628117777814\n",
      "Accuracy: 48.11%\n",
      "Iteration 287:\n",
      "Training Loss: 0.5153628117777814, Training Accuracy: 48.35%\n",
      "Test Loss: 0.5306325653171617, Test Accuracy: 46.79%\n",
      "----------------------------------------\n",
      "Epoch 578, Loss: 0.5148120398540129\n",
      "Accuracy: 48.12%\n",
      "Iteration 288:\n",
      "Training Loss: 0.5148120398540129, Training Accuracy: 48.40%\n",
      "Test Loss: 0.5308248730094921, Test Accuracy: 46.77%\n",
      "----------------------------------------\n",
      "Epoch 580, Loss: 0.5144755125645257\n",
      "Accuracy: 48.12%\n",
      "Iteration 289:\n",
      "Training Loss: 0.5144755125645257, Training Accuracy: 48.44%\n",
      "Test Loss: 0.5304402576248768, Test Accuracy: 46.81%\n",
      "----------------------------------------\n",
      "Epoch 582, Loss: 0.514987806775258\n",
      "Accuracy: 48.12%\n",
      "Iteration 290:\n",
      "Training Loss: 0.514987806775258, Training Accuracy: 48.38%\n",
      "Test Loss: 0.5299432221691651, Test Accuracy: 46.85%\n",
      "----------------------------------------\n",
      "Epoch 584, Loss: 0.5146234712041513\n",
      "Accuracy: 48.15%\n",
      "Iteration 291:\n",
      "Training Loss: 0.5146234712041513, Training Accuracy: 48.41%\n",
      "Test Loss: 0.5297509144768574, Test Accuracy: 46.87%\n",
      "----------------------------------------\n",
      "Epoch 586, Loss: 0.514221681515099\n",
      "Accuracy: 48.17%\n",
      "Iteration 292:\n",
      "Training Loss: 0.514221681515099, Training Accuracy: 48.45%\n",
      "Test Loss: 0.5295104430857009, Test Accuracy: 46.87%\n",
      "----------------------------------------\n",
      "Epoch 588, Loss: 0.5139344613220765\n",
      "Accuracy: 48.20%\n",
      "Iteration 293:\n",
      "Training Loss: 0.5139344613220765, Training Accuracy: 48.48%\n",
      "Test Loss: 0.5287811437067176, Test Accuracy: 46.94%\n",
      "----------------------------------------\n",
      "Epoch 590, Loss: 0.5136378775627091\n",
      "Accuracy: 48.18%\n",
      "Iteration 294:\n",
      "Training Loss: 0.5136378775627091, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5287811437069219, Test Accuracy: 46.94%\n",
      "----------------------------------------\n",
      "Epoch 592, Loss: 0.5136920404127899\n",
      "Accuracy: 48.18%\n",
      "Iteration 295:\n",
      "Training Loss: 0.5136920404127899, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5287811437069219, Test Accuracy: 46.94%\n",
      "----------------------------------------\n",
      "Epoch 594, Loss: 0.513494443498987\n",
      "Accuracy: 48.19%\n",
      "Iteration 296:\n",
      "Training Loss: 0.513494443498987, Training Accuracy: 48.51%\n",
      "Test Loss: 0.5285563035646474, Test Accuracy: 46.94%\n",
      "----------------------------------------\n",
      "Epoch 596, Loss: 0.5135024692377355\n",
      "Accuracy: 48.22%\n",
      "Iteration 297:\n",
      "Training Loss: 0.5135024692377355, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5282515758013246, Test Accuracy: 46.96%\n",
      "----------------------------------------\n",
      "Epoch 598, Loss: 0.513485003169252\n",
      "Accuracy: 48.22%\n",
      "Iteration 298:\n",
      "Training Loss: 0.513485003169252, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5282515758013254, Test Accuracy: 46.96%\n",
      "----------------------------------------\n",
      "Epoch 600, Loss: 0.5132251118805903\n",
      "Accuracy: 48.23%\n",
      "Iteration 299:\n",
      "Training Loss: 0.5132251118805903, Training Accuracy: 48.52%\n",
      "Test Loss: 0.5275066451735823, Test Accuracy: 47.04%\n",
      "----------------------------------------\n",
      "Epoch 602, Loss: 0.5132731994576722\n",
      "Accuracy: 48.23%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iteration 300:\n",
      "Training Loss: 0.5132731994576722, Training Accuracy: 48.52%\n",
      "Test Loss: 0.5268173020255629, Test Accuracy: 47.10%\n",
      "----------------------------------------\n",
      "Epoch 604, Loss: 0.5133473203834648\n",
      "Accuracy: 48.23%\n",
      "Iteration 301:\n",
      "Training Loss: 0.5133473203834648, Training Accuracy: 48.51%\n",
      "Test Loss: 0.5262403789486398, Test Accuracy: 47.15%\n",
      "----------------------------------------\n",
      "Epoch 606, Loss: 0.5131550368332051\n",
      "Accuracy: 48.20%\n",
      "Iteration 302:\n",
      "Training Loss: 0.5131550368332051, Training Accuracy: 48.53%\n",
      "Test Loss: 0.5262403789486398, Test Accuracy: 47.15%\n",
      "----------------------------------------\n",
      "Epoch 608, Loss: 0.5133872985985407\n",
      "Accuracy: 48.18%\n",
      "Iteration 303:\n",
      "Training Loss: 0.5133872985985407, Training Accuracy: 48.50%\n",
      "Test Loss: 0.5266249943332553, Test Accuracy: 47.12%\n",
      "----------------------------------------\n",
      "Epoch 610, Loss: 0.5132010617211699\n",
      "Accuracy: 48.20%\n",
      "Iteration 304:\n",
      "Training Loss: 0.5132010617211699, Training Accuracy: 48.52%\n",
      "Test Loss: 0.5263202665698512, Test Accuracy: 47.13%\n",
      "----------------------------------------\n",
      "Epoch 612, Loss: 0.5128845215789735\n",
      "Accuracy: 48.22%\n",
      "Iteration 305:\n",
      "Training Loss: 0.5128845215789735, Training Accuracy: 48.55%\n",
      "Test Loss: 0.5263202665698512, Test Accuracy: 47.13%\n",
      "----------------------------------------\n",
      "Epoch 614, Loss: 0.5126640949030605\n",
      "Accuracy: 48.27%\n",
      "Iteration 306:\n",
      "Training Loss: 0.5126640949030605, Training Accuracy: 48.57%\n",
      "Test Loss: 0.5275539679441601, Test Accuracy: 47.00%\n",
      "----------------------------------------\n",
      "Epoch 616, Loss: 0.51225020654742\n",
      "Accuracy: 48.27%\n",
      "Iteration 307:\n",
      "Training Loss: 0.51225020654742, Training Accuracy: 48.60%\n",
      "Test Loss: 0.5272167401315834, Test Accuracy: 47.00%\n",
      "----------------------------------------\n",
      "Epoch 618, Loss: 0.5122975209633666\n",
      "Accuracy: 48.28%\n",
      "Iteration 308:\n",
      "Training Loss: 0.5122975209633666, Training Accuracy: 48.60%\n",
      "Test Loss: 0.5268321247470046, Test Accuracy: 47.04%\n",
      "----------------------------------------\n",
      "Epoch 620, Loss: 0.5122554460781816\n",
      "Accuracy: 48.27%\n",
      "Iteration 309:\n",
      "Training Loss: 0.5122554460781816, Training Accuracy: 48.60%\n",
      "Test Loss: 0.5259504739066776, Test Accuracy: 47.12%\n",
      "----------------------------------------\n",
      "Epoch 622, Loss: 0.5121833438089749\n",
      "Accuracy: 48.29%\n",
      "Iteration 310:\n",
      "Training Loss: 0.5121833438089749, Training Accuracy: 48.60%\n",
      "Test Loss: 0.5261427815267679, Test Accuracy: 47.10%\n",
      "----------------------------------------\n",
      "Epoch 624, Loss: 0.5122093473689281\n",
      "Accuracy: 48.26%\n",
      "Iteration 311:\n",
      "Training Loss: 0.5122093473689281, Training Accuracy: 48.60%\n",
      "Test Loss: 0.5257581662143699, Test Accuracy: 47.13%\n",
      "----------------------------------------\n",
      "Epoch 626, Loss: 0.5121126667586764\n",
      "Accuracy: 48.27%\n",
      "Iteration 312:\n",
      "Training Loss: 0.5121126667586764, Training Accuracy: 48.61%\n",
      "Test Loss: 0.5253978509712423, Test Accuracy: 47.17%\n",
      "----------------------------------------\n",
      "Epoch 628, Loss: 0.5118707723152837\n",
      "Accuracy: 48.29%\n",
      "Iteration 313:\n",
      "Training Loss: 0.5118707723152837, Training Accuracy: 48.62%\n",
      "Test Loss: 0.5246529203434992, Test Accuracy: 47.25%\n",
      "----------------------------------------\n",
      "Epoch 630, Loss: 0.5115411880630193\n",
      "Accuracy: 48.30%\n",
      "Iteration 314:\n",
      "Training Loss: 0.5115411880630193, Training Accuracy: 48.66%\n",
      "Test Loss: 0.5248312717862059, Test Accuracy: 47.23%\n",
      "----------------------------------------\n",
      "Epoch 632, Loss: 0.5111197871400932\n",
      "Accuracy: 48.31%\n",
      "Iteration 315:\n",
      "Training Loss: 0.5111197871400932, Training Accuracy: 48.70%\n",
      "Test Loss: 0.5242683049628475, Test Accuracy: 47.29%\n",
      "----------------------------------------\n",
      "Epoch 634, Loss: 0.511007688002286\n",
      "Accuracy: 48.35%\n",
      "Iteration 316:\n",
      "Training Loss: 0.511007688002286, Training Accuracy: 48.71%\n",
      "Test Loss: 0.5238518625269779, Test Accuracy: 47.31%\n",
      "----------------------------------------\n",
      "Epoch 636, Loss: 0.5109795880517497\n",
      "Accuracy: 48.34%\n",
      "Iteration 317:\n",
      "Training Loss: 0.5109795880517497, Training Accuracy: 48.71%\n",
      "Test Loss: 0.5236588494320761, Test Accuracy: 47.33%\n",
      "----------------------------------------\n",
      "Epoch 638, Loss: 0.5106410054742527\n",
      "Accuracy: 48.34%\n",
      "Iteration 318:\n",
      "Training Loss: 0.5106410054742527, Training Accuracy: 48.74%\n",
      "Test Loss: 0.5235464710308553, Test Accuracy: 47.33%\n",
      "----------------------------------------\n",
      "Epoch 640, Loss: 0.5108216965454621\n",
      "Accuracy: 48.35%\n",
      "Iteration 319:\n",
      "Training Loss: 0.5108216965454621, Training Accuracy: 48.72%\n",
      "Test Loss: 0.523546429360903, Test Accuracy: 47.33%\n",
      "----------------------------------------\n",
      "Epoch 642, Loss: 0.5103625960665867\n",
      "Accuracy: 48.38%\n",
      "Iteration 320:\n",
      "Training Loss: 0.5103625960665867, Training Accuracy: 48.76%\n",
      "Test Loss: 0.5232417015975763, Test Accuracy: 47.35%\n",
      "----------------------------------------\n",
      "Epoch 644, Loss: 0.5100563123141375\n",
      "Accuracy: 48.42%\n",
      "Iteration 321:\n",
      "Training Loss: 0.5100563123141375, Training Accuracy: 48.79%\n",
      "Test Loss: 0.5232417015981802, Test Accuracy: 47.35%\n",
      "----------------------------------------\n",
      "Epoch 646, Loss: 0.5096736442426382\n",
      "Accuracy: 48.45%\n",
      "Iteration 322:\n",
      "Training Loss: 0.5096736442426382, Training Accuracy: 48.83%\n",
      "Test Loss: 0.5228570862129617, Test Accuracy: 47.38%\n",
      "----------------------------------------\n",
      "Epoch 648, Loss: 0.5091908266516872\n",
      "Accuracy: 48.50%\n",
      "Iteration 323:\n",
      "Training Loss: 0.5091908266516872, Training Accuracy: 48.88%\n",
      "Test Loss: 0.5224724708283468, Test Accuracy: 47.42%\n",
      "----------------------------------------\n",
      "Epoch 650, Loss: 0.5087306260757117\n",
      "Accuracy: 48.50%\n",
      "Iteration 324:\n",
      "Training Loss: 0.5087306260757117, Training Accuracy: 48.92%\n",
      "Test Loss: 0.5228536949369114, Test Accuracy: 47.38%\n",
      "----------------------------------------\n",
      "Epoch 652, Loss: 0.5083237145476756\n",
      "Accuracy: 48.50%\n",
      "Iteration 325:\n",
      "Training Loss: 0.5083237145476756, Training Accuracy: 48.96%\n",
      "Test Loss: 0.5223044632775645, Test Accuracy: 47.44%\n",
      "----------------------------------------\n",
      "Epoch 654, Loss: 0.5082170678606526\n",
      "Accuracy: 48.52%\n",
      "Iteration 326:\n",
      "Training Loss: 0.5082170678606526, Training Accuracy: 48.96%\n",
      "Test Loss: 0.5219198478929153, Test Accuracy: 47.48%\n",
      "----------------------------------------\n",
      "Epoch 656, Loss: 0.5084574619899422\n",
      "Accuracy: 48.51%\n",
      "Iteration 327:\n",
      "Training Loss: 0.5084574619899422, Training Accuracy: 48.93%\n",
      "Test Loss: 0.5219441480344059, Test Accuracy: 47.48%\n",
      "----------------------------------------\n",
      "Epoch 658, Loss: 0.508319094040392\n",
      "Accuracy: 48.52%\n",
      "Iteration 328:\n",
      "Training Loss: 0.508319094040392, Training Accuracy: 48.95%\n",
      "Test Loss: 0.5217518403421025, Test Accuracy: 47.50%\n",
      "----------------------------------------\n",
      "Epoch 660, Loss: 0.5081229253622858\n",
      "Accuracy: 48.56%\n",
      "Iteration 329:\n",
      "Training Loss: 0.5081229253622858, Training Accuracy: 48.97%\n",
      "Test Loss: 0.5207903018805704, Test Accuracy: 47.60%\n",
      "----------------------------------------\n",
      "Epoch 662, Loss: 0.5082834058001146\n",
      "Accuracy: 48.53%\n",
      "Iteration 330:\n",
      "Training Loss: 0.5082834058001146, Training Accuracy: 48.95%\n",
      "Test Loss: 0.5207903018805802, Test Accuracy: 47.60%\n",
      "----------------------------------------\n",
      "Epoch 664, Loss: 0.5078828468296339\n",
      "Accuracy: 48.54%\n",
      "Iteration 331:\n",
      "Training Loss: 0.5078828468296339, Training Accuracy: 48.99%\n",
      "Test Loss: 0.5206778818094989, Test Accuracy: 47.60%\n",
      "----------------------------------------\n",
      "Epoch 666, Loss: 0.5078245881565354\n",
      "Accuracy: 48.55%\n",
      "Iteration 332:\n",
      "Training Loss: 0.5078245881565354, Training Accuracy: 48.99%\n",
      "Test Loss: 0.5208701895018297, Test Accuracy: 47.58%\n",
      "----------------------------------------\n",
      "Epoch 668, Loss: 0.5075080309498582\n",
      "Accuracy: 48.60%\n",
      "Iteration 333:\n",
      "Training Loss: 0.5075080309498582, Training Accuracy: 49.01%\n",
      "Test Loss: 0.5204855741172498, Test Accuracy: 47.62%\n",
      "----------------------------------------\n",
      "Epoch 670, Loss: 0.5071074698604902\n",
      "Accuracy: 48.61%\n",
      "Iteration 334:\n",
      "Training Loss: 0.5071074698604902, Training Accuracy: 49.05%\n",
      "Test Loss: 0.5204855741173049, Test Accuracy: 47.62%\n",
      "----------------------------------------\n",
      "Epoch 672, Loss: 0.5067749574867204\n",
      "Accuracy: 48.60%\n",
      "Iteration 335:\n",
      "Training Loss: 0.5067749574867204, Training Accuracy: 49.08%\n",
      "Test Loss: 0.5195483359833218, Test Accuracy: 47.71%\n",
      "----------------------------------------\n",
      "Epoch 674, Loss: 0.5064643612585451\n",
      "Accuracy: 48.63%\n",
      "Iteration 336:\n",
      "Training Loss: 0.5064643612585451, Training Accuracy: 49.11%\n",
      "Test Loss: 0.5197892442291758, Test Accuracy: 47.69%\n",
      "----------------------------------------\n",
      "Epoch 676, Loss: 0.5063004021196053\n",
      "Accuracy: 48.62%\n",
      "Iteration 337:\n",
      "Training Loss: 0.5063004021196053, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5194289285298108, Test Accuracy: 47.73%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 678, Loss: 0.5061154988870303\n",
      "Accuracy: 48.64%\n",
      "Iteration 338:\n",
      "Training Loss: 0.5061154988870303, Training Accuracy: 49.13%\n",
      "Test Loss: 0.51885200545324, Test Accuracy: 47.79%\n",
      "----------------------------------------\n",
      "Epoch 680, Loss: 0.506167737810825\n",
      "Accuracy: 48.64%\n",
      "Iteration 339:\n",
      "Training Loss: 0.506167737810825, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5175301517490847, Test Accuracy: 47.92%\n",
      "----------------------------------------\n",
      "Epoch 682, Loss: 0.50600155797128\n",
      "Accuracy: 48.65%\n",
      "Iteration 340:\n",
      "Training Loss: 0.50600155797128, Training Accuracy: 49.13%\n",
      "Test Loss: 0.5176899269923103, Test Accuracy: 47.88%\n",
      "----------------------------------------\n",
      "Epoch 684, Loss: 0.5063269168777055\n",
      "Accuracy: 48.62%\n",
      "Iteration 341:\n",
      "Training Loss: 0.5063269168777055, Training Accuracy: 49.10%\n",
      "Test Loss: 0.5175219194427528, Test Accuracy: 47.90%\n",
      "----------------------------------------\n",
      "Epoch 686, Loss: 0.5064724657171157\n",
      "Accuracy: 48.59%\n",
      "Iteration 342:\n",
      "Training Loss: 0.5064724657171157, Training Accuracy: 49.09%\n",
      "Test Loss: 0.5173296117524316, Test Accuracy: 47.92%\n",
      "----------------------------------------\n",
      "Epoch 688, Loss: 0.5064924159781167\n",
      "Accuracy: 48.59%\n",
      "Iteration 343:\n",
      "Training Loss: 0.5064924159781167, Training Accuracy: 49.08%\n",
      "Test Loss: 0.5172171916842542, Test Accuracy: 47.92%\n",
      "----------------------------------------\n",
      "Epoch 690, Loss: 0.5063500579305776\n",
      "Accuracy: 48.62%\n",
      "Iteration 344:\n",
      "Training Loss: 0.5063500579305776, Training Accuracy: 49.10%\n",
      "Test Loss: 0.5172171916893986, Test Accuracy: 47.92%\n",
      "----------------------------------------\n",
      "Epoch 692, Loss: 0.5058735408552567\n",
      "Accuracy: 48.61%\n",
      "Iteration 345:\n",
      "Training Loss: 0.5058735408552567, Training Accuracy: 49.14%\n",
      "Test Loss: 0.5167931642784931, Test Accuracy: 47.96%\n",
      "----------------------------------------\n",
      "Epoch 694, Loss: 0.5061174558102534\n",
      "Accuracy: 48.59%\n",
      "Iteration 346:\n",
      "Training Loss: 0.5061174558102534, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5166402686325885, Test Accuracy: 47.98%\n",
      "----------------------------------------\n",
      "Epoch 696, Loss: 0.5058154198944605\n",
      "Accuracy: 48.59%\n",
      "Iteration 347:\n",
      "Training Loss: 0.5058154198944605, Training Accuracy: 49.14%\n",
      "Test Loss: 0.515950925504052, Test Accuracy: 48.04%\n",
      "----------------------------------------\n",
      "Epoch 698, Loss: 0.5060315425056439\n",
      "Accuracy: 48.60%\n",
      "Iteration 348:\n",
      "Training Loss: 0.5060315425056439, Training Accuracy: 49.11%\n",
      "Test Loss: 0.5155663101502457, Test Accuracy: 48.08%\n",
      "----------------------------------------\n",
      "Epoch 700, Loss: 0.5059415144505836\n",
      "Accuracy: 48.56%\n",
      "Iteration 349:\n",
      "Training Loss: 0.5059415144505836, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5155580778699527, Test Accuracy: 48.06%\n",
      "----------------------------------------\n",
      "Epoch 702, Loss: 0.505775274454322\n",
      "Accuracy: 48.56%\n",
      "Iteration 350:\n",
      "Training Loss: 0.505775274454322, Training Accuracy: 49.13%\n",
      "Test Loss: 0.5155580779678451, Test Accuracy: 48.06%\n",
      "----------------------------------------\n",
      "Epoch 704, Loss: 0.5055528126554866\n",
      "Accuracy: 48.59%\n",
      "Iteration 351:\n",
      "Training Loss: 0.5055528126554866, Training Accuracy: 49.15%\n",
      "Test Loss: 0.5147888473209757, Test Accuracy: 48.13%\n",
      "----------------------------------------\n",
      "Epoch 706, Loss: 0.5054016661114235\n",
      "Accuracy: 48.60%\n",
      "Iteration 352:\n",
      "Training Loss: 0.5054016661114235, Training Accuracy: 49.15%\n",
      "Test Loss: 0.5150610428286593, Test Accuracy: 48.10%\n",
      "----------------------------------------\n",
      "Epoch 708, Loss: 0.505296248207651\n",
      "Accuracy: 48.60%\n",
      "Iteration 353:\n",
      "Training Loss: 0.505296248207651, Training Accuracy: 49.16%\n",
      "Test Loss: 0.5147007278938059, Test Accuracy: 48.13%\n",
      "----------------------------------------\n",
      "Epoch 710, Loss: 0.5053097659304131\n",
      "Accuracy: 48.59%\n",
      "Iteration 354:\n",
      "Training Loss: 0.5053097659304131, Training Accuracy: 49.15%\n",
      "Test Loss: 0.5144758882410483, Test Accuracy: 48.13%\n",
      "----------------------------------------\n",
      "Epoch 712, Loss: 0.5055543941578371\n",
      "Accuracy: 48.59%\n",
      "Iteration 355:\n",
      "Training Loss: 0.5055543941578371, Training Accuracy: 49.12%\n",
      "Test Loss: 0.514078932877918, Test Accuracy: 48.15%\n",
      "----------------------------------------\n",
      "Epoch 714, Loss: 0.5057246837570959\n",
      "Accuracy: 48.56%\n",
      "Iteration 356:\n",
      "Training Loss: 0.5057246837570959, Training Accuracy: 49.10%\n",
      "Test Loss: 0.5141629300726237, Test Accuracy: 48.13%\n",
      "----------------------------------------\n",
      "Epoch 716, Loss: 0.5055042799862598\n",
      "Accuracy: 48.56%\n",
      "Iteration 357:\n",
      "Training Loss: 0.5055042799862598, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5139706330297127, Test Accuracy: 48.15%\n",
      "----------------------------------------\n",
      "Epoch 718, Loss: 0.5052198443175461\n",
      "Accuracy: 48.55%\n",
      "Iteration 358:\n",
      "Training Loss: 0.5052198443175461, Training Accuracy: 49.14%\n",
      "Test Loss: 0.5138026200241022, Test Accuracy: 48.17%\n",
      "----------------------------------------\n",
      "Epoch 720, Loss: 0.5053399959170188\n",
      "Accuracy: 48.55%\n",
      "Iteration 359:\n",
      "Training Loss: 0.5053399959170188, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5138264537545899, Test Accuracy: 48.17%\n",
      "----------------------------------------\n",
      "Epoch 722, Loss: 0.5054226816501747\n",
      "Accuracy: 48.53%\n",
      "Iteration 360:\n",
      "Training Loss: 0.5054226816501747, Training Accuracy: 49.11%\n",
      "Test Loss: 0.5134423106610065, Test Accuracy: 48.21%\n",
      "----------------------------------------\n",
      "Epoch 724, Loss: 0.5055823571096615\n",
      "Accuracy: 48.52%\n",
      "Iteration 361:\n",
      "Training Loss: 0.5055823571096615, Training Accuracy: 49.09%\n",
      "Test Loss: 0.5134423302088004, Test Accuracy: 48.21%\n",
      "----------------------------------------\n",
      "Epoch 726, Loss: 0.5053995817729058\n",
      "Accuracy: 48.50%\n",
      "Iteration 362:\n",
      "Training Loss: 0.5053995817729058, Training Accuracy: 49.11%\n",
      "Test Loss: 0.5130577331536004, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 728, Loss: 0.5051979715121001\n",
      "Accuracy: 48.49%\n",
      "Iteration 363:\n",
      "Training Loss: 0.5051979715121001, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5134666746453489, Test Accuracy: 48.21%\n",
      "----------------------------------------\n",
      "Epoch 730, Loss: 0.5051339484473395\n",
      "Accuracy: 48.52%\n",
      "Iteration 364:\n",
      "Training Loss: 0.5051339484473395, Training Accuracy: 49.12%\n",
      "Test Loss: 0.5132744030094633, Test Accuracy: 48.23%\n",
      "----------------------------------------\n",
      "Epoch 732, Loss: 0.5048776176814529\n",
      "Accuracy: 48.51%\n",
      "Iteration 365:\n",
      "Training Loss: 0.5048776176814529, Training Accuracy: 49.15%\n",
      "Test Loss: 0.513274451954403, Test Accuracy: 48.23%\n",
      "----------------------------------------\n",
      "Epoch 734, Loss: 0.5046754368189632\n",
      "Accuracy: 48.50%\n",
      "Iteration 366:\n",
      "Training Loss: 0.5046754368189632, Training Accuracy: 49.17%\n",
      "Test Loss: 0.5128261035500018, Test Accuracy: 48.27%\n",
      "----------------------------------------\n",
      "Epoch 736, Loss: 0.5045132241905963\n",
      "Accuracy: 48.50%\n",
      "Iteration 367:\n",
      "Training Loss: 0.5045132241905963, Training Accuracy: 49.18%\n",
      "Test Loss: 0.5129407950372218, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 738, Loss: 0.504132599437404\n",
      "Accuracy: 48.51%\n",
      "Iteration 368:\n",
      "Training Loss: 0.504132599437404, Training Accuracy: 49.21%\n",
      "Test Loss: 0.5128263371256968, Test Accuracy: 48.27%\n",
      "----------------------------------------\n",
      "Epoch 740, Loss: 0.504392979345396\n",
      "Accuracy: 48.50%\n",
      "Iteration 369:\n",
      "Training Loss: 0.504392979345396, Training Accuracy: 49.18%\n",
      "Test Loss: 0.5122737964530254, Test Accuracy: 48.33%\n",
      "----------------------------------------\n",
      "Epoch 742, Loss: 0.5041586789411574\n",
      "Accuracy: 48.51%\n",
      "Iteration 370:\n",
      "Training Loss: 0.5041586789411574, Training Accuracy: 49.21%\n",
      "Test Loss: 0.5125461876050084, Test Accuracy: 48.29%\n",
      "----------------------------------------\n",
      "Epoch 744, Loss: 0.5041024866123408\n",
      "Accuracy: 48.50%\n",
      "Iteration 371:\n",
      "Training Loss: 0.5041024866123408, Training Accuracy: 49.21%\n",
      "Test Loss: 0.512354159403297, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 746, Loss: 0.5036596021941184\n",
      "Accuracy: 48.50%\n",
      "Iteration 372:\n",
      "Training Loss: 0.5036596021941184, Training Accuracy: 49.24%\n",
      "Test Loss: 0.5128187071596685, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 748, Loss: 0.5039460142732513\n",
      "Accuracy: 48.50%\n",
      "Iteration 373:\n",
      "Training Loss: 0.5039460142732513, Training Accuracy: 49.21%\n",
      "Test Loss: 0.5128196242118208, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 750, Loss: 0.5038877547358008\n",
      "Accuracy: 48.49%\n",
      "Iteration 374:\n",
      "Training Loss: 0.5038877547358008, Training Accuracy: 49.20%\n",
      "Test Loss: 0.5123233964118301, Test Accuracy: 48.29%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 752, Loss: 0.5039618870783753\n",
      "Accuracy: 48.49%\n",
      "Iteration 375:\n",
      "Training Loss: 0.5039618870783753, Training Accuracy: 49.19%\n",
      "Test Loss: 0.5123245566885336, Test Accuracy: 48.29%\n",
      "----------------------------------------\n",
      "Epoch 754, Loss: 0.5036574791880382\n",
      "Accuracy: 48.48%\n",
      "Iteration 376:\n",
      "Training Loss: 0.5036574791880382, Training Accuracy: 49.22%\n",
      "Test Loss: 0.5122936806792358, Test Accuracy: 48.27%\n",
      "----------------------------------------\n",
      "Epoch 756, Loss: 0.5033738861061654\n",
      "Accuracy: 48.47%\n",
      "Iteration 377:\n",
      "Training Loss: 0.5033738861061654, Training Accuracy: 49.25%\n",
      "Test Loss: 0.5124558010508498, Test Accuracy: 48.23%\n",
      "----------------------------------------\n",
      "Epoch 758, Loss: 0.5030967403582856\n",
      "Accuracy: 48.51%\n",
      "Iteration 378:\n",
      "Training Loss: 0.5030967403582856, Training Accuracy: 49.27%\n",
      "Test Loss: 0.5126513829203363, Test Accuracy: 48.21%\n",
      "----------------------------------------\n",
      "Epoch 760, Loss: 0.502772887324586\n",
      "Accuracy: 48.52%\n",
      "Iteration 379:\n",
      "Training Loss: 0.502772887324586, Training Accuracy: 49.29%\n",
      "Test Loss: 0.5124878592587198, Test Accuracy: 48.23%\n",
      "----------------------------------------\n",
      "Epoch 762, Loss: 0.5027479725825883\n",
      "Accuracy: 48.53%\n",
      "Iteration 380:\n",
      "Training Loss: 0.5027479725825883, Training Accuracy: 49.29%\n",
      "Test Loss: 0.5121009929075414, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 764, Loss: 0.5025776005684719\n",
      "Accuracy: 48.53%\n",
      "Iteration 381:\n",
      "Training Loss: 0.5025776005684719, Training Accuracy: 49.30%\n",
      "Test Loss: 0.5118787531693513, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 766, Loss: 0.5023814962371932\n",
      "Accuracy: 48.50%\n",
      "Iteration 382:\n",
      "Training Loss: 0.5023814962371932, Training Accuracy: 49.32%\n",
      "Test Loss: 0.5116798485424983, Test Accuracy: 48.27%\n",
      "----------------------------------------\n",
      "Epoch 768, Loss: 0.502426968166256\n",
      "Accuracy: 48.51%\n",
      "Iteration 383:\n",
      "Training Loss: 0.502426968166256, Training Accuracy: 49.31%\n",
      "Test Loss: 0.5113817281139624, Test Accuracy: 48.29%\n",
      "----------------------------------------\n",
      "Epoch 770, Loss: 0.5022135273140514\n",
      "Accuracy: 48.50%\n",
      "Iteration 384:\n",
      "Training Loss: 0.5022135273140514, Training Accuracy: 49.33%\n",
      "Test Loss: 0.5111894281144477, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 772, Loss: 0.5022108972738334\n",
      "Accuracy: 48.55%\n",
      "Iteration 385:\n",
      "Training Loss: 0.5022108972738334, Training Accuracy: 49.32%\n",
      "Test Loss: 0.5116539410932123, Test Accuracy: 48.25%\n",
      "----------------------------------------\n",
      "Epoch 774, Loss: 0.5021610173523289\n",
      "Accuracy: 48.52%\n",
      "Iteration 386:\n",
      "Training Loss: 0.5021610173523289, Training Accuracy: 49.32%\n",
      "Test Loss: 0.510739770544785, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 776, Loss: 0.501960584763639\n",
      "Accuracy: 48.52%\n",
      "Iteration 387:\n",
      "Training Loss: 0.501960584763639, Training Accuracy: 49.34%\n",
      "Test Loss: 0.5107640885175724, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 778, Loss: 0.5016762691403849\n",
      "Accuracy: 48.50%\n",
      "Iteration 388:\n",
      "Training Loss: 0.5016762691403849, Training Accuracy: 49.36%\n",
      "Test Loss: 0.5105717987426951, Test Accuracy: 48.33%\n",
      "----------------------------------------\n",
      "Epoch 780, Loss: 0.5018100248220397\n",
      "Accuracy: 48.51%\n",
      "Iteration 389:\n",
      "Training Loss: 0.5018100248220397, Training Accuracy: 49.34%\n",
      "Test Loss: 0.5104594211246313, Test Accuracy: 48.33%\n",
      "----------------------------------------\n",
      "Epoch 782, Loss: 0.5016799960339215\n",
      "Accuracy: 48.52%\n",
      "Iteration 390:\n",
      "Training Loss: 0.5016799960339215, Training Accuracy: 49.36%\n",
      "Test Loss: 0.51065177886952, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 784, Loss: 0.5014087154494248\n",
      "Accuracy: 48.54%\n",
      "Iteration 391:\n",
      "Training Loss: 0.5014087154494248, Training Accuracy: 49.38%\n",
      "Test Loss: 0.5106518500088277, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 786, Loss: 0.501028378662506\n",
      "Accuracy: 48.56%\n",
      "Iteration 392:\n",
      "Training Loss: 0.501028378662506, Training Accuracy: 49.41%\n",
      "Test Loss: 0.5102673359667844, Test Accuracy: 48.35%\n",
      "----------------------------------------\n",
      "Epoch 788, Loss: 0.5008749489154737\n",
      "Accuracy: 48.57%\n",
      "Iteration 393:\n",
      "Training Loss: 0.5008749489154737, Training Accuracy: 49.42%\n",
      "Test Loss: 0.5106520855785651, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 790, Loss: 0.5008390300710277\n",
      "Accuracy: 48.57%\n",
      "Iteration 394:\n",
      "Training Loss: 0.5008390300710277, Training Accuracy: 49.43%\n",
      "Test Loss: 0.5105069411668058, Test Accuracy: 48.29%\n",
      "----------------------------------------\n",
      "Epoch 792, Loss: 0.5007604500474617\n",
      "Accuracy: 48.57%\n",
      "Iteration 395:\n",
      "Training Loss: 0.5007604500474617, Training Accuracy: 49.44%\n",
      "Test Loss: 0.5097626038684465, Test Accuracy: 48.37%\n",
      "----------------------------------------\n",
      "Epoch 794, Loss: 0.5006967721150284\n",
      "Accuracy: 48.55%\n",
      "Iteration 396:\n",
      "Training Loss: 0.5006967721150284, Training Accuracy: 49.43%\n",
      "Test Loss: 0.5098984296022901, Test Accuracy: 48.35%\n",
      "----------------------------------------\n",
      "Epoch 796, Loss: 0.5004627141124484\n",
      "Accuracy: 48.55%\n",
      "Iteration 397:\n",
      "Training Loss: 0.5004627141124484, Training Accuracy: 49.45%\n",
      "Test Loss: 0.5099473510974128, Test Accuracy: 48.33%\n",
      "----------------------------------------\n",
      "Epoch 798, Loss: 0.5004825232974583\n",
      "Accuracy: 48.56%\n",
      "Iteration 398:\n",
      "Training Loss: 0.5004825232974583, Training Accuracy: 49.44%\n",
      "Test Loss: 0.5099242623505917, Test Accuracy: 48.31%\n",
      "----------------------------------------\n",
      "Epoch 800, Loss: 0.5002402291610795\n",
      "Accuracy: 48.54%\n",
      "Iteration 399:\n",
      "Training Loss: 0.5002402291610795, Training Accuracy: 49.47%\n",
      "Test Loss: 0.509251036733711, Test Accuracy: 48.37%\n",
      "----------------------------------------\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "def sigmoid(x):\n",
    "    return 1 / (1 + np.exp(-x))\n",
    "\n",
    "def sigmoid_derivative(x):\n",
    "    return x * (1 - x)\n",
    "\n",
    "class AdamOptimizer:\n",
    "    def __init__(self,x,hidden1, out_layer ,learning_rate=0.001, beta1=0.9, beta2=0.999, epsilon=1e-8):\n",
    "        self.learning_rate = learning_rate\n",
    "        self.beta1 = beta1\n",
    "        self.beta2 = beta2\n",
    "        self.epsilon = epsilon\n",
    "        self.m_w1 = np.zeros([x.shape[1],hidden1])\n",
    "        self.v_w1 = np.zeros([x.shape[1],hidden1])\n",
    "        self.m_w2 = np.zeros([hidden1,out_layer])\n",
    "        self.v_w2 = np.zeros([hidden1,out_layer])\n",
    "        self.t = 0\n",
    "\n",
    "    def initialize(self, parameters):\n",
    "        self.m = np.zeros_like(parameters)\n",
    "        self.v = np.zeros_like(parameters)\n",
    "\n",
    "    def update(self, parameters, gradients):\n",
    "        self.t += 1\n",
    "        if(parameters.shape==self.m_w1.shape):\n",
    "            self.m_w1 = self.beta1 * self.m_w1 + (1 - self.beta1) * gradients\n",
    "            self.v_w1 = self.beta2 * self.v_w1 + (1 - self.beta2) * (gradients ** 2)\n",
    "\n",
    "            m_hat_w1 = self.m_w1 / (1 - self.beta1 ** self.t)\n",
    "            v_hat_w1 = self.v_w1 / (1 - self.beta2 ** self.t)\n",
    "            parameters += self.learning_rate * m_hat_w1 / (np.sqrt(v_hat_w1) + self.epsilon)\n",
    "        else:\n",
    "            self.m_w2 = self.beta1 * self.m_w2 + (1 - self.beta1) * gradients\n",
    "            self.v_w2 = self.beta2 * self.v_w2 + (1 - self.beta2) * (gradients ** 2)\n",
    "\n",
    "            m_hat_w2 = self.m_w2 / (1 - self.beta1 ** self.t)\n",
    "            v_hat_w2 = self.v_w2 / (1 - self.beta2 ** self.t)\n",
    "            parameters += self.learning_rate * m_hat_w2 / (np.sqrt(v_hat_w2) + self.epsilon)\n",
    "\n",
    "        \n",
    "\n",
    "        return parameters\n",
    "\n",
    "def init_weights(num_features, num_neuron):\n",
    "    weights = np.random.normal(0, 1, size=(num_features, num_neuron))\n",
    "    return weights\n",
    "\n",
    "def mean_squared_error(y_true, y_pred):\n",
    "    return np.mean((y_true - y_pred)**2)\n",
    "\n",
    "class BayesianNeuralNetwork:\n",
    "    def __init__(self, x, y, hidden1, out_layer, b):\n",
    "        self.input = x\n",
    "        self.weights1_prior = init_weights(self.input.shape[1], hidden1)\n",
    "        self.weights2_prior = init_weights(hidden1, out_layer)\n",
    "        self.y = y\n",
    "        self.output = np.zeros(self.y.shape)\n",
    "        self.loss = []\n",
    "        self.bias = b\n",
    "        self.optimizer = AdamOptimizer(x,hidden1,out_layer)\n",
    "\n",
    "    def feedforward(self):\n",
    "        self.layer1_out = relu(np.dot(self.input, self.weights1_prior) + self.bias[0]) \n",
    "        self.output = sigmoid(np.dot(self.layer1_out, self.weights2_prior) + self.bias[1])\n",
    "\n",
    "    def backprop(self):\n",
    "        gradients_w2 = np.dot(self.layer1_out.T, (self.output - self.y))\n",
    "        gradients_w1 = np.dot(self.input.T, np.dot((self.output - self.y), self.weights2_prior.T) * self.layer1_out * (1 - self.layer1_out))\n",
    "\n",
    "        # Update weights using Adam optimizer\n",
    "        self.weights2_prior = self.optimizer.update(self.weights2_prior, gradients_w2)\n",
    "        self.weights1_prior = self.optimizer.update(self.weights1_prior, gradients_w1)\n",
    "\n",
    "        # Calculate and print the loss\n",
    "        cost = mean_squared_error(self.y, self.output)\n",
    "        self.loss.append(cost)\n",
    "        print(f\"Epoch {self.optimizer.t}, Loss: {cost}\")\n",
    "\n",
    "# Example usage\n",
    "# Assuming you have X_train, Y_train, initial_weights, and other necessary variables\n",
    "# ...\n",
    "\n",
    "# Create an instance of BayesianNeuralNetwork\n",
    "bias = np.random.normal(0.6,0.05,size=(2,1))\n",
    "model = BayesianNeuralNetwork(X_train, y_train, 100, 1, bias)\n",
    "model.loss = []\n",
    "iterations = []\n",
    "train_losses = []\n",
    "test_losses = []\n",
    "train_accuracies = []\n",
    "test_accuracies = []\n",
    "# Training loop\n",
    "for epoch in range(400):\n",
    "    # Forward and backward pass\n",
    "    model.feedforward()\n",
    "    model.backprop()\n",
    "    iterations.append(epoch)\n",
    "    accuracy = np.mean(model.output == y_train)*100\n",
    "    print(f\"Accuracy: {accuracy:.2f}%\")\n",
    "    #print(model.output)\n",
    "     # Training set evaluation\n",
    "    train_loss = model.loss[-1]\n",
    "    train_losses.append(train_loss)\n",
    "    train_accuracy = np.mean((model.output >= 0.5) == y_train) * 100\n",
    "    train_accuracies.append(train_accuracy)\n",
    "    # Test set evaluation\n",
    "    model.input = X_test\n",
    "    model.y = y_test\n",
    "    model.feedforward()\n",
    "    test_loss = mean_squared_error(model.y , model.output)\n",
    "    test_losses.append(test_loss)\n",
    "    test_accuracy = np.mean((model.output >= 0.5) == y_test) * 100\n",
    "    test_accuracies.append(test_accuracy)\n",
    "    # Reset input and y for the training set\n",
    "    model.input = X_train\n",
    "    model.y = y_train\n",
    "    print(f'Iteration {epoch}:')\n",
    "    print(f'Training Loss: {train_loss}, Training Accuracy: {train_accuracy:.2f}%')\n",
    "    print(f'Test Loss: {test_loss}, Test Accuracy: {test_accuracy:.2f}%')\n",
    "    print('-' * 40)\n",
    "\n",
    "plt.figure(figsize=(12, 5))\n",
    "\n",
    "# Plot training and test losses\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(iterations, train_losses, label='Training Loss')\n",
    "plt.plot(iterations, test_losses, label='Test Loss')\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('Loss')\n",
    "plt.title('Training and Test Losses over Iterations')\n",
    "plt.legend()\n",
    "\n",
    "# Plot training and test accuracies\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(iterations, train_accuracies, label='Training Accuracy')\n",
    "plt.plot(iterations, test_accuracies, label='Test Accuracy')\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('Accuracy (%)')\n",
    "plt.title('Training and Test Accuracies over Iterations')\n",
    "plt.legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()\n",
    "plt.plot(iterations, model.loss, label='loss value vs iteration')\n",
    "plt.xlabel('Iterations')\n",
    "plt.ylabel('loss function')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ba3a6ca2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2af7642c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}