VGG16_transfer3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#load important liberars\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.applications.vgg16 import VGG16\n",
    "import numpy as np\n",
    "from tensorflow.keras.layers import Flatten, Dense\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Input\n",
    "from tensorflow.keras.utils import plot_model\n",
    "from keras.optimizers import adam\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from sklearn.metrics import accuracy_score\n",
    "import glob\n",
    "from PIL import Image\n",
    "import numpy as np\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint\n",
    "import cv2\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import np_utils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the train and test datsets\n",
    "x_train=np.load('x_train.npy')\n",
    "x_test= np.load('x_test.npy')\n",
    "y_train=np.load('y_train_231.npy')\n",
    "y_test=np.load('y_test_231.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]],\n",
       "\n",
       "       [[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]],\n",
       "\n",
       "       [[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]],\n",
       "\n",
       "       ...,\n",
       "\n",
       "       [[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]],\n",
       "\n",
       "       [[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]],\n",
       "\n",
       "       [[255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        ...,\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.],\n",
       "        [255., 255., 255.]]], dtype=float32)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test[1]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from numpy import where\n",
    "from matplotlib import pyplot\n",
    "from sklearn.datasets import make_blobs\n",
    "# generate dataset\n",
    "X, y = make_blobs(n_samples=8000, centers=2, n_features=8000, random_state=1, cluster_std=3)\n",
    "# create scatter plot for samples from each class\n",
    "for class_value in range(231):\n",
    "\t# get row indexes for samples with this class\n",
    "\trow_ix = where(y == class_value)\n",
    "\t# create scatter of these samples\n",
    "\tpyplot.scatter(X[row_ix, 0], X[row_ix, 1])\n",
    "# show the plot\n",
    "pyplot.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "#load lables of the training and test datset interms of row/col\n",
    "y_trainr=np.load('y_trainr.npy')\n",
    "y_trainc=np.load('y_trainc.npy')\n",
    "y_testr=np.load('y_testr.npy')\n",
    "y_testc=np.load('y_testc.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# one-hoet-encoding of 231 classes\n",
    "Y_train =np_utils.to_categorical(y_train, 231)\n",
    "Y_test = np_utils.to_categorical(y_test, 231)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "#just one-hot-encoding (row/col--33/7 class)\n",
    "Y_trainr =np_utils.to_categorical(y_trainr, 33)\n",
    "Y_testr = np_utils.to_categorical(y_testr, 33)\n",
    "\n",
    "Y_trainc =np_utils.to_categorical(y_trainc, 7)\n",
    "Y_testc = np_utils.to_categorical(y_testc, 7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-14 15:01:29.011246: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "#since we set the include_top  as \"False\" the fully-connected output layers of the model is not loaded\n",
    "inputshape=Input(shape=(32,32,3))\n",
    "model=VGG16(weights=\"imagenet\",include_top=False,input_tensor=inputshape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to freezing the layers of the orginal as per our need... we just set\n",
    "# the tune_layer value for the model layers that should not be freez--->for 231 class\n",
    "\n",
    "def fine_tune(tune_layer=0):\n",
    "    if tune_layer>0:\n",
    "        for layer in model.layers[:-tune_layer]:\n",
    "            layer.trainable=False\n",
    "    else:\n",
    "        for layer in model.layers:\n",
    "            layer.trainable = False\n",
    "            \n",
    "            \n",
    "    last_conv_out= model.output\n",
    "    flatten= Flatten()(last_conv_out)\n",
    "    fc1= Dense(1024, activation='relu')(flatten)\n",
    "    fc2= Dense(2048, activation='relu')(fc1)\n",
    "    outputs_231=Dense(231, activation='softmax')(fc2)\n",
    "    \n",
    "    model_231=Model(inputs=model.inputs, outputs=outputs_231)\n",
    "    \n",
    "    model_231.compile(optimizer='adam',loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "    \n",
    "    return model_231"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans_model_231=fine_tune(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         [(None, 32, 32, 3)]       0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 32, 32, 64)        1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 32, 32, 64)        36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 16, 16, 64)        0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 16, 16, 128)       73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 16, 16, 128)       147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 8, 8, 128)         0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 8, 8, 256)         295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 8, 8, 256)         590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 4, 4, 256)         0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 4, 4, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 4, 4, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 2, 2, 512)         0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 2, 2, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1024)              525312    \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 2048)              2099200   \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 231)               473319    \n",
      "=================================================================\n",
      "Total params: 17,812,519\n",
      "Trainable params: 7,817,447\n",
      "Non-trainable params: 9,995,072\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "trans_model_231.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n"
     ]
    }
   ],
   "source": [
    "# Early Stopping to prevent overfitting and checkpoint to save model at each epoch231\n",
    "early_stopper = tf.keras.callbacks.EarlyStopping(monitor=\"val_loss\", \n",
    "                                                 patience=5, verbose=2, \n",
    "                                                 restore_best_weights=True)\n",
    "checkpoint = ModelCheckpoint('./vgg/vgg231_fin4{epoch:01d}.hdf5', period=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-14 15:01:47.784224: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:176] None of the MLIR Optimization Passes are enabled (registered 2)\n",
      "2022-05-14 15:01:47.802096: I tensorflow/core/platform/profile_utils/cpu_utils.cc:114] CPU Frequency: 3199980000 Hz\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2025/2025 [==============================] - 339s 167ms/step - loss: 5.5291 - accuracy: 0.0042 - val_loss: 5.4431 - val_accuracy: 0.0046\n",
      "Epoch 2/10\n",
      "2025/2025 [==============================] - 334s 165ms/step - loss: 5.4419 - accuracy: 0.0050 - val_loss: 5.4438 - val_accuracy: 0.0046\n",
      "Epoch 3/10\n",
      "2025/2025 [==============================] - 346s 171ms/step - loss: 5.4417 - accuracy: 0.0050 - val_loss: 5.4441 - val_accuracy: 0.0046\n",
      "Epoch 4/10\n",
      "2025/2025 [==============================] - 335s 165ms/step - loss: 5.4417 - accuracy: 0.0052 - val_loss: 5.4442 - val_accuracy: 0.0046\n",
      "Epoch 5/10\n",
      "2025/2025 [==============================] - 347s 172ms/step - loss: 5.4416 - accuracy: 0.0046 - val_loss: 5.4444 - val_accuracy: 0.0046\n",
      "Epoch 6/10\n",
      "2025/2025 [==============================] - 535s 264ms/step - loss: 5.4417 - accuracy: 0.0050 - val_loss: 5.4443 - val_accuracy: 0.0046\n",
      "Epoch 7/10\n",
      "2025/2025 [==============================] - 739s 365ms/step - loss: 5.4416 - accuracy: 0.0050 - val_loss: 5.4444 - val_accuracy: 0.0047\n",
      "Epoch 8/10\n",
      "2025/2025 [==============================] - 746s 369ms/step - loss: 5.4417 - accuracy: 0.0049 - val_loss: 5.4445 - val_accuracy: 0.0047\n",
      "Epoch 9/10\n",
      "2025/2025 [==============================] - 729s 360ms/step - loss: 5.4416 - accuracy: 0.0051 - val_loss: 5.4446 - val_accuracy: 0.0047\n",
      "Epoch 10/10\n",
      "2025/2025 [==============================] - 849s 419ms/step - loss: 5.4416 - accuracy: 0.0048 - val_loss: 5.4445 - val_accuracy: 0.0046\n"
     ]
    }
   ],
   "source": [
    "hist = trans_model_231.fit(x_train, Y_train, batch_size=32, epochs=10,\n",
    "                  verbose=1, validation_split=0.1,shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "500/500 [==============================] - 17s 34ms/step - loss: 1.0972 - accuracy: 0.8126\n"
     ]
    }
   ],
   "source": [
    "score = trans_model_231.evaluate(x_test,Y_test, batch_size=16, verbose=1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# visualizing  accuracy per  train1\n",
    "train_acc = hist.history['accuracy']\n",
    "val_acc = hist.history['val_accuracy']\n",
    "epochs = range(10)\n",
    "\n",
    "# to print training  accuracy vs validation accuracy\n",
    "plt.plot(epochs, train_acc)\n",
    "plt.plot(epochs, val_acc)\n",
    "plt.xlabel('num of Epochs')\n",
    "plt.ylabel('accuracy')\n",
    "plt.title('train_acc vs val_acc (3 layers Fine-tuned)')\n",
    "plt.grid(True)\n",
    "plt.legend(['train', 'val'], loc=4)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.89      0.93        38\n",
      "           1       1.00      0.89      0.94        36\n",
      "           2       0.73      0.57      0.64        28\n",
      "           3       0.84      0.81      0.82        26\n",
      "           4       0.91      0.79      0.85        38\n",
      "           5       0.86      0.97      0.91        32\n",
      "           6       0.89      0.87      0.88        38\n",
      "           7       0.89      0.69      0.77        35\n",
      "           8       0.83      0.74      0.78        27\n",
      "           9       0.76      0.85      0.80        40\n",
      "          10       0.81      0.93      0.87        42\n",
      "          11       0.58      0.76      0.66        25\n",
      "          12       0.89      0.97      0.93        34\n",
      "          13       0.96      0.87      0.92        31\n",
      "          14       0.84      0.89      0.86        36\n",
      "          15       0.90      0.90      0.90        31\n",
      "          16       0.89      0.78      0.83        32\n",
      "          17       0.65      0.96      0.77        25\n",
      "          18       0.83      0.92      0.88        38\n",
      "          19       0.94      0.94      0.94        35\n",
      "          20       0.91      0.85      0.88        34\n",
      "          21       0.84      0.91      0.87        23\n",
      "          22       1.00      0.89      0.94        38\n",
      "          23       0.83      0.89      0.86        38\n",
      "          24       0.95      0.82      0.88        22\n",
      "          25       0.80      0.82      0.81        34\n",
      "          26       0.97      0.97      0.97        33\n",
      "          27       0.86      0.93      0.89        27\n",
      "          28       0.97      0.91      0.94        34\n",
      "          29       0.83      0.94      0.88        31\n",
      "          30       0.69      0.89      0.78        28\n",
      "          31       0.90      0.92      0.91        39\n",
      "          32       0.93      0.68      0.79        38\n",
      "          33       0.97      0.94      0.96        36\n",
      "          34       0.74      0.96      0.84        24\n",
      "          35       0.94      0.94      0.94        31\n",
      "          36       0.95      0.90      0.92        40\n",
      "          37       0.75      0.84      0.79        32\n",
      "          38       0.94      0.62      0.75        24\n",
      "          39       0.93      0.90      0.92        31\n",
      "          40       0.86      0.86      0.86        28\n",
      "          41       0.73      0.77      0.75        31\n",
      "          42       0.79      0.74      0.76        50\n",
      "          43       0.74      0.85      0.79        27\n",
      "          44       0.82      0.68      0.75        41\n",
      "          45       0.86      0.88      0.87        34\n",
      "          46       0.60      0.80      0.69        41\n",
      "          47       0.83      0.71      0.77        35\n",
      "          48       0.81      0.89      0.85        28\n",
      "          49       0.64      0.67      0.65        24\n",
      "          50       0.89      0.82      0.85        39\n",
      "          51       0.83      0.78      0.81        37\n",
      "          52       0.78      0.76      0.77        41\n",
      "          53       0.76      0.83      0.79        41\n",
      "          54       0.82      0.74      0.78        31\n",
      "          55       0.79      0.87      0.82        38\n",
      "          56       0.90      0.97      0.94        38\n",
      "          57       0.84      0.73      0.78        37\n",
      "          58       0.69      0.76      0.72        29\n",
      "          59       1.00      0.86      0.93        44\n",
      "          60       0.60      0.64      0.62        33\n",
      "          61       0.92      0.73      0.81        33\n",
      "          62       0.97      0.87      0.92        45\n",
      "          63       0.56      0.79      0.66        29\n",
      "          64       0.88      0.83      0.86        36\n",
      "          65       0.78      0.75      0.77        24\n",
      "          66       0.74      0.78      0.76        32\n",
      "          67       0.91      0.46      0.61        46\n",
      "          68       0.81      0.87      0.84        30\n",
      "          69       0.79      0.70      0.74        33\n",
      "          70       0.91      0.76      0.83        41\n",
      "          71       0.62      0.96      0.75        25\n",
      "          72       0.87      0.87      0.87        47\n",
      "          73       0.86      0.90      0.88        41\n",
      "          74       0.71      0.67      0.69        33\n",
      "          75       0.81      0.91      0.85        32\n",
      "          76       0.97      0.85      0.91        40\n",
      "          77       0.71      0.76      0.74        46\n",
      "          78       0.70      0.34      0.46        41\n",
      "          79       0.93      0.36      0.52        39\n",
      "          80       0.89      0.86      0.88        29\n",
      "          81       0.37      0.80      0.51        30\n",
      "          82       0.68      1.00      0.81        38\n",
      "          83       0.96      0.93      0.95        29\n",
      "          84       0.85      0.73      0.79        30\n",
      "          85       0.75      0.87      0.80        38\n",
      "          86       0.66      0.81      0.72        26\n",
      "          87       0.95      0.93      0.94        44\n",
      "          88       0.71      0.84      0.77        32\n",
      "          89       0.97      0.86      0.91        42\n",
      "          90       0.88      0.88      0.88        41\n",
      "          91       0.96      0.73      0.83        33\n",
      "          92       0.95      0.64      0.76        33\n",
      "          93       0.81      0.75      0.78        28\n",
      "          94       0.90      0.73      0.81        37\n",
      "          95       0.79      0.74      0.77        31\n",
      "          96       0.79      0.96      0.87        24\n",
      "          97       0.75      0.92      0.83        36\n",
      "          98       0.97      0.83      0.90        36\n",
      "          99       0.89      0.86      0.88        37\n",
      "         100       0.84      0.91      0.88        35\n",
      "         101       0.76      0.97      0.85        29\n",
      "         102       0.66      0.79      0.72        29\n",
      "         103       0.83      0.97      0.89        39\n",
      "         104       1.00      0.89      0.94        38\n",
      "         105       0.89      0.69      0.78        45\n",
      "         106       0.84      0.71      0.77        38\n",
      "         107       0.91      0.83      0.87        35\n",
      "         108       0.88      0.69      0.77        32\n",
      "         109       0.65      0.91      0.76        35\n",
      "         110       0.64      0.97      0.77        29\n",
      "         111       0.76      0.97      0.85        39\n",
      "         112       0.68      0.62      0.65        37\n",
      "         113       0.78      0.78      0.78        37\n",
      "         114       0.95      0.93      0.94        42\n",
      "         115       0.75      0.87      0.80        38\n",
      "         116       0.70      0.64      0.67        36\n",
      "         117       0.75      0.67      0.71        27\n",
      "         118       0.57      0.64      0.60        39\n",
      "         119       0.65      0.67      0.66        30\n",
      "         120       0.83      0.76      0.79        33\n",
      "         121       0.72      0.79      0.75        29\n",
      "         122       0.74      0.78      0.76        36\n",
      "         123       0.78      0.89      0.83        36\n",
      "         124       0.79      0.68      0.73        34\n",
      "         125       0.82      0.71      0.76        38\n",
      "         126       0.85      1.00      0.92        22\n",
      "         127       0.90      0.93      0.92        41\n",
      "         128       0.91      0.72      0.81        40\n",
      "         129       0.97      0.97      0.97        34\n",
      "         130       0.81      0.78      0.79        37\n",
      "         131       0.84      0.84      0.84        37\n",
      "         132       0.94      0.87      0.90        53\n",
      "         133       0.84      0.65      0.73        48\n",
      "         134       0.88      0.77      0.82        48\n",
      "         135       0.84      0.76      0.80        34\n",
      "         136       0.77      0.80      0.79        30\n",
      "         137       0.78      0.85      0.81        41\n",
      "         138       0.75      0.69      0.72        26\n",
      "         139       0.67      0.67      0.67        30\n",
      "         140       0.70      0.70      0.70        30\n",
      "         141       0.79      0.82      0.80        40\n",
      "         142       0.81      0.74      0.77        34\n",
      "         143       0.68      0.72      0.70        32\n",
      "         144       0.79      0.79      0.79        34\n",
      "         145       0.96      0.69      0.80        35\n",
      "         146       0.48      0.92      0.63        24\n",
      "         147       0.82      0.87      0.85        38\n",
      "         148       0.86      0.69      0.77        26\n",
      "         149       0.88      0.82      0.85        28\n",
      "         150       0.78      0.90      0.84        31\n",
      "         151       0.69      0.85      0.76        34\n",
      "         152       0.88      0.76      0.81        29\n",
      "         153       0.75      0.78      0.76        27\n",
      "         154       0.92      0.80      0.86        30\n",
      "         155       0.93      0.86      0.89        29\n",
      "         156       0.83      0.67      0.74        43\n",
      "         157       0.94      0.89      0.92        37\n",
      "         158       0.71      0.71      0.71        42\n",
      "         159       0.91      0.88      0.89        33\n",
      "         160       1.00      0.80      0.89        44\n",
      "         161       0.89      0.84      0.86        37\n",
      "         162       0.83      0.88      0.85        43\n",
      "         163       0.89      0.91      0.90        44\n",
      "         164       0.89      0.81      0.85        42\n",
      "         165       0.91      0.97      0.94        32\n",
      "         166       0.68      0.80      0.74        35\n",
      "         167       0.79      0.79      0.79        39\n",
      "         168       0.81      0.86      0.83        29\n",
      "         169       0.86      0.97      0.91        32\n",
      "         170       0.93      0.81      0.86        31\n",
      "         171       0.91      0.86      0.88        35\n",
      "         172       0.86      0.86      0.86        35\n",
      "         173       0.69      0.85      0.76        26\n",
      "         174       0.94      0.82      0.88        39\n",
      "         175       0.86      0.93      0.89        40\n",
      "         176       0.93      0.68      0.78        37\n",
      "         177       0.79      0.81      0.80        32\n",
      "         178       0.97      0.94      0.96        35\n",
      "         179       0.76      0.82      0.79        34\n",
      "         180       0.96      0.85      0.90        26\n",
      "         181       0.88      0.72      0.79        40\n",
      "         182       0.89      0.80      0.85        41\n",
      "         183       0.82      0.93      0.87        30\n",
      "         184       0.65      0.91      0.75        22\n",
      "         185       0.86      0.92      0.89        39\n",
      "         186       0.97      0.77      0.86        43\n",
      "         187       0.96      0.93      0.95        29\n",
      "         188       0.82      0.95      0.88        38\n",
      "         189       0.85      0.88      0.86        25\n",
      "         190       0.69      0.81      0.75        27\n",
      "         191       0.98      0.87      0.92        53\n",
      "         192       0.81      0.65      0.72        26\n",
      "         193       0.95      0.83      0.89        24\n",
      "         194       0.73      0.95      0.82        37\n",
      "         195       0.88      0.90      0.89        40\n",
      "         196       0.78      0.73      0.75        48\n",
      "         197       0.86      0.86      0.86        44\n",
      "         198       0.79      0.87      0.83        31\n",
      "         199       0.65      0.79      0.71        14\n",
      "         200       0.87      0.70      0.78        37\n",
      "         201       1.00      0.85      0.92        40\n",
      "         202       0.84      0.67      0.74        39\n",
      "         203       0.35      0.78      0.49        36\n",
      "         204       0.53      0.84      0.65        32\n",
      "         205       0.85      0.73      0.79        56\n",
      "         206       0.94      0.64      0.76        25\n",
      "         207       0.93      0.50      0.65        28\n",
      "         208       0.90      0.88      0.89        32\n",
      "         209       0.84      0.79      0.81        33\n",
      "         210       0.71      0.89      0.79        46\n",
      "         211       0.90      0.79      0.84        33\n",
      "         212       0.89      0.67      0.77        46\n",
      "         213       0.91      0.77      0.83        26\n",
      "         214       0.73      0.65      0.69        34\n",
      "         215       0.71      0.74      0.73        27\n",
      "         216       0.77      0.62      0.69        37\n",
      "         217       0.76      0.86      0.81        44\n",
      "         218       0.82      0.82      0.82        33\n",
      "         219       0.83      0.88      0.85        40\n",
      "         220       1.00      0.82      0.90        28\n",
      "         221       0.83      0.89      0.86        27\n",
      "         222       0.85      0.87      0.86        39\n",
      "         223       0.88      0.91      0.89        32\n",
      "         224       0.86      0.89      0.87        35\n",
      "         225       0.89      0.74      0.81        34\n",
      "         226       0.83      0.88      0.85        40\n",
      "         227       0.87      0.97      0.92        40\n",
      "         228       0.77      0.88      0.82        34\n",
      "         229       0.85      0.85      0.85        27\n",
      "         230       0.82      0.84      0.83        32\n",
      "\n",
      "    accuracy                           0.81      8000\n",
      "   macro avg       0.83      0.82      0.81      8000\n",
      "weighted avg       0.83      0.81      0.82      8000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# model_0 fine tune\n",
    "from sklearn.metrics import classification_report\n",
    "from sklearn.metrics import accuracy_score\n",
    "from sklearn.metrics import confusion_matrix\n",
    "print(classification_report(y_true, y_pred))\n",
    "matrix = confusion_matrix(y_true, y_pred)\n",
    "acceach0=matrix.diagonal()/matrix.sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 0.8947368421052632\n",
      "1 0.8888888888888888\n",
      "2 0.5714285714285714\n",
      "3 0.8076923076923077\n",
      "4 0.7894736842105263\n",
      "5 0.96875\n",
      "6 0.868421052631579\n",
      "7 0.6857142857142857\n",
      "8 0.7407407407407407\n",
      "9 0.85\n",
      "10 0.9285714285714286\n",
      "11 0.76\n",
      "12 0.9705882352941176\n",
      "13 0.8709677419354839\n",
      "14 0.8888888888888888\n",
      "15 0.9032258064516129\n",
      "16 0.78125\n",
      "17 0.96\n",
      "18 0.9210526315789473\n",
      "19 0.9428571428571428\n",
      "20 0.8529411764705882\n",
      "21 0.9130434782608695\n",
      "22 0.8947368421052632\n",
      "23 0.8947368421052632\n",
      "24 0.8181818181818182\n",
      "25 0.8235294117647058\n",
      "26 0.9696969696969697\n",
      "27 0.9259259259259259\n",
      "28 0.9117647058823529\n",
      "29 0.9354838709677419\n",
      "30 0.8928571428571429\n",
      "31 0.9230769230769231\n",
      "32 0.6842105263157895\n",
      "33 0.9444444444444444\n",
      "34 0.9583333333333334\n",
      "35 0.9354838709677419\n",
      "36 0.9\n",
      "37 0.84375\n",
      "38 0.625\n",
      "39 0.9032258064516129\n",
      "40 0.8571428571428571\n",
      "41 0.7741935483870968\n",
      "42 0.74\n",
      "43 0.8518518518518519\n",
      "44 0.6829268292682927\n",
      "45 0.8823529411764706\n",
      "46 0.8048780487804879\n",
      "47 0.7142857142857143\n",
      "48 0.8928571428571429\n",
      "49 0.6666666666666666\n",
      "50 0.8205128205128205\n",
      "51 0.7837837837837838\n",
      "52 0.7560975609756098\n",
      "53 0.8292682926829268\n",
      "54 0.7419354838709677\n",
      "55 0.868421052631579\n",
      "56 0.9736842105263158\n",
      "57 0.7297297297297297\n",
      "58 0.7586206896551724\n",
      "59 0.8636363636363636\n",
      "60 0.6363636363636364\n",
      "61 0.7272727272727273\n",
      "62 0.8666666666666667\n",
      "63 0.7931034482758621\n",
      "64 0.8333333333333334\n",
      "65 0.75\n",
      "66 0.78125\n",
      "67 0.45652173913043476\n",
      "68 0.8666666666666667\n",
      "69 0.696969696969697\n",
      "70 0.7560975609756098\n",
      "71 0.96\n",
      "72 0.8723404255319149\n",
      "73 0.9024390243902439\n",
      "74 0.6666666666666666\n",
      "75 0.90625\n",
      "76 0.85\n",
      "77 0.7608695652173914\n",
      "78 0.34146341463414637\n",
      "79 0.358974358974359\n",
      "80 0.8620689655172413\n",
      "81 0.8\n",
      "82 1.0\n",
      "83 0.9310344827586207\n",
      "84 0.7333333333333333\n",
      "85 0.868421052631579\n",
      "86 0.8076923076923077\n",
      "87 0.9318181818181818\n",
      "88 0.84375\n",
      "89 0.8571428571428571\n",
      "90 0.8780487804878049\n",
      "91 0.7272727272727273\n",
      "92 0.6363636363636364\n",
      "93 0.75\n",
      "94 0.7297297297297297\n",
      "95 0.7419354838709677\n",
      "96 0.9583333333333334\n",
      "97 0.9166666666666666\n",
      "98 0.8333333333333334\n",
      "99 0.8648648648648649\n",
      "100 0.9142857142857143\n",
      "101 0.9655172413793104\n",
      "102 0.7931034482758621\n",
      "103 0.9743589743589743\n",
      "104 0.8947368421052632\n",
      "105 0.6888888888888889\n",
      "106 0.7105263157894737\n",
      "107 0.8285714285714286\n",
      "108 0.6875\n",
      "109 0.9142857142857143\n",
      "110 0.9655172413793104\n",
      "111 0.9743589743589743\n",
      "112 0.6216216216216216\n",
      "113 0.7837837837837838\n",
      "114 0.9285714285714286\n",
      "115 0.868421052631579\n",
      "116 0.6388888888888888\n",
      "117 0.6666666666666666\n",
      "118 0.6410256410256411\n",
      "119 0.6666666666666666\n",
      "120 0.7575757575757576\n",
      "121 0.7931034482758621\n",
      "122 0.7777777777777778\n",
      "123 0.8888888888888888\n",
      "124 0.6764705882352942\n",
      "125 0.7105263157894737\n",
      "126 1.0\n",
      "127 0.926829268292683\n",
      "128 0.725\n",
      "129 0.9705882352941176\n",
      "130 0.7837837837837838\n",
      "131 0.8378378378378378\n",
      "132 0.8679245283018868\n",
      "133 0.6458333333333334\n",
      "134 0.7708333333333334\n",
      "135 0.7647058823529411\n",
      "136 0.8\n",
      "137 0.8536585365853658\n",
      "138 0.6923076923076923\n",
      "139 0.6666666666666666\n",
      "140 0.7\n",
      "141 0.825\n",
      "142 0.7352941176470589\n",
      "143 0.71875\n",
      "144 0.7941176470588235\n",
      "145 0.6857142857142857\n",
      "146 0.9166666666666666\n",
      "147 0.868421052631579\n",
      "148 0.6923076923076923\n",
      "149 0.8214285714285714\n",
      "150 0.9032258064516129\n",
      "151 0.8529411764705882\n",
      "152 0.7586206896551724\n",
      "153 0.7777777777777778\n",
      "154 0.8\n",
      "155 0.8620689655172413\n",
      "156 0.6744186046511628\n",
      "157 0.8918918918918919\n",
      "158 0.7142857142857143\n",
      "159 0.8787878787878788\n",
      "160 0.7954545454545454\n",
      "161 0.8378378378378378\n",
      "162 0.8837209302325582\n",
      "163 0.9090909090909091\n",
      "164 0.8095238095238095\n",
      "165 0.96875\n",
      "166 0.8\n",
      "167 0.7948717948717948\n",
      "168 0.8620689655172413\n",
      "169 0.96875\n",
      "170 0.8064516129032258\n",
      "171 0.8571428571428571\n",
      "172 0.8571428571428571\n",
      "173 0.8461538461538461\n",
      "174 0.8205128205128205\n",
      "175 0.925\n",
      "176 0.6756756756756757\n",
      "177 0.8125\n",
      "178 0.9428571428571428\n",
      "179 0.8235294117647058\n",
      "180 0.8461538461538461\n",
      "181 0.725\n",
      "182 0.8048780487804879\n",
      "183 0.9333333333333333\n",
      "184 0.9090909090909091\n",
      "185 0.9230769230769231\n",
      "186 0.7674418604651163\n",
      "187 0.9310344827586207\n",
      "188 0.9473684210526315\n",
      "189 0.88\n",
      "190 0.8148148148148148\n",
      "191 0.8679245283018868\n",
      "192 0.6538461538461539\n",
      "193 0.8333333333333334\n",
      "194 0.9459459459459459\n",
      "195 0.9\n",
      "196 0.7291666666666666\n",
      "197 0.8636363636363636\n",
      "198 0.8709677419354839\n",
      "199 0.7857142857142857\n",
      "200 0.7027027027027027\n",
      "201 0.85\n",
      "202 0.6666666666666666\n",
      "203 0.7777777777777778\n",
      "204 0.84375\n",
      "205 0.7321428571428571\n",
      "206 0.64\n",
      "207 0.5\n",
      "208 0.875\n",
      "209 0.7878787878787878\n",
      "210 0.8913043478260869\n",
      "211 0.7878787878787878\n",
      "212 0.6739130434782609\n",
      "213 0.7692307692307693\n",
      "214 0.6470588235294118\n",
      "215 0.7407407407407407\n",
      "216 0.6216216216216216\n",
      "217 0.8636363636363636\n",
      "218 0.8181818181818182\n",
      "219 0.875\n",
      "220 0.8214285714285714\n",
      "221 0.8888888888888888\n",
      "222 0.8717948717948718\n",
      "223 0.90625\n",
      "224 0.8857142857142857\n",
      "225 0.7352941176470589\n",
      "226 0.875\n",
      "227 0.975\n",
      "228 0.8823529411764706\n",
      "229 0.8518518518518519\n",
      "230 0.84375\n"
     ]
    }
   ],
   "source": [
    "#finetune0\n",
    "for i in range(len(acceach)):\n",
    "    print(i, acceach[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = trans_model_231.predict(x_test)\n",
    "y_pred = pred.argmax(axis=-1)\n",
    "y_true = [np.argmax(p) for p in Y_test]\n",
    "#finetune 1\n",
    "print(classification_report(y_true, y_pred))\n",
    "matrix = confusion_matrix(y_true, y_pred)\n",
    "acceach=matrix.diagonal()/matrix.sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.89473684, 0.88888889, 0.57142857, 0.80769231, 0.78947368,\n",
       "       0.96875   , 0.86842105, 0.68571429, 0.74074074, 0.85      ,\n",
       "       0.92857143, 0.76      , 0.97058824, 0.87096774, 0.88888889,\n",
       "       0.90322581, 0.78125   , 0.96      , 0.92105263, 0.94285714,\n",
       "       0.85294118, 0.91304348, 0.89473684, 0.89473684, 0.81818182,\n",
       "       0.82352941, 0.96969697, 0.92592593, 0.91176471, 0.93548387,\n",
       "       0.89285714, 0.92307692, 0.68421053, 0.94444444, 0.95833333,\n",
       "       0.93548387, 0.9       , 0.84375   , 0.625     , 0.90322581,\n",
       "       0.85714286, 0.77419355, 0.74      , 0.85185185, 0.68292683,\n",
       "       0.88235294, 0.80487805, 0.71428571, 0.89285714, 0.66666667,\n",
       "       0.82051282, 0.78378378, 0.75609756, 0.82926829, 0.74193548,\n",
       "       0.86842105, 0.97368421, 0.72972973, 0.75862069, 0.86363636,\n",
       "       0.63636364, 0.72727273, 0.86666667, 0.79310345, 0.83333333,\n",
       "       0.75      , 0.78125   , 0.45652174, 0.86666667, 0.6969697 ,\n",
       "       0.75609756, 0.96      , 0.87234043, 0.90243902, 0.66666667,\n",
       "       0.90625   , 0.85      , 0.76086957, 0.34146341, 0.35897436,\n",
       "       0.86206897, 0.8       , 1.        , 0.93103448, 0.73333333,\n",
       "       0.86842105, 0.80769231, 0.93181818, 0.84375   , 0.85714286,\n",
       "       0.87804878, 0.72727273, 0.63636364, 0.75      , 0.72972973,\n",
       "       0.74193548, 0.95833333, 0.91666667, 0.83333333, 0.86486486,\n",
       "       0.91428571, 0.96551724, 0.79310345, 0.97435897, 0.89473684,\n",
       "       0.68888889, 0.71052632, 0.82857143, 0.6875    , 0.91428571,\n",
       "       0.96551724, 0.97435897, 0.62162162, 0.78378378, 0.92857143,\n",
       "       0.86842105, 0.63888889, 0.66666667, 0.64102564, 0.66666667,\n",
       "       0.75757576, 0.79310345, 0.77777778, 0.88888889, 0.67647059,\n",
       "       0.71052632, 1.        , 0.92682927, 0.725     , 0.97058824,\n",
       "       0.78378378, 0.83783784, 0.86792453, 0.64583333, 0.77083333,\n",
       "       0.76470588, 0.8       , 0.85365854, 0.69230769, 0.66666667,\n",
       "       0.7       , 0.825     , 0.73529412, 0.71875   , 0.79411765,\n",
       "       0.68571429, 0.91666667, 0.86842105, 0.69230769, 0.82142857,\n",
       "       0.90322581, 0.85294118, 0.75862069, 0.77777778, 0.8       ,\n",
       "       0.86206897, 0.6744186 , 0.89189189, 0.71428571, 0.87878788,\n",
       "       0.79545455, 0.83783784, 0.88372093, 0.90909091, 0.80952381,\n",
       "       0.96875   , 0.8       , 0.79487179, 0.86206897, 0.96875   ,\n",
       "       0.80645161, 0.85714286, 0.85714286, 0.84615385, 0.82051282,\n",
       "       0.925     , 0.67567568, 0.8125    , 0.94285714, 0.82352941,\n",
       "       0.84615385, 0.725     , 0.80487805, 0.93333333, 0.90909091,\n",
       "       0.92307692, 0.76744186, 0.93103448, 0.94736842, 0.88      ,\n",
       "       0.81481481, 0.86792453, 0.65384615, 0.83333333, 0.94594595,\n",
       "       0.9       , 0.72916667, 0.86363636, 0.87096774, 0.78571429,\n",
       "       0.7027027 , 0.85      , 0.66666667, 0.77777778, 0.84375   ,\n",
       "       0.73214286, 0.64      , 0.5       , 0.875     , 0.78787879,\n",
       "       0.89130435, 0.78787879, 0.67391304, 0.76923077, 0.64705882,\n",
       "       0.74074074, 0.62162162, 0.86363636, 0.81818182, 0.875     ,\n",
       "       0.82142857, 0.88888889, 0.87179487, 0.90625   , 0.88571429,\n",
       "       0.73529412, 0.875     , 0.975     , 0.88235294, 0.85185185,\n",
       "       0.84375   ])"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acceach"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.85      0.92      0.89        38\n",
      "           1       0.86      0.86      0.86        36\n",
      "           2       0.70      0.75      0.72        28\n",
      "           3       0.60      0.96      0.74        26\n",
      "           4       0.80      0.84      0.82        38\n",
      "           5       0.91      0.97      0.94        32\n",
      "           6       0.94      0.84      0.89        38\n",
      "           7       0.92      0.66      0.77        35\n",
      "           8       0.96      0.89      0.92        27\n",
      "           9       0.87      0.82      0.85        40\n",
      "          10       0.79      0.90      0.84        42\n",
      "          11       0.70      0.84      0.76        25\n",
      "          12       0.97      0.88      0.92        34\n",
      "          13       0.87      0.87      0.87        31\n",
      "          14       0.82      0.92      0.87        36\n",
      "          15       0.91      0.94      0.92        31\n",
      "          16       0.84      0.81      0.83        32\n",
      "          17       0.63      0.96      0.76        25\n",
      "          18       0.88      0.92      0.90        38\n",
      "          19       1.00      0.97      0.99        35\n",
      "          20       1.00      0.82      0.90        34\n",
      "          21       0.88      0.91      0.89        23\n",
      "          22       0.95      0.92      0.93        38\n",
      "          23       0.92      0.89      0.91        38\n",
      "          24       0.95      0.86      0.90        22\n",
      "          25       0.84      0.91      0.87        34\n",
      "          26       0.86      0.97      0.91        33\n",
      "          27       0.96      0.89      0.92        27\n",
      "          28       1.00      0.88      0.94        34\n",
      "          29       0.88      0.94      0.91        31\n",
      "          30       0.59      0.82      0.69        28\n",
      "          31       0.90      0.95      0.92        39\n",
      "          32       0.88      0.55      0.68        38\n",
      "          33       0.92      0.94      0.93        36\n",
      "          34       0.92      0.92      0.92        24\n",
      "          35       0.90      0.90      0.90        31\n",
      "          36       0.84      0.93      0.88        40\n",
      "          37       0.81      0.78      0.79        32\n",
      "          38       0.76      0.79      0.78        24\n",
      "          39       0.80      0.90      0.85        31\n",
      "          40       1.00      0.93      0.96        28\n",
      "          41       0.79      0.84      0.81        31\n",
      "          42       0.85      0.58      0.69        50\n",
      "          43       0.86      0.89      0.87        27\n",
      "          44       0.83      0.83      0.83        41\n",
      "          45       0.77      0.88      0.82        34\n",
      "          46       0.87      0.83      0.85        41\n",
      "          47       0.71      0.86      0.78        35\n",
      "          48       0.78      0.75      0.76        28\n",
      "          49       0.71      0.62      0.67        24\n",
      "          50       0.84      0.82      0.83        39\n",
      "          51       0.78      0.78      0.78        37\n",
      "          52       0.89      0.83      0.86        41\n",
      "          53       0.68      0.68      0.68        41\n",
      "          54       0.96      0.77      0.86        31\n",
      "          55       0.66      0.82      0.73        38\n",
      "          56       0.93      0.97      0.95        38\n",
      "          57       0.89      0.68      0.77        37\n",
      "          58       0.68      0.79      0.73        29\n",
      "          59       0.95      0.84      0.89        44\n",
      "          60       0.60      0.64      0.62        33\n",
      "          61       0.81      0.79      0.80        33\n",
      "          62       0.94      0.73      0.83        45\n",
      "          63       0.51      0.83      0.63        29\n",
      "          64       0.84      0.86      0.85        36\n",
      "          65       0.70      0.88      0.78        24\n",
      "          66       0.74      0.44      0.55        32\n",
      "          67       0.86      0.67      0.76        46\n",
      "          68       0.86      0.83      0.85        30\n",
      "          69       0.66      0.64      0.65        33\n",
      "          70       0.80      0.80      0.80        41\n",
      "          71       0.61      0.76      0.68        25\n",
      "          72       0.58      0.91      0.71        47\n",
      "          73       0.97      0.85      0.91        41\n",
      "          74       0.65      0.39      0.49        33\n",
      "          75       0.88      0.91      0.89        32\n",
      "          76       0.94      0.82      0.88        40\n",
      "          77       0.70      0.80      0.75        46\n",
      "          78       0.72      0.68      0.70        41\n",
      "          79       0.80      0.90      0.84        39\n",
      "          80       0.84      0.90      0.87        29\n",
      "          81       0.82      0.30      0.44        30\n",
      "          82       0.81      1.00      0.89        38\n",
      "          83       0.83      1.00      0.91        29\n",
      "          84       0.67      0.80      0.73        30\n",
      "          85       0.87      0.89      0.88        38\n",
      "          86       0.71      0.85      0.77        26\n",
      "          87       0.95      0.93      0.94        44\n",
      "          88       0.83      0.75      0.79        32\n",
      "          89       0.83      0.90      0.86        42\n",
      "          90       0.86      0.93      0.89        41\n",
      "          91       0.68      0.52      0.59        33\n",
      "          92       0.85      0.85      0.85        33\n",
      "          93       0.63      0.96      0.76        28\n",
      "          94       0.97      0.92      0.94        37\n",
      "          95       0.86      0.58      0.69        31\n",
      "          96       0.96      0.92      0.94        24\n",
      "          97       0.84      0.89      0.86        36\n",
      "          98       0.88      0.97      0.92        36\n",
      "          99       0.94      0.78      0.85        37\n",
      "         100       0.70      0.91      0.79        35\n",
      "         101       0.75      0.93      0.83        29\n",
      "         102       0.79      0.52      0.62        29\n",
      "         103       0.95      1.00      0.97        39\n",
      "         104       1.00      0.79      0.88        38\n",
      "         105       0.95      0.42      0.58        45\n",
      "         106       0.73      0.29      0.42        38\n",
      "         107       0.69      0.89      0.78        35\n",
      "         108       0.72      0.72      0.72        32\n",
      "         109       0.49      0.83      0.62        35\n",
      "         110       0.68      0.93      0.78        29\n",
      "         111       0.65      0.95      0.77        39\n",
      "         112       0.68      0.51      0.58        37\n",
      "         113       0.83      0.51      0.63        37\n",
      "         114       0.82      0.88      0.85        42\n",
      "         115       0.64      0.61      0.62        38\n",
      "         116       0.82      0.78      0.80        36\n",
      "         117       0.51      0.70      0.59        27\n",
      "         118       0.73      0.62      0.67        39\n",
      "         119       0.53      0.70      0.60        30\n",
      "         120       0.88      0.85      0.86        33\n",
      "         121       0.92      0.79      0.85        29\n",
      "         122       1.00      0.69      0.82        36\n",
      "         123       0.85      0.92      0.88        36\n",
      "         124       0.73      0.65      0.69        34\n",
      "         125       0.71      0.66      0.68        38\n",
      "         126       0.96      1.00      0.98        22\n",
      "         127       0.93      0.90      0.91        41\n",
      "         128       0.86      0.80      0.83        40\n",
      "         129       0.97      0.91      0.94        34\n",
      "         130       0.79      0.89      0.84        37\n",
      "         131       0.84      0.84      0.84        37\n",
      "         132       0.98      0.89      0.93        53\n",
      "         133       0.94      0.71      0.81        48\n",
      "         134       0.76      0.81      0.79        48\n",
      "         135       0.93      0.82      0.87        34\n",
      "         136       0.92      0.80      0.86        30\n",
      "         137       0.90      0.90      0.90        41\n",
      "         138       0.59      0.73      0.66        26\n",
      "         139       0.42      0.33      0.37        30\n",
      "         140       0.53      0.27      0.36        30\n",
      "         141       0.91      0.75      0.82        40\n",
      "         142       0.95      0.62      0.75        34\n",
      "         143       0.44      0.94      0.60        32\n",
      "         144       0.56      0.88      0.68        34\n",
      "         145       0.94      0.43      0.59        35\n",
      "         146       0.59      0.42      0.49        24\n",
      "         147       0.83      0.92      0.88        38\n",
      "         148       0.79      0.85      0.81        26\n",
      "         149       0.95      0.75      0.84        28\n",
      "         150       0.90      0.87      0.89        31\n",
      "         151       0.76      0.85      0.81        34\n",
      "         152       0.96      0.83      0.89        29\n",
      "         153       0.74      0.85      0.79        27\n",
      "         154       0.86      0.63      0.73        30\n",
      "         155       0.90      0.90      0.90        29\n",
      "         156       0.77      0.79      0.78        43\n",
      "         157       1.00      0.92      0.96        37\n",
      "         158       0.88      0.88      0.88        42\n",
      "         159       0.93      0.85      0.89        33\n",
      "         160       0.97      0.86      0.92        44\n",
      "         161       0.86      0.84      0.85        37\n",
      "         162       0.86      0.86      0.86        43\n",
      "         163       0.93      0.86      0.89        44\n",
      "         164       0.83      0.83      0.83        42\n",
      "         165       0.89      0.97      0.93        32\n",
      "         166       0.82      0.77      0.79        35\n",
      "         167       0.97      0.90      0.93        39\n",
      "         168       0.75      0.83      0.79        29\n",
      "         169       0.97      0.97      0.97        32\n",
      "         170       0.86      0.97      0.91        31\n",
      "         171       0.89      0.89      0.89        35\n",
      "         172       0.93      0.80      0.86        35\n",
      "         173       1.00      0.81      0.89        26\n",
      "         174       0.88      0.92      0.90        39\n",
      "         175       1.00      0.72      0.84        40\n",
      "         176       0.83      0.78      0.81        37\n",
      "         177       0.69      0.91      0.78        32\n",
      "         178       0.76      0.91      0.83        35\n",
      "         179       0.93      0.76      0.84        34\n",
      "         180       0.96      0.88      0.92        26\n",
      "         181       0.83      0.50      0.62        40\n",
      "         182       0.80      0.88      0.84        41\n",
      "         183       0.88      0.77      0.82        30\n",
      "         184       0.77      0.91      0.83        22\n",
      "         185       0.79      0.95      0.86        39\n",
      "         186       0.74      0.79      0.76        43\n",
      "         187       0.88      0.97      0.92        29\n",
      "         188       0.92      0.92      0.92        38\n",
      "         189       0.74      1.00      0.85        25\n",
      "         190       0.90      0.70      0.79        27\n",
      "         191       0.98      0.79      0.88        53\n",
      "         192       0.87      0.77      0.82        26\n",
      "         193       0.81      0.92      0.86        24\n",
      "         194       0.97      0.84      0.90        37\n",
      "         195       0.84      0.93      0.88        40\n",
      "         196       0.88      0.60      0.72        48\n",
      "         197       0.86      0.84      0.85        44\n",
      "         198       0.66      0.87      0.75        31\n",
      "         199       0.57      0.93      0.70        14\n",
      "         200       0.86      0.68      0.76        37\n",
      "         201       0.97      0.82      0.89        40\n",
      "         202       0.79      0.85      0.81        39\n",
      "         203       0.85      0.61      0.71        36\n",
      "         204       0.76      0.78      0.77        32\n",
      "         205       0.89      0.73      0.80        56\n",
      "         206       0.86      0.72      0.78        25\n",
      "         207       0.84      0.93      0.88        28\n",
      "         208       0.97      0.91      0.94        32\n",
      "         209       0.76      0.88      0.82        33\n",
      "         210       0.75      0.89      0.81        46\n",
      "         211       0.91      0.61      0.73        33\n",
      "         212       0.52      0.87      0.65        46\n",
      "         213       0.69      0.92      0.79        26\n",
      "         214       0.76      0.38      0.51        34\n",
      "         215       0.71      0.44      0.55        27\n",
      "         216       0.45      0.70      0.55        37\n",
      "         217       0.77      0.82      0.79        44\n",
      "         218       0.54      0.76      0.63        33\n",
      "         219       0.95      0.88      0.91        40\n",
      "         220       0.79      0.93      0.85        28\n",
      "         221       0.96      0.85      0.90        27\n",
      "         222       0.85      0.87      0.86        39\n",
      "         223       1.00      0.88      0.93        32\n",
      "         224       0.67      0.91      0.77        35\n",
      "         225       0.74      0.82      0.78        34\n",
      "         226       0.92      0.90      0.91        40\n",
      "         227       0.93      0.97      0.95        40\n",
      "         228       0.87      0.59      0.70        34\n",
      "         229       0.71      0.89      0.79        27\n",
      "         230       0.91      0.94      0.92        32\n",
      "\n",
      "    accuracy                           0.81      8000\n",
      "   macro avg       0.82      0.81      0.81      8000\n",
      "weighted avg       0.83      0.81      0.81      8000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "#finetune 1\n",
    "print(classification_report(y_true, y_pred))\n",
    "matrix = confusion_matrix(y_true, y_pred)\n",
    "acceach1=matrix.diagonal()/matrix.sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 0.9210526315789473\n",
      "1 0.8611111111111112\n",
      "2 0.75\n",
      "3 0.9615384615384616\n",
      "4 0.8421052631578947\n",
      "5 0.96875\n",
      "6 0.8421052631578947\n",
      "7 0.6571428571428571\n",
      "8 0.8888888888888888\n",
      "9 0.825\n",
      "10 0.9047619047619048\n",
      "11 0.84\n",
      "12 0.8823529411764706\n",
      "13 0.8709677419354839\n",
      "14 0.9166666666666666\n",
      "15 0.9354838709677419\n",
      "16 0.8125\n",
      "17 0.96\n",
      "18 0.9210526315789473\n",
      "19 0.9714285714285714\n",
      "20 0.8235294117647058\n",
      "21 0.9130434782608695\n",
      "22 0.9210526315789473\n",
      "23 0.8947368421052632\n",
      "24 0.8636363636363636\n",
      "25 0.9117647058823529\n",
      "26 0.9696969696969697\n",
      "27 0.8888888888888888\n",
      "28 0.8823529411764706\n",
      "29 0.9354838709677419\n",
      "30 0.8214285714285714\n",
      "31 0.9487179487179487\n",
      "32 0.5526315789473685\n",
      "33 0.9444444444444444\n",
      "34 0.9166666666666666\n",
      "35 0.9032258064516129\n",
      "36 0.925\n",
      "37 0.78125\n",
      "38 0.7916666666666666\n",
      "39 0.9032258064516129\n",
      "40 0.9285714285714286\n",
      "41 0.8387096774193549\n",
      "42 0.58\n",
      "43 0.8888888888888888\n",
      "44 0.8292682926829268\n",
      "45 0.8823529411764706\n",
      "46 0.8292682926829268\n",
      "47 0.8571428571428571\n",
      "48 0.75\n",
      "49 0.625\n",
      "50 0.8205128205128205\n",
      "51 0.7837837837837838\n",
      "52 0.8292682926829268\n",
      "53 0.6829268292682927\n",
      "54 0.7741935483870968\n",
      "55 0.8157894736842105\n",
      "56 0.9736842105263158\n",
      "57 0.6756756756756757\n",
      "58 0.7931034482758621\n",
      "59 0.8409090909090909\n",
      "60 0.6363636363636364\n",
      "61 0.7878787878787878\n",
      "62 0.7333333333333333\n",
      "63 0.8275862068965517\n",
      "64 0.8611111111111112\n",
      "65 0.875\n",
      "66 0.4375\n",
      "67 0.6739130434782609\n",
      "68 0.8333333333333334\n",
      "69 0.6363636363636364\n",
      "70 0.8048780487804879\n",
      "71 0.76\n",
      "72 0.9148936170212766\n",
      "73 0.8536585365853658\n",
      "74 0.3939393939393939\n",
      "75 0.90625\n",
      "76 0.825\n",
      "77 0.8043478260869565\n",
      "78 0.6829268292682927\n",
      "79 0.8974358974358975\n",
      "80 0.896551724137931\n",
      "81 0.3\n",
      "82 1.0\n",
      "83 1.0\n",
      "84 0.8\n",
      "85 0.8947368421052632\n",
      "86 0.8461538461538461\n",
      "87 0.9318181818181818\n",
      "88 0.75\n",
      "89 0.9047619047619048\n",
      "90 0.926829268292683\n",
      "91 0.5151515151515151\n",
      "92 0.8484848484848485\n",
      "93 0.9642857142857143\n",
      "94 0.918918918918919\n",
      "95 0.5806451612903226\n",
      "96 0.9166666666666666\n",
      "97 0.8888888888888888\n",
      "98 0.9722222222222222\n",
      "99 0.7837837837837838\n",
      "100 0.9142857142857143\n",
      "101 0.9310344827586207\n",
      "102 0.5172413793103449\n",
      "103 1.0\n",
      "104 0.7894736842105263\n",
      "105 0.4222222222222222\n",
      "106 0.2894736842105263\n",
      "107 0.8857142857142857\n",
      "108 0.71875\n",
      "109 0.8285714285714286\n",
      "110 0.9310344827586207\n",
      "111 0.9487179487179487\n",
      "112 0.5135135135135135\n",
      "113 0.5135135135135135\n",
      "114 0.8809523809523809\n",
      "115 0.6052631578947368\n",
      "116 0.7777777777777778\n",
      "117 0.7037037037037037\n",
      "118 0.6153846153846154\n",
      "119 0.7\n",
      "120 0.8484848484848485\n",
      "121 0.7931034482758621\n",
      "122 0.6944444444444444\n",
      "123 0.9166666666666666\n",
      "124 0.6470588235294118\n",
      "125 0.6578947368421053\n",
      "126 1.0\n",
      "127 0.9024390243902439\n",
      "128 0.8\n",
      "129 0.9117647058823529\n",
      "130 0.8918918918918919\n",
      "131 0.8378378378378378\n",
      "132 0.8867924528301887\n",
      "133 0.7083333333333334\n",
      "134 0.8125\n",
      "135 0.8235294117647058\n",
      "136 0.8\n",
      "137 0.9024390243902439\n",
      "138 0.7307692307692307\n",
      "139 0.3333333333333333\n",
      "140 0.26666666666666666\n",
      "141 0.75\n",
      "142 0.6176470588235294\n",
      "143 0.9375\n",
      "144 0.8823529411764706\n",
      "145 0.42857142857142855\n",
      "146 0.4166666666666667\n",
      "147 0.9210526315789473\n",
      "148 0.8461538461538461\n",
      "149 0.75\n",
      "150 0.8709677419354839\n",
      "151 0.8529411764705882\n",
      "152 0.8275862068965517\n",
      "153 0.8518518518518519\n",
      "154 0.6333333333333333\n",
      "155 0.896551724137931\n",
      "156 0.7906976744186046\n",
      "157 0.918918918918919\n",
      "158 0.8809523809523809\n",
      "159 0.8484848484848485\n",
      "160 0.8636363636363636\n",
      "161 0.8378378378378378\n",
      "162 0.8604651162790697\n",
      "163 0.8636363636363636\n",
      "164 0.8333333333333334\n",
      "165 0.96875\n",
      "166 0.7714285714285715\n",
      "167 0.8974358974358975\n",
      "168 0.8275862068965517\n",
      "169 0.96875\n",
      "170 0.967741935483871\n",
      "171 0.8857142857142857\n",
      "172 0.8\n",
      "173 0.8076923076923077\n",
      "174 0.9230769230769231\n",
      "175 0.725\n",
      "176 0.7837837837837838\n",
      "177 0.90625\n",
      "178 0.9142857142857143\n",
      "179 0.7647058823529411\n",
      "180 0.8846153846153846\n",
      "181 0.5\n",
      "182 0.8780487804878049\n",
      "183 0.7666666666666667\n",
      "184 0.9090909090909091\n",
      "185 0.9487179487179487\n",
      "186 0.7906976744186046\n",
      "187 0.9655172413793104\n",
      "188 0.9210526315789473\n",
      "189 1.0\n",
      "190 0.7037037037037037\n",
      "191 0.7924528301886793\n",
      "192 0.7692307692307693\n",
      "193 0.9166666666666666\n",
      "194 0.8378378378378378\n",
      "195 0.925\n",
      "196 0.6041666666666666\n",
      "197 0.8409090909090909\n",
      "198 0.8709677419354839\n",
      "199 0.9285714285714286\n",
      "200 0.6756756756756757\n",
      "201 0.825\n",
      "202 0.8461538461538461\n",
      "203 0.6111111111111112\n",
      "204 0.78125\n",
      "205 0.7321428571428571\n",
      "206 0.72\n",
      "207 0.9285714285714286\n",
      "208 0.90625\n",
      "209 0.8787878787878788\n",
      "210 0.8913043478260869\n",
      "211 0.6060606060606061\n",
      "212 0.8695652173913043\n",
      "213 0.9230769230769231\n",
      "214 0.38235294117647056\n",
      "215 0.4444444444444444\n",
      "216 0.7027027027027027\n",
      "217 0.8181818181818182\n",
      "218 0.7575757575757576\n",
      "219 0.875\n",
      "220 0.9285714285714286\n",
      "221 0.8518518518518519\n",
      "222 0.8717948717948718\n",
      "223 0.875\n",
      "224 0.9142857142857143\n",
      "225 0.8235294117647058\n",
      "226 0.9\n",
      "227 0.975\n",
      "228 0.5882352941176471\n",
      "229 0.8888888888888888\n",
      "230 0.9375\n",
      "187.0397124758341\n"
     ]
    }
   ],
   "source": [
    "#finetune1\n",
    "sum=0\n",
    "for i in range(len(acceach1)):\n",
    "    sum=sum+acceach1[i]\n",
    "    print(i, acceach1[i])\n",
    "\n",
    "print(sum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.97      0.89      0.93        38\n",
      "           1       0.96      0.75      0.84        36\n",
      "           2       0.79      0.79      0.79        28\n",
      "           3       0.89      0.92      0.91        26\n",
      "           4       0.94      0.87      0.90        38\n",
      "           5       0.97      0.94      0.95        32\n",
      "           6       0.91      0.82      0.86        38\n",
      "           7       0.81      0.83      0.82        35\n",
      "           8       0.60      0.96      0.74        27\n",
      "           9       0.82      0.93      0.87        40\n",
      "          10       0.97      0.86      0.91        42\n",
      "          11       0.94      0.68      0.79        25\n",
      "          12       0.86      0.88      0.87        34\n",
      "          13       0.82      0.87      0.84        31\n",
      "          14       0.91      0.83      0.87        36\n",
      "          15       0.78      0.90      0.84        31\n",
      "          16       0.82      0.84      0.83        32\n",
      "          17       0.68      0.92      0.78        25\n",
      "          18       0.96      0.66      0.78        38\n",
      "          19       0.94      0.86      0.90        35\n",
      "          20       0.79      0.88      0.83        34\n",
      "          21       1.00      0.87      0.93        23\n",
      "          22       0.92      0.89      0.91        38\n",
      "          23       0.78      0.84      0.81        38\n",
      "          24       0.83      0.86      0.84        22\n",
      "          25       0.88      0.88      0.88        34\n",
      "          26       0.91      0.97      0.94        33\n",
      "          27       0.87      0.96      0.91        27\n",
      "          28       1.00      0.88      0.94        34\n",
      "          29       0.97      0.94      0.95        31\n",
      "          30       0.73      0.79      0.76        28\n",
      "          31       0.88      0.95      0.91        39\n",
      "          32       0.89      0.63      0.74        38\n",
      "          33       0.83      0.94      0.88        36\n",
      "          34       0.92      0.96      0.94        24\n",
      "          35       0.89      0.81      0.85        31\n",
      "          36       0.92      0.90      0.91        40\n",
      "          37       0.79      0.81      0.80        32\n",
      "          38       0.87      0.83      0.85        24\n",
      "          39       0.89      0.81      0.85        31\n",
      "          40       0.92      0.82      0.87        28\n",
      "          41       0.87      0.87      0.87        31\n",
      "          42       0.82      0.66      0.73        50\n",
      "          43       0.92      0.85      0.88        27\n",
      "          44       0.82      0.80      0.81        41\n",
      "          45       0.86      0.88      0.87        34\n",
      "          46       0.88      0.71      0.78        41\n",
      "          47       0.75      0.77      0.76        35\n",
      "          48       0.71      0.79      0.75        28\n",
      "          49       0.62      0.96      0.75        24\n",
      "          50       0.85      0.87      0.86        39\n",
      "          51       0.79      0.70      0.74        37\n",
      "          52       0.82      0.80      0.81        41\n",
      "          53       0.91      0.71      0.79        41\n",
      "          54       0.65      0.77      0.71        31\n",
      "          55       0.87      0.71      0.78        38\n",
      "          56       0.93      1.00      0.96        38\n",
      "          57       0.86      0.68      0.76        37\n",
      "          58       0.83      0.66      0.73        29\n",
      "          59       0.78      0.86      0.82        44\n",
      "          60       0.86      0.73      0.79        33\n",
      "          61       0.75      0.82      0.78        33\n",
      "          62       0.85      0.89      0.87        45\n",
      "          63       0.70      0.72      0.71        29\n",
      "          64       0.71      0.83      0.77        36\n",
      "          65       0.94      0.67      0.78        24\n",
      "          66       0.65      0.75      0.70        32\n",
      "          67       0.84      0.70      0.76        46\n",
      "          68       0.79      0.90      0.84        30\n",
      "          69       0.68      0.79      0.73        33\n",
      "          70       0.95      0.85      0.90        41\n",
      "          71       0.57      0.84      0.68        25\n",
      "          72       0.89      0.87      0.88        47\n",
      "          73       0.97      0.90      0.94        41\n",
      "          74       0.80      0.73      0.76        33\n",
      "          75       0.87      0.84      0.86        32\n",
      "          76       0.95      0.88      0.91        40\n",
      "          77       0.95      0.76      0.84        46\n",
      "          78       0.92      0.80      0.86        41\n",
      "          79       0.89      0.87      0.88        39\n",
      "          80       0.90      0.93      0.92        29\n",
      "          81       0.72      0.77      0.74        30\n",
      "          82       0.95      0.95      0.95        38\n",
      "          83       0.97      0.97      0.97        29\n",
      "          84       0.84      0.90      0.87        30\n",
      "          85       0.94      0.82      0.87        38\n",
      "          86       0.92      0.42      0.58        26\n",
      "          87       0.91      0.93      0.92        44\n",
      "          88       0.60      0.81      0.69        32\n",
      "          89       0.97      0.86      0.91        42\n",
      "          90       0.90      0.90      0.90        41\n",
      "          91       0.90      0.85      0.88        33\n",
      "          92       0.88      0.85      0.86        33\n",
      "          93       0.96      0.79      0.86        28\n",
      "          94       0.89      0.86      0.88        37\n",
      "          95       0.87      0.87      0.87        31\n",
      "          96       0.71      1.00      0.83        24\n",
      "          97       0.65      0.94      0.77        36\n",
      "          98       1.00      0.92      0.96        36\n",
      "          99       0.89      0.86      0.88        37\n",
      "         100       0.79      0.89      0.84        35\n",
      "         101       0.88      0.97      0.92        29\n",
      "         102       0.76      0.66      0.70        29\n",
      "         103       0.95      1.00      0.97        39\n",
      "         104       0.92      0.92      0.92        38\n",
      "         105       0.77      0.76      0.76        45\n",
      "         106       0.89      0.82      0.85        38\n",
      "         107       0.94      0.91      0.93        35\n",
      "         108       0.79      0.81      0.80        32\n",
      "         109       0.84      0.89      0.86        35\n",
      "         110       0.67      0.76      0.71        29\n",
      "         111       0.77      0.85      0.80        39\n",
      "         112       0.53      0.62      0.57        37\n",
      "         113       0.62      0.84      0.71        37\n",
      "         114       1.00      0.86      0.92        42\n",
      "         115       0.68      0.79      0.73        38\n",
      "         116       0.76      0.72      0.74        36\n",
      "         117       0.77      0.63      0.69        27\n",
      "         118       0.81      0.54      0.65        39\n",
      "         119       0.68      0.63      0.66        30\n",
      "         120       0.97      0.85      0.90        33\n",
      "         121       0.76      0.86      0.81        29\n",
      "         122       0.84      0.72      0.78        36\n",
      "         123       0.81      0.72      0.76        36\n",
      "         124       0.95      0.56      0.70        34\n",
      "         125       0.84      0.82      0.83        38\n",
      "         126       0.85      1.00      0.92        22\n",
      "         127       0.93      0.95      0.94        41\n",
      "         128       0.87      0.82      0.85        40\n",
      "         129       0.89      1.00      0.94        34\n",
      "         130       0.86      0.86      0.86        37\n",
      "         131       0.97      0.86      0.91        37\n",
      "         132       1.00      0.89      0.94        53\n",
      "         133       0.80      0.90      0.84        48\n",
      "         134       0.86      0.77      0.81        48\n",
      "         135       0.89      0.74      0.81        34\n",
      "         136       0.74      0.87      0.80        30\n",
      "         137       0.86      0.90      0.88        41\n",
      "         138       0.64      0.81      0.71        26\n",
      "         139       0.64      0.70      0.67        30\n",
      "         140       0.73      0.73      0.73        30\n",
      "         141       0.73      0.82      0.78        40\n",
      "         142       0.92      0.71      0.80        34\n",
      "         143       0.73      0.59      0.66        32\n",
      "         144       0.72      0.82      0.77        34\n",
      "         145       0.81      0.71      0.76        35\n",
      "         146       0.55      0.75      0.63        24\n",
      "         147       0.69      0.89      0.78        38\n",
      "         148       0.87      0.77      0.82        26\n",
      "         149       0.96      0.86      0.91        28\n",
      "         150       0.69      0.87      0.77        31\n",
      "         151       0.71      0.85      0.77        34\n",
      "         152       0.92      0.83      0.87        29\n",
      "         153       0.83      0.74      0.78        27\n",
      "         154       0.87      0.87      0.87        30\n",
      "         155       0.77      0.83      0.80        29\n",
      "         156       0.81      0.79      0.80        43\n",
      "         157       0.92      0.95      0.93        37\n",
      "         158       0.91      0.93      0.92        42\n",
      "         159       0.84      0.82      0.83        33\n",
      "         160       0.90      0.84      0.87        44\n",
      "         161       0.93      0.76      0.84        37\n",
      "         162       0.80      0.81      0.80        43\n",
      "         163       0.86      0.95      0.90        44\n",
      "         164       0.73      0.83      0.78        42\n",
      "         165       0.91      0.94      0.92        32\n",
      "         166       0.71      0.77      0.74        35\n",
      "         167       0.76      0.90      0.82        39\n",
      "         168       0.81      0.76      0.79        29\n",
      "         169       0.86      0.97      0.91        32\n",
      "         170       0.84      0.87      0.86        31\n",
      "         171       0.97      0.80      0.88        35\n",
      "         172       0.96      0.74      0.84        35\n",
      "         173       0.85      0.85      0.85        26\n",
      "         174       0.94      0.85      0.89        39\n",
      "         175       0.90      0.90      0.90        40\n",
      "         176       0.70      0.76      0.73        37\n",
      "         177       1.00      0.59      0.75        32\n",
      "         178       0.97      0.91      0.94        35\n",
      "         179       0.71      0.85      0.77        34\n",
      "         180       0.84      0.81      0.82        26\n",
      "         181       0.84      0.65      0.73        40\n",
      "         182       0.68      0.88      0.77        41\n",
      "         183       0.83      0.80      0.81        30\n",
      "         184       0.65      0.91      0.75        22\n",
      "         185       0.91      0.77      0.83        39\n",
      "         186       0.86      0.58      0.69        43\n",
      "         187       0.93      0.93      0.93        29\n",
      "         188       0.72      0.95      0.82        38\n",
      "         189       0.79      0.92      0.85        25\n",
      "         190       0.68      0.78      0.72        27\n",
      "         191       0.98      0.91      0.94        53\n",
      "         192       0.69      0.77      0.73        26\n",
      "         193       0.95      0.75      0.84        24\n",
      "         194       0.92      0.92      0.92        37\n",
      "         195       0.87      0.82      0.85        40\n",
      "         196       0.74      0.42      0.53        48\n",
      "         197       0.79      0.86      0.83        44\n",
      "         198       0.80      0.90      0.85        31\n",
      "         199       0.45      0.36      0.40        14\n",
      "         200       1.00      0.65      0.79        37\n",
      "         201       0.95      0.93      0.94        40\n",
      "         202       0.66      0.74      0.70        39\n",
      "         203       0.73      0.61      0.67        36\n",
      "         204       0.67      0.81      0.73        32\n",
      "         205       0.88      0.41      0.56        56\n",
      "         206       0.54      0.80      0.65        25\n",
      "         207       0.50      0.93      0.65        28\n",
      "         208       0.83      0.91      0.87        32\n",
      "         209       0.55      0.88      0.67        33\n",
      "         210       0.97      0.74      0.84        46\n",
      "         211       0.81      0.79      0.80        33\n",
      "         212       0.63      0.59      0.61        46\n",
      "         213       0.69      0.92      0.79        26\n",
      "         214       0.61      0.79      0.69        34\n",
      "         215       0.74      0.63      0.68        27\n",
      "         216       0.69      0.54      0.61        37\n",
      "         217       0.86      0.82      0.84        44\n",
      "         218       0.66      0.88      0.75        33\n",
      "         219       0.73      0.90      0.81        40\n",
      "         220       0.93      0.89      0.91        28\n",
      "         221       0.88      0.85      0.87        27\n",
      "         222       0.78      0.97      0.86        39\n",
      "         223       0.97      0.94      0.95        32\n",
      "         224       0.70      0.91      0.79        35\n",
      "         225       0.87      0.79      0.83        34\n",
      "         226       0.85      0.88      0.86        40\n",
      "         227       0.78      0.97      0.87        40\n",
      "         228       0.78      0.74      0.76        34\n",
      "         229       0.93      0.93      0.93        27\n",
      "         230       0.90      0.88      0.89        32\n",
      "\n",
      "    accuracy                           0.82      8000\n",
      "   macro avg       0.83      0.82      0.82      8000\n",
      "weighted avg       0.83      0.82      0.82      8000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "pred = trans_model_231.predict(x_test)\n",
    "y_pred = pred.argmax(axis=-1)\n",
    "y_true = [np.argmax(p) for p in Y_test]\n",
    "#finetune 2\n",
    "print(classification_report(y_true, y_pred))\n",
    "matrix = confusion_matrix(y_true, y_pred)\n",
    "acceach2=matrix.diagonal()/matrix.sum(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 0.8947368421052632\n",
      "1 0.75\n",
      "2 0.7857142857142857\n",
      "3 0.9230769230769231\n",
      "4 0.868421052631579\n",
      "5 0.9375\n",
      "6 0.8157894736842105\n",
      "7 0.8285714285714286\n",
      "8 0.9629629629629629\n",
      "9 0.925\n",
      "10 0.8571428571428571\n",
      "11 0.68\n",
      "12 0.8823529411764706\n",
      "13 0.8709677419354839\n",
      "14 0.8333333333333334\n",
      "15 0.9032258064516129\n",
      "16 0.84375\n",
      "17 0.92\n",
      "18 0.6578947368421053\n",
      "19 0.8571428571428571\n",
      "20 0.8823529411764706\n",
      "21 0.8695652173913043\n",
      "22 0.8947368421052632\n",
      "23 0.8421052631578947\n",
      "24 0.8636363636363636\n",
      "25 0.8823529411764706\n",
      "26 0.9696969696969697\n",
      "27 0.9629629629629629\n",
      "28 0.8823529411764706\n",
      "29 0.9354838709677419\n",
      "30 0.7857142857142857\n",
      "31 0.9487179487179487\n",
      "32 0.631578947368421\n",
      "33 0.9444444444444444\n",
      "34 0.9583333333333334\n",
      "35 0.8064516129032258\n",
      "36 0.9\n",
      "37 0.8125\n",
      "38 0.8333333333333334\n",
      "39 0.8064516129032258\n",
      "40 0.8214285714285714\n",
      "41 0.8709677419354839\n",
      "42 0.66\n",
      "43 0.8518518518518519\n",
      "44 0.8048780487804879\n",
      "45 0.8823529411764706\n",
      "46 0.7073170731707317\n",
      "47 0.7714285714285715\n",
      "48 0.7857142857142857\n",
      "49 0.9583333333333334\n",
      "50 0.8717948717948718\n",
      "51 0.7027027027027027\n",
      "52 0.8048780487804879\n",
      "53 0.7073170731707317\n",
      "54 0.7741935483870968\n",
      "55 0.7105263157894737\n",
      "56 1.0\n",
      "57 0.6756756756756757\n",
      "58 0.6551724137931034\n",
      "59 0.8636363636363636\n",
      "60 0.7272727272727273\n",
      "61 0.8181818181818182\n",
      "62 0.8888888888888888\n",
      "63 0.7241379310344828\n",
      "64 0.8333333333333334\n",
      "65 0.6666666666666666\n",
      "66 0.75\n",
      "67 0.6956521739130435\n",
      "68 0.9\n",
      "69 0.7878787878787878\n",
      "70 0.8536585365853658\n",
      "71 0.84\n",
      "72 0.8723404255319149\n",
      "73 0.9024390243902439\n",
      "74 0.7272727272727273\n",
      "75 0.84375\n",
      "76 0.875\n",
      "77 0.7608695652173914\n",
      "78 0.8048780487804879\n",
      "79 0.8717948717948718\n",
      "80 0.9310344827586207\n",
      "81 0.7666666666666667\n",
      "82 0.9473684210526315\n",
      "83 0.9655172413793104\n",
      "84 0.9\n",
      "85 0.8157894736842105\n",
      "86 0.4230769230769231\n",
      "87 0.9318181818181818\n",
      "88 0.8125\n",
      "89 0.8571428571428571\n",
      "90 0.9024390243902439\n",
      "91 0.8484848484848485\n",
      "92 0.8484848484848485\n",
      "93 0.7857142857142857\n",
      "94 0.8648648648648649\n",
      "95 0.8709677419354839\n",
      "96 1.0\n",
      "97 0.9444444444444444\n",
      "98 0.9166666666666666\n",
      "99 0.8648648648648649\n",
      "100 0.8857142857142857\n",
      "101 0.9655172413793104\n",
      "102 0.6551724137931034\n",
      "103 1.0\n",
      "104 0.9210526315789473\n",
      "105 0.7555555555555555\n",
      "106 0.8157894736842105\n",
      "107 0.9142857142857143\n",
      "108 0.8125\n",
      "109 0.8857142857142857\n",
      "110 0.7586206896551724\n",
      "111 0.8461538461538461\n",
      "112 0.6216216216216216\n",
      "113 0.8378378378378378\n",
      "114 0.8571428571428571\n",
      "115 0.7894736842105263\n",
      "116 0.7222222222222222\n",
      "117 0.6296296296296297\n",
      "118 0.5384615384615384\n",
      "119 0.6333333333333333\n",
      "120 0.8484848484848485\n",
      "121 0.8620689655172413\n",
      "122 0.7222222222222222\n",
      "123 0.7222222222222222\n",
      "124 0.5588235294117647\n",
      "125 0.8157894736842105\n",
      "126 1.0\n",
      "127 0.9512195121951219\n",
      "128 0.825\n",
      "129 1.0\n",
      "130 0.8648648648648649\n",
      "131 0.8648648648648649\n",
      "132 0.8867924528301887\n",
      "133 0.8958333333333334\n",
      "134 0.7708333333333334\n",
      "135 0.7352941176470589\n",
      "136 0.8666666666666667\n",
      "137 0.9024390243902439\n",
      "138 0.8076923076923077\n",
      "139 0.7\n",
      "140 0.7333333333333333\n",
      "141 0.825\n",
      "142 0.7058823529411765\n",
      "143 0.59375\n",
      "144 0.8235294117647058\n",
      "145 0.7142857142857143\n",
      "146 0.75\n",
      "147 0.8947368421052632\n",
      "148 0.7692307692307693\n",
      "149 0.8571428571428571\n",
      "150 0.8709677419354839\n",
      "151 0.8529411764705882\n",
      "152 0.8275862068965517\n",
      "153 0.7407407407407407\n",
      "154 0.8666666666666667\n",
      "155 0.8275862068965517\n",
      "156 0.7906976744186046\n",
      "157 0.9459459459459459\n",
      "158 0.9285714285714286\n",
      "159 0.8181818181818182\n",
      "160 0.8409090909090909\n",
      "161 0.7567567567567568\n",
      "162 0.813953488372093\n",
      "163 0.9545454545454546\n",
      "164 0.8333333333333334\n",
      "165 0.9375\n",
      "166 0.7714285714285715\n",
      "167 0.8974358974358975\n",
      "168 0.7586206896551724\n",
      "169 0.96875\n",
      "170 0.8709677419354839\n",
      "171 0.8\n",
      "172 0.7428571428571429\n",
      "173 0.8461538461538461\n",
      "174 0.8461538461538461\n",
      "175 0.9\n",
      "176 0.7567567567567568\n",
      "177 0.59375\n",
      "178 0.9142857142857143\n",
      "179 0.8529411764705882\n",
      "180 0.8076923076923077\n",
      "181 0.65\n",
      "182 0.8780487804878049\n",
      "183 0.8\n",
      "184 0.9090909090909091\n",
      "185 0.7692307692307693\n",
      "186 0.5813953488372093\n",
      "187 0.9310344827586207\n",
      "188 0.9473684210526315\n",
      "189 0.92\n",
      "190 0.7777777777777778\n",
      "191 0.9056603773584906\n",
      "192 0.7692307692307693\n",
      "193 0.75\n",
      "194 0.918918918918919\n",
      "195 0.825\n",
      "196 0.4166666666666667\n",
      "197 0.8636363636363636\n",
      "198 0.9032258064516129\n",
      "199 0.35714285714285715\n",
      "200 0.6486486486486487\n",
      "201 0.925\n",
      "202 0.7435897435897436\n",
      "203 0.6111111111111112\n",
      "204 0.8125\n",
      "205 0.4107142857142857\n",
      "206 0.8\n",
      "207 0.9285714285714286\n",
      "208 0.90625\n",
      "209 0.8787878787878788\n",
      "210 0.7391304347826086\n",
      "211 0.7878787878787878\n",
      "212 0.5869565217391305\n",
      "213 0.9230769230769231\n",
      "214 0.7941176470588235\n",
      "215 0.6296296296296297\n",
      "216 0.5405405405405406\n",
      "217 0.8181818181818182\n",
      "218 0.8787878787878788\n",
      "219 0.9\n",
      "220 0.8928571428571429\n",
      "221 0.8518518518518519\n",
      "222 0.9743589743589743\n",
      "223 0.9375\n",
      "224 0.9142857142857143\n",
      "225 0.7941176470588235\n",
      "226 0.875\n",
      "227 0.975\n",
      "228 0.7352941176470589\n",
      "229 0.9259259259259259\n",
      "230 0.875\n",
      "189.5689580759914\n"
     ]
    }
   ],
   "source": [
    "#finetune1\n",
    "sum=0\n",
    "for i in range(len(acceach2)):\n",
    "    sum=sum+acceach2[i]\n",
    "    print(i, acceach2[i])\n",
    "\n",
    "print(sum)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# function to freezing the layers of the orginal as per our need... we just set\n",
    "# the tune_layer value for the model layers that should not be freez--->for 33/7class\n",
    "\n",
    "def fine_tunerc(tune_layerrc=0):\n",
    "    if tune_layerrc>0:\n",
    "        for layer in model.layers[:-tune_layerrc]:\n",
    "            layer.trainable=False\n",
    "    else:\n",
    "        for layer in model.layers:\n",
    "            layer.trainable = False\n",
    "            \n",
    "            \n",
    "    last_conv_out= model.output\n",
    "    flatten= Flatten()(last_conv_out)\n",
    "    fc1= Dense(1024, activation='relu')(flatten)\n",
    "    fc2= Dense(2048, activation='relu')(fc1)\n",
    "    out_r =Dense(33 ,activation=\"softmax\")(fc2)\n",
    "    out_c =Dense(7,activation=\"softmax\")(fc2)\n",
    "\n",
    "    model_rc=Model(inputs=model.inputs, outputs=[out_r, out_c])\n",
    "    \n",
    "    model_rc.compile(optimizer='adam',loss=['categorical_crossentropy','categorical_crossentropy'],\n",
    "                     metrics=['accuracy'])\n",
    "    \n",
    "    return model_rc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans_model_rc=fine_tunerc(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Early Stopping to prevent overfitting and checkpoint to save model at each epoch231\n",
    "early_stopper = tf.keras.callbacks.EarlyStopping(monitor=\"val_loss\", \n",
    "                                                 patience=5, verbose=2, \n",
    "                                                 restore_best_weights=True)\n",
    "checkpoint = ModelCheckpoint('./vgg/modelVGGrc{epoch:01d}.hdf5', period=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "2025/2025 [==============================] - 146s 72ms/step - loss: 1.6904 - dense_63_loss: 0.7222 - dense_64_loss: 0.9682 - dense_63_accuracy: 0.8409 - dense_64_accuracy: 0.7141 - val_loss: 1.1689 - val_dense_63_loss: 0.4714 - val_dense_64_loss: 0.6974 - val_dense_63_accuracy: 0.8739 - val_dense_64_accuracy: 0.7544\n",
      "Epoch 2/10\n",
      "2025/2025 [==============================] - 145s 72ms/step - loss: 0.9349 - dense_63_loss: 0.3958 - dense_64_loss: 0.5391 - dense_63_accuracy: 0.8890 - dense_64_accuracy: 0.8091 - val_loss: 0.9326 - val_dense_63_loss: 0.4148 - val_dense_64_loss: 0.5178 - val_dense_63_accuracy: 0.8861 - val_dense_64_accuracy: 0.8194\n",
      "Epoch 3/10\n",
      "2025/2025 [==============================] - 145s 71ms/step - loss: 0.8467 - dense_63_loss: 0.3649 - dense_64_loss: 0.4817 - dense_63_accuracy: 0.8967 - dense_64_accuracy: 0.8297 - val_loss: 0.9137 - val_dense_63_loss: 0.4100 - val_dense_64_loss: 0.5036 - val_dense_63_accuracy: 0.8935 - val_dense_64_accuracy: 0.8283\n",
      "Epoch 4/10\n",
      "2025/2025 [==============================] - 145s 72ms/step - loss: 0.7808 - dense_63_loss: 0.3390 - dense_64_loss: 0.4419 - dense_63_accuracy: 0.9052 - dense_64_accuracy: 0.8444 - val_loss: 0.9119 - val_dense_63_loss: 0.4226 - val_dense_64_loss: 0.4893 - val_dense_63_accuracy: 0.8949 - val_dense_64_accuracy: 0.8392\n",
      "Epoch 5/10\n",
      "2025/2025 [==============================] - 144s 71ms/step - loss: 0.7349 - dense_63_loss: 0.3200 - dense_64_loss: 0.4150 - dense_63_accuracy: 0.9095 - dense_64_accuracy: 0.8537 - val_loss: 0.9532 - val_dense_63_loss: 0.4042 - val_dense_64_loss: 0.5490 - val_dense_63_accuracy: 0.8963 - val_dense_64_accuracy: 0.8174\n",
      "Epoch 6/10\n",
      "2025/2025 [==============================] - 144s 71ms/step - loss: 0.6891 - dense_63_loss: 0.2956 - dense_64_loss: 0.3935 - dense_63_accuracy: 0.9161 - dense_64_accuracy: 0.8620 - val_loss: 0.9516 - val_dense_63_loss: 0.4671 - val_dense_64_loss: 0.4845 - val_dense_63_accuracy: 0.8936 - val_dense_64_accuracy: 0.8429\n",
      "Epoch 7/10\n",
      "2025/2025 [==============================] - 146s 72ms/step - loss: 0.6646 - dense_63_loss: 0.2858 - dense_64_loss: 0.3788 - dense_63_accuracy: 0.9196 - dense_64_accuracy: 0.8691 - val_loss: 0.9910 - val_dense_63_loss: 0.4412 - val_dense_64_loss: 0.5498 - val_dense_63_accuracy: 0.8913 - val_dense_64_accuracy: 0.8196\n",
      "Epoch 8/10\n",
      "2025/2025 [==============================] - 148s 73ms/step - loss: 0.6368 - dense_63_loss: 0.2754 - dense_64_loss: 0.3614 - dense_63_accuracy: 0.9224 - dense_64_accuracy: 0.8739 - val_loss: 0.9720 - val_dense_63_loss: 0.4825 - val_dense_64_loss: 0.4895 - val_dense_63_accuracy: 0.8982 - val_dense_64_accuracy: 0.8554\n",
      "Epoch 9/10\n",
      "2025/2025 [==============================] - 143s 70ms/step - loss: 0.6175 - dense_63_loss: 0.2658 - dense_64_loss: 0.3517 - dense_63_accuracy: 0.9256 - dense_64_accuracy: 0.8780 - val_loss: 0.9420 - val_dense_63_loss: 0.4422 - val_dense_64_loss: 0.4998 - val_dense_63_accuracy: 0.8918 - val_dense_64_accuracy: 0.8342\n",
      "Epoch 10/10\n",
      "2025/2025 [==============================] - 147s 73ms/step - loss: 0.5886 - dense_63_loss: 0.2507 - dense_64_loss: 0.3380 - dense_63_accuracy: 0.9301 - dense_64_accuracy: 0.8836 - val_loss: 1.0274 - val_dense_63_loss: 0.4907 - val_dense_64_loss: 0.5367 - val_dense_63_accuracy: 0.9008 - val_dense_64_accuracy: 0.8467\n"
     ]
    }
   ],
   "source": [
    "hist_rc = trans_model_rc.fit(x_train, [Y_trainr, Y_trainc], batch_size=32, epochs=10,\n",
    "                  verbose=1, validation_split=0.1,shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [],
   "source": [
    "trans_model_rc.save('trans_model_rc_0.hdf5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "500/500 [==============================] - 18s 34ms/step - loss: 0.9941 - dense_63_loss: 0.4734 - dense_64_loss: 0.5207 - dense_63_accuracy: 0.8988 - dense_64_accuracy: 0.8482\n"
     ]
    }
   ],
   "source": [
    "score = trans_model_rc.evaluate(x_test,[Y_testr,Y_testc],batch_size=16, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test lossr: 0.4734340310096741\n",
      "Test lossc: 0.5206609964370728\n",
      "Test accuracyr: 0.8987500071525574\n",
      "Test accuracyc: 0.8482499718666077\n"
     ]
    }
   ],
   "source": [
    "print('Test lossr:', score[1])\n",
    "print('Test lossc:', score[2])\n",
    "print('Test accuracyr:', score[3])\n",
    "print('Test accuracyc:', score[4])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0.9789667129516602,\n",
       " 0.494941383600235,\n",
       " 0.4840252101421356,\n",
       " 0.9017500281333923,\n",
       " 0.8622499704360962]"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = trans_model_rc.predict(x_test)[1]\n",
    "y_pred = [np.argmax(p) for p in pred]\n",
    "y_true = [np.argmax(p) for p in Y_testc]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.86225"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(y_true,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<tensorflow.python.keras.engine.input_layer.InputLayer at 0x7f0e9788eee0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e9788e1c0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917b7fa0>,\n",
       " <tensorflow.python.keras.layers.pooling.MaxPooling2D at 0x7f0e97821160>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0eb930f190>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e91796c70>,\n",
       " <tensorflow.python.keras.layers.pooling.MaxPooling2D at 0x7f0eb92def40>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e91791dc0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917a4d00>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917a47f0>,\n",
       " <tensorflow.python.keras.layers.pooling.MaxPooling2D at 0x7f0e91791bb0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917a4520>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917b09d0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917b0fa0>,\n",
       " <tensorflow.python.keras.layers.pooling.MaxPooling2D at 0x7f0e917b08e0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917388b0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e917916d0>,\n",
       " <tensorflow.python.keras.layers.convolutional.Conv2D at 0x7f0e91738430>]"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.layers[:-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [],
   "source": [
    "#plot accuracy as a function of fine-tuned model\n",
    "# visualizing  accuracy per epoch\n",
    "accuracy_231 = [0.8172500133514404,0.8261250257492065,0.8240000009536743,0.002749999985098839]\n",
    "accuracy_row = [0.8987500071525574 ,0.8981249928474426,0.9017500281333923,0.028999999165534973]\n",
    "accuracy_col = [0.8482499718666077 ,0.8482499718666077,0.8622499704360962,0.1469999998807907]\n",
    "no_of_fine_tune =range(0,4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAABKXklEQVR4nO3dd3hUVfrA8e+bTgi9BGkJnVBC7x1EmsIqFlRQ/K11VSy7CgjSBBVhFVxRcVcsWMC2ykqX3qv0gLRQAkgoAQIkpJzfH/cmTJJJMoRMJiHv53nmYe695577nrnhvrfMnCPGGJRSShVeXp4OQCmllGdpIlBKqUJOE4FSShVymgiUUqqQ00SglFKFnCYCpZQq5DQRKOVARAaLyGoPbv8ZEflTRGJFpEy6ZaEiYkTEx1PxZUcsn4nIeRHZKCIdRGSfp+NSWcu3f1BKFTYi4gu8C7Q2xmz3dDw51B7oDlQ2xly259Vxx4ZEZAxQ0xgz0B31FyZ6RVAI2Wdtuu/dLAdn7sFAALDbDeHkSA7+VkKASIckoAoAPRh4iIgME5GDInJJRPaIyN3plj8hIhEOy5va86uIyE8iEi0iZ0XkA3v+GBH5ymH9NLcRRGS5iEwQkTXAFaC6iDzmsI1DIvJUuhj6icg2Eblox9pTRO4TkS3pyr0sIr84aeMDIrI53byXRGSO/b633bZLIhIlIv/I5LMaLCKrRWSyfcvhsIj0clgeKSK3O0ynfhYOn8NjInLMXv9pEWkhIjtEJCblM0y7SflARC6IyF4R6eawoISIfCoiJ+2Yx4uIt0Oca0TkPRE5C4xx0hZ/EZkiIifs1xR7Xm0g5RZKjIgsdfZZpKsr0/0nIrtE5C6HaV8ROSMiTezp1iKy1m7/dhHp7FDW2d/KYHsbl+zP/2En8fwV+A/QRqxbW2NFpLOIHE+3r/5hf/YXRGS2iAQ4LL/T/puLseMLz6TtPYHXgAfsbW13qD+7v4VHReSo/XmMcCjrJdf/X54Vke9EpHR2++GWYIzRlwdewH1ARaxk/ABwGbjNYVkU0AIQoCbWmZY3sB14DyiKdfbY3l5nDPCVQ/2hgAF87OnlwFGgPtYtQV+gD1DD3kYnrP/0Te3yLYELWJf5XkAloC7gD5wDwhy29TvQ30kbA4FLQC2HeZuAAfb7k0AH+32plG07qWcwkAA8YX8GzwAnALGXRwK3O5RP/SwcPoeP7c/rDiAO+Bkob7frNNDJYVuJwEv2Z/SA/TmUtpf/F5huf/7lgY3AU+nWfd7+jIs4acs4YL29bjlgLfCGs33mZN30+zSr/fcqMNth3X7ATvt9JeAs0Nvet93t6XKZ/K2UAC4CdezltwH1s9hXqx2mOwPHHaYj7c+sIlAaiACetpc1sfdFK3s/P2qX989kW6n7OV392f0t/BsoAjQC4rH/loEX7H1TGevvfDrwraePFXlyPPJ0APqydwRsA/rZ7xcCLzgp0waIdnagSP+fwslBYzkwLpsYfk7Zrv2f4L1Myn0ETLDf1wfOZ/Gf9StglP2+FlZiCLSnjwJPAcWziWswcMBhOtBuWwV72pX//JUclp8FHnCY/hF40WFbqUnGnrcRGIR16yYehwM88CCwzGHdo9m05SDQ22G6B9atlAz7zMm62S133H8V7c+6uD39A/Cq/X4oMDPduguBR539rWAlvRigP06Sm5N9lV0iGOgw/Q7wscPf1Rvp6tuHnaSz+5u/gb+Fyun2bcqJSQTQzWHZbVgnIE4/71vppbeGPEREHnG4BI4BGgBl7cVVsA4Y6VUBjhhjEnO42WPpYuglIutF5JwdQ28XYgD4AnhIRATrAPmdMSY+k7LfYB0sAR4CfjbGXLGn+9vbPCIiK0SkTRaxn0p547B+UBbl0/vT4f1VJ9OOdUUZ+0hgO4J1YA3Buko46bDfpmOd3adI8xk7UdGuL33dNyyr/WeMOQGsAfqLSEmgF/C1vWoIcF9KG+x122Md+DK0w1j3+x8AnsZq+1wRqZuTmG2nHN5f4fpnHwL8PV1cVYCKIvKwfQsoVkTm38S2s9v+fx22HQEkYZ0A3NL0W0MeICIhWJen3YB1xpgkEdmGdYkP1n/CGk5WPQZUFREfJ8ngMtaZcooKTtZPPbiJiD/WmfAjwC/GmAQR+dmFGDDGrBeRa0AHrIP7Q5k0FWAxUE5EGmMlhJcc6tkE9BPr2zLPAd9h/ce/Ua60/UZUEhFxSAZVgTlYn0k8UDaLZJxdd74nsA44KQ+Eq9rzbogL+w+shP041v/zdcaYKHv+Mawrgiey2ESadhhjFgILRaQIMB7r77fDjcadjWNYV5oTMln+dbppZ5/1zfwtHAP+zxiz5gbWuSXoFYFnFMX6I44G66Ef1hVBiv8A/xCRZmKpaSePjVj31d8WkaIiEiAi7ex1tgEdRaSqiJQAhmcTgx/WfdBoIFGsh693OCz/FHhMRLrZD9EqpTsL/BL4AEgwxmT6vXtjTALwPTAJ657wYrvNfvZZXgm7zEUgOZuYM7MNGGA/EG0O3JvDelKUB4bY9d0HhAHzjDEngUXAP0WkuP251BCRTjdQ97fASBEpJyJlgVFYt89uVHb7D6xbRU2x7n1/6TD/K+AuEekhIt7231FnEansbEMiEizWFweKYiXCWHK+r7Lyb+BpEWll/90XFZE+IlIsk/J/AqGS9ltN28j538LHwAT7/xr2PuqXg3YUOJoIPMAYswf4J7AO64+5IdZlfMry74EJWLdVLmH9hy5tjEkC7sJ6eHwUOI51yY4xZjEwG9gBbAF+zSaGS8AQrLPw81hn9XMclm8EHsN6MH0BWIF1JptiJlbycuUg9g1wO/B9ujPpQUCkiFzEuu2Q4ZsoLnod6+rlPDDW3t7N2ID1POMM1n641xhz1l72CNZBeI+9vR9Ie0slO+OBzVj7aSew1Z53Q7Lbf3aZq1hXDdWAnxzmH8N6ePwaViI5BrxC5scDL+BlrCuXc1gPpp+50ZizY4zZjPWFgA+w2nQA65lDZr63/z0rIlvt9zfztzAV6zNcJCKXsB4ct7qB9QssMSa7K1mlMrJvEZzG+pbKfk/Ho5wTkVFAbaM/ulJZ0GcEKqeeATZpEsi/7O/A/xXrykupTGkiUDdMRCKxHkr+xbORqMyIyBPAFKyHwis9HI7K5/TWkFJKFXL6sFgppQq5AndrqGzZsiY0NDRH616+fJmiRYvmbkAeom3Jf26VdoC2Jb+6mbZs2bLljDGmnLNlBS4RhIaGsnnz5uwLOrF8+XI6d+6cuwF5iLYl/7lV2gHalvzqZtoiIkcyW6a3hpRSqpDTRKCUUoWcJgKllCrkCtwzAqXUrSEhIYHjx48TFxfn1u2UKFGCiIgIt24jr7jSloCAACpXroyvr6/L9WoiUEp5xPHjxylWrBihoaFYPZq7x6VLlyhWLLN+6wqW7NpijOHs2bMcP36catWquVyv3hpSSnlEXFwcZcqUcWsSKGxEhDJlytzwVZYmAqWUx2gSyH05+Uz11pBS6Rljv5IzeSU5Xe4fFw0xRx3mZVVHbpbJ/e2UP52ANcqkKgwKTyLY8gUtN7wNO1MGL3LImmky6I3Md6UsmczP6Tat+U0vXoT9xZ2Xv8m6nc8nk/nZ1ONC2fCzZ+FoyVw88N3k8mwHGXOuDVg92N8C6gGc6AcVG3s4kvwtKCiI2NhYIiMjWbt2LQ89ZA3Wt3nzZr788kvef//9TNdN+XFs2bJlMy2TVwpPIggqz6ViNQgsX946EKRyeH8j8zPrrC836nZhfuKVJAgo7qT8jdZNJvPN9ekcx+5aWZ/Ey3DNF8Tr+svLGyTdvDQvyWKZq2Vyt459f+ynTlg9D8WRi2XiL5IwuT6+S9+AgT86/1tRaURGRvLNN9+kJoLmzZvTvHlzD0d1A3I66r2nXs2aNTM5tWzZshyvm99oW/KfW6Udxhhz4IshxowubszhVW7bxp49e9xWt6OLFy9muuzw4cOmTp065tFHHzW1atUyDz30kFm8eLFp27atqVmzptmwYYMZPXq0mTRpUuo69evXN4cPHzbGGFO0aFFjjDGtWrUyxYsXN40aNTLvvvuuWbZsmenTp48xxphLly6ZwYMHmwYNGpiGDRuaH374wRhjTEhIiImOjjbGGNOvXz/TtGlTU69ePTN9+nRjjDGJiYnm0UcfNfXr1zcNGjQw7777rrl48aKZOnWqCQsLMw0bNjQPPPCA03Y5+2yBzSaT42rhuSJQSrksqlJvakQvgt/Gwl8XZX6bMJeM/d9u9py4mKt11qtYnNF31c+23IEDB/j++++ZMWMGLVq04JtvvmH16tXMmTOHN998k8aNG2dbx9tvv83kyZP59VdrhNjly5enLnvjjTcoUaIEO3fuBOD8+fMZ1p8xYwalS5fm6tWrtGjRgv79+xMZGUlUVBS7du0CICYmJnVbhw8fxt/fP3XezdJvDSmlMkj29odOr8LxjfDHAk+H41bVqlWjYcOGeHl5Ub9+fbp164aI0LBhQyIjI2+6/t9++41nn302dbpUqVIZyrz//vs0atSI1q1bc+zYMfbv30/16tU5dOgQzz//PAsWLKB4cetWcHh4OA8//DBfffUVPj65cy6vVwRKKeeaDIS1/4Il46DWHdZzGzdx5czdXfz9/VPfe3l5pU57eXmRmJiIj48PycnJqWVy+5fQy5cv57fffmPdunUEBgbSuXNn4uLiKFWqFNu3b2fhwoV8/PHHfPfdd0ydOpW5c+eycuVK/ve//zFhwgR27tx50wlBrwiUUs55+0LXEXB6D+z8wdPReExoaChbt24FYOvWrRw+fDhDmWLFinHp0iWn63fv3p1p06alTqe/NXThwgVKlSpFYGAge/fuZf1666tnZ86cITk5mf79+zN+/Hi2bt1KcnIyx44do0uXLkycOJELFy4QGxt7023URKCUyly9u6FCQ1g2ARKveToaj+jfvz/nzp2jfv36fPDBB9SuXTtDmfDwcLy9vWnUqBHvvfdemmUjR47k/PnzNGjQgEaNGrFs2bI0y3v27EliYiJhYWEMGzaM1q1bAxAVFUXnzp1p3LgxAwcO5K233iIpKYmBAwfSsGFDmjRpwpAhQyhZsuRNt1FvDSlVQCQlJxGfFM+1pGvWv8nXrr93+DflfXxSPAnJCanvHZdlV0ex+GK0T26Pj5cPdBsDX/eHrV9Ayyc8/THkqtDQ0NSHsQCff/6502WLFi1yun7K2bivry9Lly5NsyxlAJmgoCC++OKLDOs6Pn+YP3++0/pTrkRSXLp0idWrVztvzE3QRKCUCxwPwteSnR98d1/dTdKRpAzL45PiSUhKd0BOVyb9cmcH6SSTdNPt8PXyxd/bHz9vP/y8/VLf+3tZ/xbxKUKATwAbzm3gyz1f8n8N/g9qdoOQdrDiHWj8EPjdGsM+qus0Eah8Lyk5Kc1B0dWz2/TLMztDzmp5ykE60SS6FuzpzBdldhD287LeB/gEUMK/RNrlXn7O13FyEE+Z57SMtz++Xr54SfZ3g40xDPx+IB9u+5BuVbsRUjwEuo2GGXfAho+hw99d3HOqoCg0iWDB4QV8cuoTZi6c6elQcsX58+dvibYYDGfOn+GjXz9yegZ9Lema6wfhLPh4+aQ9SHqlPXAG+ARQ3L94hgNnVgdWx4O4n7cfu7fvpnWL1mnqdTz4u3IQzg9EhPtL38/E0xMZs3YMn/b4FK+qraB2L1g9FZo9BoGlPR2mykWFJhEkm2SSSSYx+eYPKvnBrdQWX/GlbJGyGQ6cqQdTb98MB/EMB+T0B2mvtPPy4iB8ee9l6pau6/bt5IUSPiX4e/O/M2bdGH7c/yP31b4Pur0OH7WDNVOh+1hPh6hyUaFJBL2r9ybwaGDqA5yCbvny5doW5Vb31LqHeYfn8e7md+lYqSPBwfWh4X2wYTq0ehqK3+bpEFUuKRjXqkqpPCcijG4zmoTkBCZsmIAxBrq8BskJsPIdT4encpEmAqVUpqoWr8qzjZ9l2bFlLDqyCEpXg2aDYeuXcO6Qp8MrkDZv3syQIUM8HUYahebWUGGQlGxISEomMdmQmJRMQpIhMTmZxCSTYV5CkjWdmLKO4/zU5YYkJ/MyWz8h+Xo9Kdt0VndikmOchri4OALWL0UEvESsnpKxzkgFIGU+di/K2GUc52U2367Hy15O6jxr2svr+nqkKZu2HrDnp9mOYxkh+nQc/z31O4IVL45lcWibXV+G+XY9ZJhnbyddfaltsCfSz0tZD2fzHdqc4TMXiD+flDoszaB6g1gQuYA3N7xJ69taU6LjK/D717DsTej/H/f8MXtASk+cXl7uPT/Oj11UF+pEkJzscPByOJA5O5imP/BZB11nB7bsD5zX18m4flYHTsdtXomLR5YvTD2YJiQnZzrMQG4TAV8vL3y8BR8vwdc75b0Xvt6Cj7cXPl6Sdp6XFwG+dlmHdby9hD9P/UlwhdLWEAhAsjHW2DHY/zkBjON8x+XXy1z/1ypLyvKU8il1YI9BQ7I12Fi6bWKXSZ3nEAuZxGcMXLmSzMlrMRm26VgmZX6y3SjHepJN2vammY9jfA7tcKg7V/cx0KntJWoHF8PHy4dxbccx4NcBTNo0ifHtx0PrZ2D1u9DuBeuXxwVUZGQkPXr0oFWrVmzZsoWWLVuyadMmRISRI0fywAMP8Oyzz9KjRw/69u3L3XffTalSpZgxYwYzZszg4MGDTJgwIU2dDRs2ZNWqVZQoUYKyZcvy3nvv8cgjj/DII48waNAgfH19U3sqXbFiBS+88AJgJeGVK1dSrFgxJk2axHfffUd8fDx33303Y8e69+F8oUkEn605zKTfLsPSBfn2wOnr7YW3l2R54ExZ5/Spk1StUhlfb8Hb4WDr4319/fR1O67v7XW9bh9vryzXSYnd2yHG3GQ9LG6cq3V6gqcfeptMElSapOSQUFISGunmX4pLoNd7y3lzXgSfP9YSgDql6/BYg8f4985/07tab9q2GwKbP4Ulb8DD39188POHwamdN1+PowoNodfb2Rbbv38/X3zxBVFRUXz88cds376dM2fO0KJFCzp27EiHDh1YtWoVffv2JSoqipMnTwKwatUqBgwYkKG+du3asWbNGkJCQqhevTqrVq3ikUceYd26dXz00Uds2rQptezkyZOZNm0a7dq1IzY2loCAABYtWsT+/fvZuHEjxhj69u3LypUr6dixY+59Num4NRGISE9gKuAN/McY83a65VWBL4CSdplhxph57oildnAx2lXyIbRqletnrDdw4Ly+jusHzpT1c/vACbB8+Tk6d/Zcj40q/0m5rWNP5bie0kX96FvDj1n7olnxRzSdapcD4KlGT7H4yGLGrR/HT31/IrDdi7BkLBxZByFtbjp+TwkJCaF169a89NJLPPjgg3h7exMcHEynTp3YtGkTHTp0YMqUKezZs4d69epx/vx5Tp48ybp165wORdmhQwdWrlxJSEgIzzzzDJ988glRUVGUKlWKokXT/iq7Xbt2vPzyyzz88MPcc889VK5cmUWLFrFo0SKaNGkCWN1Y7N+/v2AmAhHxBqYB3YHjwCYRmWOM2eNQbCTwnTHmIxGpB8wDQt0RT7uaZUk47k/nzvXcUb1St5RuIT6sO+PDhLl7aFejAz7eXvh7+zOm7RgGLxjMv37/F0NbPW99lXTJWHhs/s0NXuPCmbu7pD84p1epUiViYmJYsGABHTt25Ny5c3z33XcEBQVRrFgxpk2bxr///W8A5s2bR8eOHZk2bRpHjx5lwoQJ/Pe//+WHH36gQ4cOGeoeNmwYffr0Yd68ebRr146FCxdijGH48OE89dRTbmmvM+58KtISOGCMOWSMuQbMAvqlK2OAlIF3SwAn3BiPUspFvl7CsJ51+ePPWL7bfDx1frPgZjxQ5wG+jviaHRcOQKdX4Og62L/Yg9Hmjg4dOjB79mySkpKIjo5m5cqVtGxp3Rpr3bo1U6ZMSb1VNHny5NQD+7PPPsu2bdvYtm0bFStWpEqVKpw5cyZ1cJn27dszefJkp2f0Bw8epGHDhgwdOpQWLVqwd+9eevTowYwZM1I7tIuKiuL06Sz6LskF7kwElYBjDtPH7XmOxgADReQ41tXA826MRyl1A3o2qECL0FK8u3gfl+ISUue/2PRFygeWZ/Ta0SQ0eghKVbMGr3EYvKUguvvuuwkPD6dRo0Z07dqVd955hwoVKgBWkkhMTKRmzZo0bdqUc+fOOT3DT9GqVavU7qo7dOhAVFQU7du3z1BuypQpNGjQgPDwcHx9fenVqxd33HEHDz30EG3atKFhw4bce++9mY51kFvEuOmJqYjcC/Q0xjxuTw8CWhljnnMo87Idwz9FpA3wKdDAGJOcrq4ngScBgoODm82aNStHMcXGxhIUFJSjdfMbbUv+c6u0A6635dCFJMati+PO6r7cW9svdfmuK7uYHj2d3iV682h8IPUi3mVP2N85Hez6fewSJUpQs2ZNd4SfRlJSEt7e7htdLS+52pYDBw5w4cKFNPO6dOmyxRjj/HurmY1qf7MvoA2w0GF6ODA8XZndQBWH6UNA+azqbdasmcmpZcuW5Xjd/Ebbkv/cKu0wJm1bXpz1u6k1Yp45du5ymjKvrHjFNP6ysTlw9g9jPmxrzJRGxiRec3kbe/bsyaVos3bx4sU82U5ecLUtzj5bYLPJ5LjqzltDm4BaIlJNRPyAAcCcdGWOAt0ARCQMCACi3RiTUuoGvdKjDgJMWrgvzfxhLYcR5BvEqPVjSOo6Es4fht8Lfo+4hZHbEoExJhF4DlgIRGB9O2i3iIwTkb52sb8DT4jIduBbYLCduZRS+UTFkkV4smN1ftl2gt+PXh9vt3RAaV5t8So7oncwK+ksVGkNyyfCtSsejFblhFt/S22MmWeMqW2MqWGMmWDPG2WMmWO/32OMaWeMaWSMaWyMcT4enFLKo57qVIOyQf6MnxuB47nandXvpH2l9kz9/X1OtH8OYk/Bxk88GKnKCe10TimVrSB/H/5xR222HDnPvJ2nUueLCK+3fh2AcUfnYmp2h9XvwdUYD0WqckITgVLKJfc1r0LdCsV4e0EE8YnXx0+uGFSRF5q+wJoTa/g1rAvExcDajL+4VfmXJgKllEu8vYSRfepx7NxVvlgbmWbZgDoDaFyuMRP3f8vZ+n1h/Udw6U/PBOomY8aMYfLkyTlat3fv3sTExORuQLlIE4FSymXta5Wla93y/GvJAc7GxqfO9/byZmzbsVxJuMLbJYMg6RqsytlB81Y0b948SpYs6ekwMqWJQCl1Q17rXZcrCUlMXbI/zfzqJavzZPiTLDixmuX1e8Lmz+B8pGeCvAFffvll6i+KBw0aRGRkJF27diU8PJxu3bpx9OjRLNefNGlSaudzL730El27dgVg6dKlPPzwwwCEhoZy5swZLl++TJ8+fWjUqBENGjRg9uzZAGzZsoVOnTrRrFkzevTokdrDaV4pNN1QK6VyR83yxXioZVW+3nCUR9qEULN8sdRlf23wVxZGLuSNuBM08/ah2LK34J7p2dY5ceNE9p7bm6tx1i1dl6Eth2ZZZvfu3YwfP561a9dStmxZzp07x6OPPpr6mjFjBkOGDOHnn3/OtI4OHTrwz3/+kyFDhrB582bi4+NJSEhg1apVGfoXWrBgARUrVmTu3LkAXLhwgYSEBJ5//nl++eUXypUrx+zZsxkxYgQzZsy46c/AVXpFoJS6YS/eXotAX2/enJf24O3r7cu4tuM4E3+OKbVbwo7Z8OeeTGrxvKVLl3LfffdRtmxZAEqXLs26det46KGHABg0aBCrV6/Oso5mzZqxZcsWLl68iL+/P23atGHz5s2sWrUqQ39EDRs2ZPHixQwdOjR18Jp9+/axa9cuunfvTuPGjRk/fjzHjx/PZGvuoVcESqkbVibIn+e61uSt+XtZvf8M7WuVTV3WsFxDHg57mJl7ZtKrWCmaL30DHvw2y/qyO3PPT44dO8Zdd90FwNNPP83TTz9NtWrV+Pzzz2nbti3h4eEsW7aMAwcOEBYWlmbd2rVrs3XrVubNm8fIkSPp1q0bd999N/Xr12fdunWeaA6gVwRKqRx6tG0oVUoXYfzcPSQlp+0Q4LnGz1EpqBJjb6tI/B/z4NhGD0WZta5du/L9999z9uxZAM6dO0fbtm1J6djy66+/znBWX6VKldRup59++mmA1K6pU7qp/vjjj2nSpEnqmNEpTpw4QWBgIAMHDuSVV15h69at1KlTh+jo6NREkJCQwO7du93d9DQ0ESilciTA15thPcPYe+oSP2w5lmZZoG8go9uMJjLhIh+Xqwi/jc39gZVzQf369RkxYgSdOnWiUaNGvPzyy/zrX//is88+Izw8nJkzZzJ16tRs6+nQoQMnT56kTZs2BAcHExAQ4LSb6p07d9KyZUsaN27M2LFjGTlyJH5+fvzwww8MHTqURo0a0bhxY9auXeuO5mZKbw0ppXKsd8MKNAspxeRFf9AnvCJB/tcPKW0qtuEvNf/CZwd+4Y6oDYQdXAI1b/dgtM6lPBh2tHTp0gzlxowZk2kd3bp1IyHh+pgNf/zxR5rlkZGRAPTo0YMePXpkWL9x48asXLnyBqLOXXpFoJTKMRFhZJ8woi/FM33FwQzL/9H8H5QMKM3o4NtIXDK2wA9ec6vSRKCUuilNqpaib6OKfLLyECdirqZZVsK/BK+1eo0IH/jyymGI+MVDUaqsaCJQSt20V3vWwQCT041ZANA9pDtdq3Thw1KlOLL8DUhKTF2mvc7nvpx8ppoIlFI3rXKpQB5vX42ffo9ix/GYNMtEhBGtR+LnE8BY70uY378CICAggLNnz2oyyEXGGM6ePUtAQMANracPi5VSueKZzjX4bvMxxv8aweynWqf56mT5wPL8veVQxqwby4/r3+HeRg9QuXJljh8/TnS0ewcljIuLu+EDY37lSlsCAgKoXLnyDdWriUAplSuKBfjycvc6vPbfnSzcfYqeDW5Ls/yeWv2Zt+db/pm8lw7rphLccRjVqlVze1zLly+nSZMmbt9OXnBXW/TWkFIq19zfvDK1g4N4a/7eNGMWgHWLaHTX90jw8mZCxBcYHbwm39BEoJTKNT7eXozoU48jZ68wc92RDMurFq/Ks7XuZ1mAD4uWFJxuJW51mgiUUrmqU+1ydKpdjveX7Of85WsZlg9qPYx6Esibp1dz4VzG3x6ovKeJQCmV60b0CSM2PjHDmAUAPl4+jG33Bhe8hEmLn/NAdCo9TQRKqVxXO7gYD7asylfrj3AwOjbD8ro17uCxIiH8EnectX/oj8w8TROBUsotXupemwBfb96a53zAmae7f0BoQiLj1r/BlYQreRydcqSJQCnlFmWD/Plblxr8FvEnaw+cybDcv3Q1xgR3IsrE86+14zwQoUqhiUAp5Tb/164alUoWYfzciAxjFgA06zqBB2Lj+DpyLjuid3ggQgWaCJRSbhTg683QXnXZc/IiP211Mvxi0TK8GPYI5RMTGb1iKAlJCRnLKLfTRKCUcqu7wm+jSdWSTFq4jyvXEjMsD2r3Iq9fSuTA5eP8Z9d/PBCh0kSglHIra8yCepy+FM/0FYcyFvAvRqfWL9Mr9jKfbJ/OwRj9bUFe00SglHK7ZiGluDP8NqavPMipC3EZCzT/P4YmBFI0OZnRa0eTlJyUsYxyG00ESqk8MbRnXZKTYZKTMQvw8adMp2EMjY5me/R2Zu2blfcBFmKaCJRSeaJK6UAeax/Kj1uPsyvqQsYC4QO4s0hl2id6MXXrVE7Ensj7IAspTQRKqTzzbJealC7qx/i5ezIOSOPtg3R9nddPHoPkRMatG6eD1uQRTQRKqTxTPMCXl7rXZv2hcyze82fGAmF3UbF8I164dI01J9bw66Ff8z7IQkgTgVIqTz3Yogo1y1tjFlxLTE67UARuH82A08doFBDMxE0TOXv1rGcCLUQ0ESil8pQ1ZkEYh89c5qv1GccsoHpnvKt1Yuzxw1xJuMLEjRPzPshCxq2JQER6isg+ETkgIsMyKXO/iOwRkd0i8o0741FK5Q+da5ejQ62yTF2yn5grGccsoNtoalw6wxPFwpgfOZ8Vx1bkfZCFiNsSgYh4A9OAXkA94EERqZeuTC1gONDOGFMfeNFd8Sil8g8RYUSfMC7FJfD+kgMZC1RuBmF38fieFdQsXo1x68cRey1jd9Yqd7jziqAlcMAYc8gYcw2YBfRLV+YJYJox5jyAMea0G+NRSuUjdSsU54EWVZi5PpLDZy5nLND1dXwTLjPOpzJnrp7hvS3v5X2QhYQ7E0El4JjD9HF7nqPaQG0RWSMi60WkpxvjUUrlMy91r42ftxdvz4/IuLBcHWj0IA23/cDD1fvx3R/fsfnU5rwPshAQd31PV0TuBXoaYx63pwcBrYwxzzmU+RVIAO4HKgMrgYbGmJh0dT0JPAkQHBzcbNasnP3qMDY2lqCgoBytm99oW/KfW6UdkLdt+d/Ba/y4P4FhLQOoW9o7zTL/uNO02vAMkcEd+b9i5/HGm2EVh+Ervi7Xr/vF0qVLly3GmOZOFxpj3PIC2gALHaaHA8PTlfkYeMxhegnQIqt6mzVrZnJq2bJlOV43v9G25D+3SjuMydu2XL2WaNq8+Zvp8/5Kk5SUnLHAvKHGjClp1kZ8bxp83sBM2TLlhurX/WIBNptMjqvuvDW0CaglItVExA8YAMxJV+ZnoDOAiJTFulXkpHtCpdStKsDXm1d71mVX1EX++3tUxgId/g6+gbTZ8T/61ejHZ7s+Y+8558NfqpxxWyIwxiQCzwELgQjgO2PMbhEZJyJ97WILgbMisgdYBrxijNFfjyhVyPRtVJFGlUswaeE+rl5L1/NoUDlo8yzs+ZlXqvSgpH9JRq0ZRWJyxrENVM649XcExph5xpjaxpgaxpgJ9rxRxpg59ntjjHnZGFPPGNPQGKNdDipVCHl5CSPvrMepi3H8e5WTmwJtnoMipSmx8l1ea/UaEeci+HLPl3kf6C1Kf1mslMoXWoSWpnfDCny0/CB/Xkw3ZkFAcesW0cGldE/2p2uVrny47UOOXHTyy2R1wzQRKKXyjaE965KUbPjnIidjFrR4HIpXQpaMY0Sr1/Dz8mPsurHaQ2ku0ESglMo3QsoU5dG2IXy/5Ti7T6Qbs8A3ADoNhajNlD+2mZebv8ymU5v4cf+Pngn2FqKJQCmVrzzXtRYli/gyYW5ExrP9xg9DmZqwZBz9a/yFFhVa8O7mdzl9RTsluBmaCJRS+UqJIr68eHtt1h48y5KIdAd4bx/oOhKi9yI7v2dMmzFcS77GhPUT9BbRTdBEoJTKdx5qVZXq5Yry5vwIEpLSjVkQ1g9uawTL36RqYDDPNn6WpceWsvjIYs8EewvQRKCUynd8vb0Y0TuMQ9GX+WbD0bQLvbyg22iIOQpbPmdQvUHUK1OPCRsmcCHeyVjIKlsuJQIR+UlE+oiIJg6lVJ7oWrc87WqWYcpvf3DhSkLahTW6QmgHWDkJn4Q4xrYdy4X4C0zaNMkzwRZwrh7YPwQeAvaLyNsiUseNMSmllDVmQe96xFxN4INl+9MvtK4KLkfDho+oW7oujzV4jF8O/sLaE2s9E3AB5lIiMMb8Zox5GGgKRAK/ichaEXlM5Aa6AVRKqRtQr2Jx7mtWmc/XRnLkbLoxC6q0gDp9YM37cOUcTzd6mtDioYxbN44rCVc8E3AB5fKtHhEpAwwGHgd+B6ZiJQZ9QqOUcpu/31EHX28v3p7vpKO5riMh/hKsfg9/b39GtxlNVGwUH2z7IO8DLcBcfUbwX2AVEAjcZYzpa4yZbYx5Hrg1OvpWSuVLwcUDeLpTDebvOsWmyHPpFtaD8Adg4ydw8QTNKzTn/tr383XE1+yM3umZgAsgV68I3rc7hnvLGHPScYHJbKADpZTKJU90qE6F4gGM/3UPycnpfi/QZTgkJ8GKiQC81OwlyhUpx6i1o0hISnBSm0rP1URQT0RKpkyISCkR+Zt7QlJKqbSK+Hnzas86bD9+gTnbT6RdWCoUmj8GW2fC2YME+QXxeuvXORBzgP/s+o9H4i1oXE0ETxiH4SONNdj8E26JSCmlnPhL40o0rFSCiQv2ZhyzoOMr4OMPyyYA0KlKJ3qF9uKTHZ9w8tpJJ7UpR64mAm8RkZQJEfEG/NwTklJKZeTlJYzoE8bJC3F8ujrdmAVB5aH132DXj3ByOwBDWw6lqG9Rvj33LUnJSU5qVClcTQQLgNki0k1EugHf2vOUUirPtK5ehh71g/lw+UFOX0o3ZkHb5yGgJCx5A4AyRcowtMVQDscfZtY+HfMqK64mgqFYQ0k+Y7+WAK+6KyillMrMsF5hJCQl897iP9IuKFIS2r8EBxZD5BoA7qx+J2EBYUzdOpUTsScyVqYA139QlmyM+cgYc6/9mm6M0WstpVSeq1a2KI+0CWX2pmNEnLyYdmHLJ6HYbbBkLBiDiDCgzAAAxq0fpz2UZsLV3xHUEpEfRGSPiBxKebk7OKWUcmZI11oUdzZmgV8gdHoVjm2APxYCUNqnNC80fYE1UWv49dCvHoo4f3P11tBnwEdAItAF+BL4yl1BKaVUVkoE+jKkay1WHzjD8n3RaRc2GQSlq8OScZBsdWE9oM4AGpVrxMRNEzl79awHIs7fXE0ERYwxSwAxxhwxxowB+rgvLKWUytrA1iFUK1uU8XP3pB2zwNsXuoyA07th1w/WLC9vxrYdy5WEK0zcONFDEedfriaCeLsL6v0i8pyI3I12LaGU8iA/Hy+G96rLwejLzNqYbsyC+vdAhYawbAKSbP26uEbJGjwR/gTzI+ez4tgKD0Scf7maCF7A6mdoCNAMGAg86q6glFLKFd3rBdO6emne+20/F+McupPw8oKuo+B8JLed/C119uMNHqdmyZq8sf4NYq/FeiDi/CnbRGD/eOwBY0ysMea4MeYxY0x/Y8z6PIhPKaUyJSKM7FOP81euMW3ZgbQLa3WHqm0JOTIbrlldWPt6+zK27VhOXznNlK1T8j7gfCrbRGB/TbR9HsSilFI3rEGlEvRvWpnPVkdy7JzDOAQicPto/K+dhw3TU2eHlwtnYL2BzN43my1/bvFAxPmPq7eGfheROSIySETuSXm5NTKllHLRP+6og7eX8PaCdGMWVG3N2dLNYc0UuHo+dfZzjZ+jUlAlxqwdQ3xSfN4Gmw+5mggCgLNAV+Au+3Wnu4JSSqkbUaFEAE92rM7cHSfZciTtmAWHqg+CuIuwZmrqvEDfQEa3GU3kxUg+3v5xXoeb77j6y+LHnLz+z93BKaWUq57qVJ3yxfx549e0PzK7HBQKDe+F9R/DpVOp89tUbEO/Gv34bNdn7D3nZPSzQsTVXxZ/JiIz0r/cHZxSSrkq0M+HV3rUYduxGP63I13X011eg+QEWDkpzexXWrxCSf+SjFozisTkxDyMNn9x9dbQr8Bc+7UEKA7od6+UUvlK/6aVqV+xOBPn7yUuwaE7tNLVoemjsOVzOHe9d5wS/iUY3mo4EecimLlnZt4HnE+4emvoR4fX18D9gA5RqZTKV1LGLIiKucqMNYfTLuz0Knj5wrK30sy+I+QOulTpwrRt0zh6Md0P0woJV68I0qsFlM/NQJRSKje0rVGW28OC+XDZQaIvOXwjqFgFaPUU7PweTu1KnS0ijGw9Ej8vP8asG1Moeyh19RnBJRG5mPIC/oc1RoFSSuU7w3vXJS4hifd+SzdmQfsXIaA4LH0jzezygeV5ufnLbDq1iR/3/5h3geYTrt4aKmaMKe7wqm2MKXyfllKqQKhRLoiBrUOYtfEoxy85dEhXpBS0ewH+WABH03aO0L9Wf1pUaMG7m9/l9JXTeRyxZ7l6RXC3iJRwmC4pIn9xYb2eIrJPRA6IyLAsyvUXESMi+txBKZUrXuhWiyB/H2bvu5Z2QaunoWh5+M0avCaFiDCmzRiuJV9jwvoJheoWkavPCEYbYy6kTBhjYoDRWa1g91E0DegF1AMeFJF6TsoVw+rUboOLsSilVLZKFfVjSLda7DyTxPJ9Dmf4fkWtB8dH18KB39KsU7V4Vf7W+G8sPbaUxUcW53HEnuNqInBWziebdVoCB4wxh4wx14BZQD8n5d4AJgJxTpYppVSOPdImlOBA4c15ESQ6jlnQ9FEoGWINaZmcnHadeo8QVjqMNze8yYX4CxQGriaCzSLyrojUsF/vAtn11lQJOOYwfdyel0pEmgJVjDFzXY5YKaVc5OfjxX21/fjjz1hmb3Y4HPn4WYPXnNoJu39Ks46Plw/j2o0jJj6GyZsn53HEnpHdWX2K54HXgdmAARYDz97Mhu2Bbt4FBrtQ9kngSYDg4GCWL1+eo23GxsbmeN38RtuS/9wq7YBbqy11isZRu5Q3b8/dRelLhyjiI9YCU5bmRUPwmjeSTdElMV5pD4ddi3Xl5wM/U+liJeoWqeuByDNy234xxrjlBbQBFjpMDweGO0yXAM4AkfYrDjgBNM+q3mbNmpmcWrZsWY7XzW+0LfnPrdIOY269tmw/dt6EDP3VvD0/Iu3CvfOMGV3cmE0zMqwXlxhn7vzpTtPjhx7m8rXLeRRt1m5mvwCbTSbHVVe/NbRYREo6TJcSkYXZrLYJqCUi1UTEDxgAzHFIQBeMMWWNMaHGmFBgPdDXGLPZlZiUUspV4ZVLck+TSny6+jDHzzuMWVC7J1RuCSsmQsLVNOv4e/szus1oomKj+GDbB3kccd5y9RlBWWN9UwgAY8x5svllsTEmEXgOWAhEAN8ZY3aLyDgR6ZvDeJVSKkf+0aMOXgLvLNh3faYI3D4GLp2EjZ9kWKd5hebcX/t+vo74mp3RO/Mu2DzmaiJIFpGqKRMiEor1rCBLxph5xvrxWQ1jzAR73ihjzBwnZTvr1YBSyl0qlizCkx2qM2f7CbYevT5IDaHtoObtsOpduBqTYb2Xmr1E2SJlGbV2FAlJCRmW3wpcTQQjgNUiMlNEvgJWYN3zV0qpAuOpTjUoV8yf8b/uSfuDsW6jIC4G1v4rwzpBfkG83vp1DsQc4NNdn+ZdsHnI1S4mFmD1NroP+Bb4O3A1y5WUUiqfKervwz/uqM3WozHM3ekwZsFtjaD+PbD+I4jN2L1E5yqd6RXai092fMLBmIN5GHHecPVh8eNY4xD8HfgHMBMY476wlFLKPe5tVoW6FYrxdvoxC7qOhMQ4WOn8twNDWw61hrhcO5qk5CSnZQoqV28NvQC0AI4YY7oATYAYdwWllFLu4u0ljOxTj+Pnr/LF2sjrC8rUgKaDYPMMOB+ZYb0yRcowtMVQtkdvZ9a+WXkWb15wNRHEGWPiAETE3xizF6jjvrCUUsp92tcqS7e65flg6QHOxjqMWdBpKHh5w/K3na53Z/U7aVexHVO3TuVE7Ik8itb9XE0Ex+3fEfwMLBaRX4Aj7gpKKaXcbXjvMK4kJDHlt/3XZxavCC2fgO2z4M89GdYREUa1GQXAuPXjbpkeSl19WHy3MSbGGDMGq6uJT4G/uDEupZRyq5rlg3i4VVW+2XiU/X9eur6g/cvgXwyWjne6XsWgirzQ9AXWRK3h10O/5lG07nXDQ1UaY1YYY+YYq0dRpZQqsF7oVotAP2/enBdxfWZgaWg7BPbNhWObnK43oM4AwsuF886mdzh79WweRes+OR2zWCmlCrwyQf4837Umy/ZFs2p/9PUFrZ+BouWsbqqd3P7x9vJmXNtxXE64zMSNE/MwYvfQRKCUKtQebRtKldJFmDA3gqRk+6DvHwQdX4HIVXBwqdP1apSswRPhTzA/cj4rjq3Iw4hznyYCpVSh5u/jzfBeYew9dYnvHccsaDYYSlSFJeOcXhUAPN7gcWqWrMkb698g9lps3gTsBpoIlFKFXq8GFWgeUorJi/4gNj7RmunjD12Gw8ltsOcXp+v5evsytu1YTl85zZStU/Is3tymiUApVeiJCCPvrMeZ2Hg+Xu7QhUT4A1CurvUNoqREp+uGlwvn4bCHmb1vNlv+zG7gxvxJE4FSSgGNq5SkX+OK/HvVIaJi7K7UvLyh6+twdj9s/ybTdZ9v8jyVgioxZu0Y4pPiMy2XX2kiUEop26s9rSEpJy3Ye31m3T5Qqbn1a+OEOKfrBfoGMqr1KCIvRjJ9+/S8CDVXaSJQSilbpZJFeLxDNX7edoLtx2KsmSJw+2i4GAWb/pPpum0rtaVfjX7M2DWDvef2ZlouP9JEoJRSDp7pXJOyQX6Mn+swZkG1jlC9C6z6J8RdzHTdV1q8Qkn/koxaM4rEZOfPFPIjTQRKKeUgyN+Hl7vXYVPkeRbsOnV9QbdRcPUcrJuW6bol/EswvNVwIs5FMHPPzDyINndoIlBKqXTub16ZOsHFeGv+XuIT7bEHKjWFev1g3Qdw+Uym694RcgddqnRh2rZpHL14NI8ivjmaCJRSKh0fby9G9Anj6LkrfLnWoaPlLiMh4Yp1iygTIsKIViPw9fJl7LqxBaKHUk0ESinlRMfa5ehcpxzvL93Puct2H5vlakPjh6yHxjHHMl03uGgwLzd/mY2nNvLT/p/yKOKc00SglFKZGNE7jCvXknh/icOYBZ2GAZLp4DUp+tfqT/Pg5vxz8z85fSXjOMj5iSYCpZTKRK3gYjzYsgoz1x/hwGm7L6GSVaDF49YPzKL3Zbqul3gxpu0YriVfY8L6Cfn6FpEmAqWUysKLt9emiK83b893GLOgw8vgG5jp4DUpQoqH8LfGf2PpsaUsPrLYzZHmnCYCpZTKQtkgf57tUpPfIk6z5oD9baGiZaHt8xAxB6Ky7l/okXqPEFY6jDc3vMmF+At5EPGN00SglFLZeKxdKJVKFmG845gFbZ6FwDJWN9VZ8PHyYVy7ccTExzB58+Q8iPbGaSJQSqlsBPh6M6xXXSJOXuTHLcetmf7FoMPf4dBy65WFuqXrMrj+YH4+8DPrTqxze7w3ShOBUkq54M7w22hStSSTF+3jcsqYBc3/CsUrZzl4TYqnGz1NSPEQxq4by5WEK3kQses0ESillAtEhNfvrMfpS/FMX3nImukbAJ2HWc8J9v6a5foBPgGMaTOGqNgoPtj2QR5E7DpNBEop5aKmVUtxZ/htfLLyICcv2GMWNHoQytaGJW9AclKW6zev0Jz7a9/P1xFfszN6Zx5E7BpNBEopdQOG9qxLsoFJC+3fEHj7QNeRcGYfbJ+V7fovNXuJskXKMmrtKBKSEtwcrWs0ESil1A2oUjqQ/2tXjZ+2RrHjeIw1M6wvVGwCy9+CxKxHKAvyC+L11q9zIOYAn+761P0Bu0ATgVJK3aC/dalBmaJ+jJ8bYf1iWMTqpvrCMdg8I9v1O1fpTM/Qnnyy4xMOxRzKg4izpolAKaVuUPEAX17qXpuNh8+xaM+f1szqXawBbFZOhvhL2dYxrOUwAn0DGb12NMkm2c0RZ00TgVJK5cCAFlWoVT6It+ZFcC0x2b4qGA1XzsD6j7Jdv0yRMrza4lW2RW9j1t7sny24k1sTgYj0FJF9InJARIY5Wf6yiOwRkR0iskREQtwZj1JK5RYfby9e6xNG5NkrzFxvj1lQuTnUvRPWvA+Xz2Zbx13V76JdxXZM2TqFE7En3Bxx5tyWCETEG5gG9ALqAQ+KSL10xX4HmhtjwoEfgHfcFY9SSuW2zrXL0aFWWd5fsp+YK/aYBV1HQsJlWP1utuuLCKPajAJg3PpxHuuh1J1XBC2BA8aYQ8aYa8AsoJ9jAWPMMmNMyk/s1gOV3RiPUkrlKhFhRJ8wLsUlMDVlzILyYRA+ADb+Gy5EZVtHxaCKvND0BdZEreHXQ1n/KM1d3JkIKgGOQ/gct+dl5q/AfDfGo5RSua5uheI80KIqM9cd4VC0PWZB52FgkmFF1oPXpBhQZwDh5cJ5Z9M7nIs758ZonRN3XYqIyL1AT2PM4/b0IKCVMeY5J2UHAs8BnYwxGb6EKyJPAk8CBAcHN5s1K2cPVmJjYwkKCsrRuvmNtiX/uVXaAdqWG3Uh3jB05RXqlfFmSNMAAGru/zeVouaxseUHXA3M6hzYcvLaSSaenEjjwMYMLjfYaZmbaUuXLl22GGOaO11ojHHLC2gDLHSYHg4Md1LudiACKO9Kvc2aNTM5tWzZshyvm99oW/KfW6UdxmhbcuKDpftNyNBfzdoDZ6wZl04bM/42Y7571OU6Pvz9Q9Pg8wZmxbEVTpffTFuAzSaT46o7bw1tAmqJSDUR8QMGAHMcC4hIE2A60NcYk78H9VRKqSz8tX01e8yCPSQnGwgqB23+Brv/Cye2uVTH4w0fp2bJmoxbN47Ya7HuDdiB2xKBMSYR63bPQqwz/u+MMbtFZJyI9LWLTQKCgO9FZJuIzMmkOqWUytcCfL15tWcddp+4yE+/2w+J2z4PRUplO3hNCl9vX8a2HcvpK6eZsnWK+4JNx62/IzDGzDPG1DbG1DDGTLDnjTLGzLHf326MCTbGNLZffbOuUSml8q+7wivSqEpJJi3cy5VriRBQAtq/DAeXwOFVLtURXi6ch8MeZva+2Wz5M+thMHOL/rJYKaVyiZeX8HqfMP68GM8nKWMWtHwCilWEJWOzHbwmxfNNnqdSUCXGrB1DfFLWndjlBk0ESimVi5qHlqZPw9uYvuIQf16MA98i0HkoHN8E+1z7hnygbyCjWo8i8mIk07dPd3PEmgiUUirXDe1Zl6Rkw+SUMQsaD4TSNWBp9oPXpGhbqS19a/Tls12fse/cPjdGq4lAKaVyXdUygQxuF8oPW4+zK+qCPXjNCDi9B3Z+73I9r7Z4leL+xRm1dhSJyYlui1cTgVJKucGzXWpSsogvE1LGLKh3N1QIh2VvQuI1l+oo4V+C4a2Gs+fsHmbumem2WDURKKWUG5QoYo1ZsO7QWX6LOA1eXlY31TFHYOsXLtfTI6QHXap0Ydq2aUQnRLslVk0ESinlJg+2rEqNckV5a14ECUnJULMbhLSHFe/Atcsu1SEijGg1ggCfAPbH7XdLnJoIlFLKTXy9vRjRJ4xDZy7z9foj1uA1t4+Gy6ddGrwmRXDRYObdM4+2xdq6JU5NBEop5UZd6pSnfc2yTFmynwtXEqBKS6jdyxq85orrPY0W9yvuthg1ESillBuJCK/1DuPC1QT+tdS+tdPtdYi/CGumeDS2FJoIlFLKzepVLM79zarwxbpIIs9chuD6EH4/bJgOF096OjxNBEoplRf+fkdtfL29eHv+XmtG5+GQnAgrPT9CryYCpZTKA+WLB/BMpxos2H2KDYfOQulq0Owx2PolnD3o0dg0ESilVB55vEN1bisRwIR5EdaYBR1fAW8/60dmHqSJQCml8kgRP2vMgh3HL/DL9igoFgytnoZdP8DJHR6LSxOBUkrloX6NKhFeuQTvLNjH1WtJ0O4FCChpdUjnIZoIlFIqD3l5CSN6h3HyQhz/WXUIipSE9i/C/kVwZJ1nYvLIVpVSqhBrVb0MPetX4KMVBzl9MQ5aPgVBFeC3MS4PXpObNBEopZQHDOtVl4SkZP656A/wC4ROr8Kx9daVQR7TRKCUUh4QWrYoj7YJ5bstx9hz4iI0fQRKVbMGuk9OztNYNBEopZSHPN+1FiWK+DJh3h6Mlw90GQF/7oJdP+ZpHJoIlFLKQ0oE+vJCt1qsOXCWZftOQ4P+ENwAlk2ApIQ8i0MTgVJKedDA1iFUL1uUCXMjSDBAt1Fw/rD1i+M8oolAKaU8yNfbi+G9wzgYfZlvNx6FWndAldb24DVX8iQGTQRKKeVht4eVp031Mry3+A8uxCXC7WMg9hRsnJ4n29dEoJRSHiYijOgTRszVBD5cdgBC2lhXBqvfg6vn3b59TQRKKZUPNKhUgnubVuazNZEcPXsFur4OcReskczcTBOBUkrlE//oUQdvL2Higr1wWzg0uBc2fAyX/nTrdjURKKVUPhFcPICnOlVn7s6TbI48B11eg6RrsHKSW7eriUAppfKRJztWJ7i4P2/MjSC5VHVoMgi2fAbnDrttm5oIlFIqHwn08+GVHnXZfiyG/+04AZ2GgpcPLH/LbdvURKCUUvnMPU0q0aBScd5ZsI+4IuWh1VOw4zuKxka6ZXuaCJRSKp+xxiyoR1TMVT5dfRjavQhB5QmKPeSe7bmlVqWUUjelTY0ydK8XzIfLDhCdVBRe2MGfFbq6ZVuaCJRSKp8a3qsu8YnJvLv4D/ANcNt23JoIRKSniOwTkQMiMszJcn8RmW0v3yAioe6MRymlCpLq5YIY1CaE2ZuOsvfURbdtx22JQES8gWlAL6Ae8KCI1EtX7K/AeWNMTeA9YKK74lFKqYLohW61KBbgy4S5EW7bhjuvCFoCB4wxh4wx14BZQL90ZfoBX9jvfwC6iYi4MSallCpQSgb6MaRbLVbtP8OO6ES3bMOdiaAScMxh+rg9z2kZY0wicAEo48aYlFKqwBnUOoQudcrh4+We82QxxrinYpF7gZ7GmMft6UFAK2PMcw5ldtlljtvTB+0yZ9LV9STwJEBwcHCzWbNm5Sim2NhYgoKCcrRufqNtyX9ulXaAtiW/upm2dOnSZYsxprmzZT43FVXWooAqDtOV7XnOyhwXER+gBHA2fUXGmE+ATwCaN29uOnfunKOAli9fTk7XzW+0LfnPrdIO0LbkV+5qiztvDW0CaolINRHxAwYAc9KVmQM8ar+/F1hq3HWJopRSyim3XREYYxJF5DlgIeANzDDG7BaRccBmY8wc4FNgpogcAM5hJQullFJ5yJ23hjDGzAPmpZs3yuF9HHCfO2NQSimVNf1lsVJKFXKaCJRSqpDTRKCUUoWcJgKllCrk3PaDMncRkWjgSA5XLwucybZUwaBtyX9ulXaAtiW/upm2hBhjyjlbUOASwc0Qkc2Z/bKuoNG25D+3SjtA25JfuastemtIKaUKOU0ESilVyBW2RPCJpwPIRdqW/OdWaQdoW/Irt7SlUD0jUEoplVFhuyJQSimVjiYCpZQq5G7JRCAiPUVkn4gcEJFhTpb7i8hse/kGEQn1QJgucaEtg0UkWkS22a/HPRFndkRkhoictgcjcrZcROR9u507RKRpXsfoKhfa0llELjjsk1HOynmaiFQRkWUiskdEdovIC07KFIj94mJbCsp+CRCRjSKy3W7LWCdlcvcYZoy5pV5YXV4fBKoDfsB2oF66Mn8DPrbfDwBmezrum2jLYOADT8fqQls6Ak2BXZks7w3MBwRoDWzwdMw30ZbOwK+ejtOFdtwGNLXfFwP+cPL3VSD2i4ttKSj7RYAg+70vsAFona5Mrh7DbsUrgpbAAWPMIWPMNWAW0C9dmX7AF/b7H4BuIuKewUBvjittKRCMMSuxxpzITD/gS2NZD5QUkdvyJrob40JbCgRjzEljzFb7/SUggozjiheI/eJiWwoE+7OOtSd97Vf6b/Xk6jHsVkwElYBjDtPHyfgHkVrGGJMIXADK5El0N8aVtgD0ty/bfxCRKk6WFwSutrWgaGNf2s8XkfqeDiY79q2FJlhnn44K3H7Joi1QQPaLiHiLyDbgNLDYGJPpfsmNY9itmAgKm/8BocaYcGAx188SlOdsxerXpRHwL+Bnz4aTNREJAn4EXjTGXPR0PDcjm7YUmP1ijEkyxjTGGuu9pYg0cOf2bsVEEAU4nhVXtuc5LSMiPkAJ4GyeRHdjsm2LMeasMSbenvwP0CyPYsttruy3AsEYczHl0t5Yo/T5ikhZD4fllIj4Yh04vzbG/OSkSIHZL9m1pSDtlxTGmBhgGdAz3aJcPYbdiolgE1BLRKqJiB/Wg5Q56crMAR61398LLDX2U5d8Jtu2pLtf2xfr3mhBNAd4xP6WSmvggjHmpKeDygkRqZByv1ZEWmL9P8t3Jxp2jJ8CEcaYdzMpViD2iyttKUD7pZyIlLTfFwG6A3vTFcvVY5hbxyz2BGNMoog8ByzE+tbNDGPMbhEZB2w2xszB+oOZKSIHsB76DfBcxJlzsS1DRKQvkIjVlsEeCzgLIvIt1rc2yorIcWA01kMwjDEfY41t3Rs4AFwBHvNMpNlzoS33As+ISCJwFRiQT0802gGDgJ32/WiA14CqUOD2iyttKSj75TbgCxHxxkpW3xljfnXnMUy7mFBKqULuVrw1pJRS6gZoIlBKqUJOE4FSShVymgiUUqqQ00SglFKFnCYCdUuyv4u9QUR+F5EO6ZYtFxGPD2YuIt/aXYO8JCLjROT2XKp3sIhUzI26XNjOB+7ejnK/W+53BErZugE7jTF53i23iHgbY5KyKVMBaGGMqemGEAYDu4ATbqhb3YL0ikC5hYiEikiEiPzb7lN9kf0ryTRn5CJSVkQi7feDReRnEVksIpEi8pyIvGyf1a8XkdKZbGepfWa9RESqikhj4B2gn1j9zhfJIs6PRGSzOPT7LiJdReRnhzLdReS/9vs7RGSdiGwVke/tvm2w450oIluB+0RkiFh94+8QkVlONr0IqGTH10FEPheRex3qGmtvY6eI1LXnFxVrLISN9meSoSdau47mwNcpbbfrK2svby4iy+33Y+z6lovIIREZ4lDPQHs720Rkuv3jJkTkMRH5Q0Q2Yv2IS90CNBEod6oFTDPG1AdigP4urNMAuAdoAUwArhhjmgDrgEeclP8X8IXd6d7XwPvGmG3AKKw+2hsbY65msb0RxpjmQDjQSUTCsfp2qSsi5ewyjwEz7IPpSOB2Y0xTYDPwskNdZ40xTY0xs4BhQBM7rqedbLcvcNCOb5WT5WfsbXwE/CMlVqyuBFoCXYBJIlLUcSVjzA92XA+70HaAukAPrC7PR4uIr4iEAQ8A7eyOz5KAh8XqzmQsVgJoD9TLpm5VQGgiUO502D4oA2wBQl1YZ5kx5pIxJhqra93/2fN3ZrJ+G+Ab+/1MrAPUjbjfPov/HaiPNZiJsesaKFafL22wBmdpjXXwW2N3Y/AoEOJQ12yH9zuwzsoHYnX/caNSOk1z/NzuAIbZ214OBGB3oXAT5hpj4o0xZ7C6PA7Guq3WDNhkb6sb1uBIrYDlxphoe3yM2ZnUqQoYfUag3Cne4X0SkHKLJpHrJyEBWayT7DCdTC7/vYpINayz7RbGmPMi8rlDPJ9hJaE44Hu73yfB6hv+wUyqvOzwvg/WSGZ3ASNEpKHdb7yrUtqdxPV2C9DfGLMvXTs+w+p//4QxpreTulz9vFO2JVhXWcPTbecvNxC/KkD0ikB5QiTXu8u+9ybrWsv1DrceBpzdZslMcayD9wURCQZ6pSwwxpzAetg6EispAKwH2olITUi9Z187faUi4gVUMcYsA4ZidREcdCONysRC4Hk7ISEiTexYH7NvA6UkgUtYwzWmiOT65+3K7bklwL0iUt7eTmkRCcEa6KWTiJQRq8vn+262QSp/0ESgPGEyVi+QvwM32x/888BjIrIDq/fJDIOWZ8YYsx3rltBerNtLa9IV+Ro4ZoyJsMtHY30j51t7e+uw7rGn5w18JSI77frft/uVv1lvYPVyukNEdtvTznwOfOzwoHwsMFVENmOd9WfJGLMHKwEustu5GLjN7n56DFa711BwuzxX6Wjvo0plQqzvyP9ujPnU07Eo5U6aCJRyQkS2YN026u4wApxStyRNBEopVcjpMwKllCrkNBEopVQhp4lAKaUKOU0ESilVyGkiUEqpQu7/Acm2ackJs9KKAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(no_of_fine_tune, accuracy_231)\n",
    "plt.plot(no_of_fine_tune, accuracy_row)\n",
    "plt.plot(no_of_fine_tune, accuracy_col)\n",
    "plt.xlabel('num of layers fine-tuned')\n",
    "plt.ylabel('accuracy')\n",
    "plt.title('accuracy vs number of layers fine-tune')\n",
    "plt.grid(True)\n",
    "plt.legend(['multiclass', 'row-wise','col-wise'], loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "range(0, 3)"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "no_of_fine_tune"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = range(1,35)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "range(1, 35)\n"
     ]
    }
   ],
   "source": [
    "print(epochs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "#summery of erros\n",
    "A0= 0.8172500133514404 # accuracy of a model by freexing all layers of the VGGnet\n",
    "A1=0.8261250257492065  #freeze all except the last one conv block\n",
    "A2=0.8240000009536743  \n",
    "A3=0.002749999985098839"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# #error\n",
    "# E0=1-A0\n",
    "# E1=1-A1\n",
    "# E2=1-A2\n",
    "# E3=1-A3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "err0p=A0+2 * np.sqrt(A0*(1-A0)/8_000)\n",
    "err1p=A1+2 * np.sqrt(A1*(1-A1)/8_000)\n",
    "err2p=A2+2 * np.sqrt(A2*(1-A2)/8_000)\n",
    "err3p=A3+2 * np.sqrt(A3*(1-A3)/8_000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "err0n=A0-2 * np.sqrt(A0*(1-A0)/8_000)\n",
    "err1n=A1-2 * np.sqrt(A1*(1-A1)/8_000)\n",
    "err2n=A2-2 * np.sqrt(A2*(1-A2)/8_000)\n",
    "err3n=A3-2 * np.sqrt(A3*(1-A3)/8_000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "errp=[err0p,err1p,err2p, err3p]\n",
    "errn=[err0n,err1n,err2n, err3n]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "A_model_0=acceach0\n",
    "A_model_1=acceach1\n",
    "A_model_2=acceach2\n",
    "A_model_3=acceach0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "A_model_0mean=np.mean(acceach)\n",
    "A_model_1mean=np.mean(acceach1)\n",
    "A_model_2mean=np.mean(acceach2)\n",
    "A_model_3mean=np.mean(acceach0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [],
   "source": [
    "A_model_0std=np.std(acceach)\n",
    "A_model_1std=np.std(acceach1)\n",
    "A_model_2std=np.std(acceach2)\n",
    "A_model_3std=np.std(acceach0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [],
   "source": [
    "#  labels, positions, barheights, standard erro(SE) and error bar heights\n",
    "labels = ['Fine-tune-o', 'Fine-tune-1', 'Fine-tune-2','Fine-tune-3']\n",
    "x_positon = np.arange(len(labels))\n",
    "b_height = [A_model_0mean, A_model_1mean, A_model_2mean,A_model_3mean]\n",
    "SE=[2*np.sqrt(A_model_0std/8000),2*A_model_1std/np.sqrt(8000),2*A_model_2std/np.sqrt(8000),2*A_model_3std/np.sqrt(8000)]\n",
    "error = [A_model_0std, A_model_1std, A_model_2std, A_model_3std]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8206448401558067"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "A_model_0mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8096957250036113"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "A_model_1mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.22323740797875527"
      ]
     },
     "execution_count": 162,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "2*A_model_0std"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0024193685021718426"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "2*A_model_0std/np.sqrt(8000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAAk50lEQVR4nO3deZgcZbn+8e9tANmCQdFRkkBQAxrxJ+gQRM+RUZATVBI5IoRFCS4RNYDiBi4QcQMXEDWKEQUEJSwKRoyEAA4eFDRBApqwGAOYEBGJQcKSQMLz+6PeCUXT01OT7pqpmbk/19XXdFW9VfXUM939VL1VXa2IwMzMrGqe1d8BmJmZ1eMCZWZmleQCZWZmleQCZWZmleQCZWZmleQCZWZmlTRkCpSkuyXt20/r7pC0vI/W1Sbpt5JWS/pGX6yzFSSNkRSSNknDnZLet5HLOlDSMkkPS9pd0iJJHa2MtxUkfVrS2f0dR08kHS7pqj5eZ7+9X6tK0rmSvtjfcfSlIVOgBooWvAinAg8A20TEx3qx3udK+pek65tYd1V8HZgWEVtHxM0R8YqI6Gx2oZKmS7qg+fAyEfHliNioItyXIuInEbFff8cx0PTmvSxp77SD1nQBkvRFSdfUjNtZ0kOSXilpiqT1aQcu/9g+tb1b0mNp3H1pO7bOLWuUpJ9JekDSfyT9RdKUZuOuxwWql5Spct52BBZH77+BfRpwWwnx9IcdgUX9HcRg0HVEOxT017ZK2hQ4E/hDixb5BeCFkt6fli/gB8DpEfHn1OaGtAOXf6zILeOAiNga2A3YHTgxN+18YBnZ++x5wLuAf7Yo9qep8gdtGfaQtFjSKknnSNocQNK2kq5IRxCr0vNRXTOl7qYvSfod8Cjw4toFp72OE+stv07bl6dlPpi6nyam8VOBw4FPpr2XX3Yz/+skzU97L/MlvS6NPxc4Mjd/oS6SNP+uwDk9tJsi6XeSzkixL02xTEldavdLOjLX/jmSfpzyeo+kz3YVd0nDJH097YUtBd7aw7rfI+m2lNu5knas0+bZkh4GhgG3SPpbGr+huygdBV2c4lqd8t+eW8b2ae/wX5LuknRsGj8B+DRwSMrtLbXLzi3/gvS8q9vySEl/T9v6mY1su4Wk89L23ybpk+qm21jS9yR9vWbcLyQdn56fIOlvafsXSzqwm//xSmB6Gnd9rk3d11+BfGwu6QJJK9PrZ76ktnrbUBP7eEk3pHn+Iek7kjZL02aopitb0mxJH03P6/4/c7FdmmJ6CJhSZ90h6aW54Q1HRUpd98q6ah9I2354mlbovZx8DLgKuL2nXOTi2FbdfGZFxFrgPcCpyo6KpgLbAl8quvwuEXEfMJesUHXZAzg3Ih6JiHWpl+LXvV120QCGxAO4G/gLMBp4LvA74Itp2vOAdwBbAsOBS4DLc/N2An8HXgFsAmzay+V3AMvT802BJWQfdpsBbwJWA7uk6ed2zdfNdjwXWEW217IJcGgafl6R+essbxjwJ+A1ZG/Q6xu0nQKsA45K830x5WUG8Gxgv7QtW6f2PwZ+kXI6BrgTeG+adjTZG7IrX78BAtgkl/P3peeTUs5enrb5s8DvG8QZwEtr/jf7pufTgTXAW9I2fAW4MU17FnATcFL637wYWAr8T27eC+r83/fNDW9ok7Y5yPZetwBeBawFXr4RbU8FriP7oBkF3Ep6TdXZ/jeQ7eEqDW8LPAZsn4bfCWyftvcQ4BHgRTX/42NSrrcg97qg59dfo3x8APgl2ftsGNlrbpsG79eu/9lrgNem9Y0hO9L/SJo2HlgBPCsNb0e2E9lW8P/5BPD21HaLAq+lc3n6+3odcDrZ63/vlMtC7+XUZkey98XWPbWvWXfDz6zU5hvANWRd/u017+NG7/N87kcBfwbOzE2/muzzbTKww8Z8Hhd9DLUjqO9ExLKI+DfZ3sShABGxMiJ+FhGPRsTqNG3vmnnPjYhFke0xPNGb5dd4LdmL8dSIeDwirgWu6KZtPW8F/hoR56dYLiT7oD+g4Py1jgX+EBE3FWx/V0ScExHrgYvICswpEbE2Iq4CHgdeKmkY2Qv4xIhYHRF3k71h3pWWczDwzVy+vtJgnUcDX4mI2yJiHfBlYLd6R1EFXR8Rc9I2nE9WDCDbM3x+RJyS/jdLyQrG5I1cT5fPR8RjEXELcEtufb1pezDw5YhYFRHLgW81WMb/kX2w/ncaPoisS2cFQERcEhErIuLJiLgI+CvZB32XFRHx7fT6eqxm2c28/p4g+2B9aUSsj4ibIuKhnmZK7W5M67sb+D7p/RkRfwT+A+yTmk8GOiPinxT7f94QEZenXNRua1GfS6//64Bfkf2vivpWmv/h3qyw4GfWZ4GXAudHxIKaaa9NR6Rdj7/VTL9c0mqyHZ37gZNz095J9hr7HHCXpIWS9uhN/EUNtQK1LPf8HrK9SCRtKen7yrqhHgJ+C4xIH7L15u3V8mtsDyyLiCdr2o4ssgFp/ntqxvVm/g3S4f+xwGd6apuT72t+DCB9GOTHbU22J7tpTaz5OLfnmfnqzo7AmV1vJuDfgNiIbU7uyz1/FNhc2fmHHYHt829csiPdHruherm+rbtr2KBtbb66fT1Gtps7i6d2eg4DftI1XdK704dK1zbuSvb/6nHZNPf6O5+su2iWpBWSvqrs/EtDyk7wX6HshP1DZDso+XjPA45Iz49I64Fi/88N2yrp13rqgoHDC2wPwKqIeCQ33N37vt52HQAMTzsJvVLkMysV3Luofz72xogYkXu8pGb62yNiONlR4svI5TvtJJ0QEa8gy+VCsoKm3m5HT4ZagRqde74DWdcAZH3AuwB7RsQ2ZF0kkH0Idily0UF3y89bAYzW0y+02AG4t+B6VpC98fLy8/fGeOBFwGJJ95GdqB2fPgiGNZ61Rw+Q7THnY83H+Q+ema/uLAM+UPOG2iIift9kjPXWc1fNeoZHxFvS9Hr/m0fIulm6vLDFMXX5B1l3S5fR3TVMLgQOSkeZewI/A0jDPwCmkXXLjSDrmi76Wu/p9ddtPiLiiYj4fESMA14HvA14dw/bAfA9sqO0sen9+emaeC8AJkl6FVk38OVpfE//T8hta0TsH09dMNBV0B/tbnuSbSVtlRvOv+97ei/vA7Sn99t9ZN2tH5H0ix7mg2KfWU1LR4Xnkl0ZW2/6A2na9mTdvy011ArUh5VdIvlcsqOGrj2X4WR7/g+maSd3t4CNXH7eH8he9J+UtKmy7+ccQLbHC9kRyjMuwsiZA+ws6TBJm0g6BBhH1k3YW78m69PfLT1OAm4GdkvdXxstzX8x8CVJw9MH4/FkHyakacemfG0LnNBgcWcBJ0p6BWy4+OKdzcTXjT8CqyV9StlFCcMk7ZrrvvgnMKZm52IhMDn9L9vJutPKcDFZDraVNJKswHQrIm4m20k4G5gbEQ+mSVuRfXD+C0DSUWRHUEX19PpbSDf5kPRGZZc5DwMeItuBeZKeDU/tH5b0MuCDNdu6HJhPduT0s1xXXU//zyIWAoeleSfwzG40gM9L2kzSf5MV3UvS+J7ey58Dduap999ssp2HowrE1arPrCK+Cbw57QAg6bSUx00kDSf7fyyJiJWtXvFQK1A/JbtaZinwN7KT/JD9A7Yge0PfCFzZ4uVvEBGPkxWk/dP6vgu8OyK6ruD5ITAudUlcXmf+lWRvgo8BK4FPAm9LezK9kvrN7+t6kPXlP5Get8IxZHvUS4HryfLzozTtB2TdPbeQXaTx8wZxXkZ2Gfys1J3xF7L8tVQqqm8j+7C4i6c+4J+TmnR98KyU9Kf0/HPAS8guFPg82TaW4RRgeYrrauBSsosoGvkpsG8+pohYTHYu8AayD9BXkp3wLqTA669RPl6Y4n6I7EKH63iqO66Rj5N1U64me93U2/E7L23LhuUV+H8WcRzZ+/VBsqvyLq+Zfh/Ztq4g60Y9uhfv5dU177/HgEfSOdmefJPmPrP20jO/B1W3cEfEv8gueDopjdoSuIwsJ0vJjqgn9nL9hXRd5WNNknQ32VVnV/d3LDb4SfogMDki6u3RDzmS3kB2dL5j9NGHWur9uCAiRvXQ1DbSUDuCMhuQJL1I0uslPUvSLmRHMJf1d1xVkC60OA44u6+Kk/UNFyizgWEzssurVwPXkn2/7Lv9GlEFSHo5WVfTi8i6vWwQcRefmZlVko+gzMyskgbcjSC32267GDNmTH+HYWZmLXLTTTc9EBHPrx0/4ArUmDFjWLCg9q4dZmY2UEmqeycZd/GZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCZmVkluUCVaPr06Uhq2WP69On9vUlmZn3GBapE06dPJyIaPvbee2/23nvvHttFhAuU9RvvbFl/GHA3i22FM+bd2d8hbLB81WNAtWL66Jt37u8QrKC+et3c8LeVLV9e2bH7dTzwDckCZWa9M+HdxzDh3cf0dxg2xJRaoCRNAM4EhgFnR8SpNdN3AM4DRqQ2J0TEnDJj6ktX/vjbXHXBdwq1PX6/XXpss98R0wbMh0Rf7dn3JsdF9FWOvXc/cPi13FiZr+XSCpSkYcAM4M3AcmC+pNkRsTjX7LPAxRHxPUnjgDnAmLJi6mve6zQz23hlHkGNB5ZExFIASbOASUC+QAWwTXr+HGBFifHYIOSdABss/Fp+pjIL1EhgWW54ObBnTZvpwFWSjgG2AvattyBJU4GpAG1tbXR2djYX2Jq1Tc0/2HV2Nr+f4Bw31oocg/PciHPcN1qV53r6+yKJQ4FzI+IbkvYCzpe0a0Q8mW8UETOBmQDt7e3R0dHR1EqrdMVcFR3c0XyfsnPcWCtyDM5zI85x32hVnusp83tQ9wKjc8Oj0ri89wIXA0TEDcDmwHYlxmRmZgNEmQVqPjBW0k6SNgMmA7Nr2vwd2AdA0svJCtS/SozJzMwGiNIKVESsA6YBc4HbyK7WWyTpFEkTU7OPAe+XdAtwITAlIqKsmMzMbOAo9RxU+k7TnJpxJ+WeLwZeX2YMZmY2MPlefGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkmlFihJEyTdIWmJpBPqTD9D0sL0uFPSg2XGY2ZmA8cmZS1Y0jBgBvBmYDkwX9LsiFjc1SYiPpprfwywe1nxmJnZwFLmEdR4YElELI2Ix4FZwKQG7Q8FLiwxHjMzG0BKO4ICRgLLcsPLgT3rNZS0I7ATcG0306cCUwHa2tro7OxsLrA1a5uaf7Dr7FzR9DKc48ZakWNwnhtxjvtGq/JcT5kFqjcmA5dGxPp6EyNiJjAToL29PTo6Oppa2Rnz7mxq/sHu4I6dm16Gc9xYK3IMznMjznHfaFWe6ymzi+9eYHRueFQaV89k3L1nZmY5ZRao+cBYSTtJ2oysCM2ubSTpZcC2wA0lxmJmZgNMaQUqItYB04C5wG3AxRGxSNIpkibmmk4GZkVElBWLmZkNPKWeg4qIOcCcmnEn1QxPLzMGMzMbmHwnCTMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzqyQXKDMzq6Ru78Un6dtAtzdwjYhjS4nIzMyMxjeLXdBnUZiZmdXotkBFxHn5YUlbRsSj5YdkZmZW4ByUpL0kLQZuT8OvkvTd0iMzM7MhrchFEt8E/gdYCRARtwBvKDEmMzOzYlfxRcSymlHrS4jFzMxsgyK/qLtM0uuAkLQpcBzZT7ibmZmVpsgR1NHAh4GRwApgtzTcI0kTJN0haYmkE7ppc7CkxZIWSfppwbjNzGyQ6/EIKiIeAA7v7YIlDQNmAG8GlgPzJc2OiMW5NmOBE4HXR8QqSS/o7XrMzGxwKnIV34sl/VLSvyTdL+kXkl5cYNnjgSURsTQiHgdmAZNq2rwfmBERqwAi4v7eboCZmQ1ORc5B/ZTsSOjANDwZuBDYs4f5RgL5iyuW15lnZwBJvwOGAdMj4sraBUmaCkwFaGtro7Ozs0DYDQJbs7ap+Qe7zs4VTS/DOW6sFTkG57kR57hvtCrP9RQpUFtGxPm54QskfaKF6x8LdACjgN9KemVEPJhvFBEzgZkA7e3t0dHR0dRKz5h3Z1PzD3YHd+zc9DKc48ZakWNwnhtxjvtGq/JcT6N78T03Pf11usBhFtm9+Q4B5hRY9r3A6NzwqDQubznwh4h4ArhL0p1kBWt+sfDNzGywanQEdRNZQVIa/kBuWpBd3NDIfGCspJ3ICtNk4LCaNpcDhwLnSNqOrMtvaaHIzcxsUGt0L76dmllwRKyTNA2YS3Z+6UcRsUjSKcCCiJidpu2XbqW0HvhERKxsZr1mZjY4FDkHhaRdgXHA5l3jIuLHPc0XEXOo6Q6MiJNyzwM4Pj3MzMw26LFASTqZ7CKGcWTFZn/geqDHAmVmZraxitxJ4iBgH+C+iDgKeBXwnFKjMjOzIa9IgXosIp4E1knaBrifp1+dZ2Zm1nJFzkEtkDQC+AHZlX0PAzeUGZSZmVmRe/F9KD09S9KVwDYRcWu5YZmZ2VDX6Iu6r240LSL+VE5IZmZmjY+gvtFgWgBvanEsZmZmGzT6ou4b+zIQMzOzvEI/+W5mZtbXXKDMzKySXKDMzKySit6LbySwY759RPy2rKDMzMyK3IvvNLLfgOq64zhkV/G5QJmZWWmKHEG9HdglIvy7x2Zm1meKnINaCmxadiBmZmZ5je4k8W2yrrxHgYWSrgE2HEVFxLHlh2dmZkNVoy6+BenvTcDsPojFzMxsg0Z3kjgPQNJWwJqIWJ+GhwHPLrJwSROAM8l+8v3siDi1ZvoU4GvAvWnUdyLi7F5ug5mZDUJFzkFdA2yRG94CuLqnmVIhm0H2C7zjgEMljavT9KKI2C09XJzMzAwoVqA2j4iHuwbS8y0LzDceWBIRSyPicWAWMGnjwjQzs6GmSIF6JP/TG5JeAzxWYL6RwLLc8PI0rtY7JN0q6VJJ/qVeMzMDin0P6iPAJZJWAAJeSPbF3Vb4JXBhRKyV9AHgPOr8jIekqcBUgLa2Njo7O5ta6cg1/kpXI52dK5pehnPcWCtyDM5zI85x32hVnusp8ou68yW9DNgljbojIp4osOx7gfwR0Sieuhiia9krc4NnA1/tJoaZwEyA9vb26OjoKLD67p0x786m5h/sDu7YuellOMeNtSLH4Dw34hz3jVbluZ5C9+IjK07jgM2BV0siIn7cwzzzgbGSdiIrTJOBw/INJL0oIv6RBicCtxWO3MzMBrUi9+I7GeggK1BzyK7Kux5oWKAiYp2kacBcssvMfxQRiySdAiyIiNnAsZImAuuAfwNTNn5TzMxsMClyBHUQ8Crg5og4SlIbcEGRhUfEHLKilh93Uu75icCJxcM1M7OhoshVfI9FxJPAOknbAPfz9HNLZmZmLVfkCGqBpBHAD8hue/QwcEOZQZmZmRW5iu9D6elZkq4EtomIW8sNy8zMhroeu/iUOULSSRFxN/CgpPHlh2ZmZkNZkXNQ3wX2Ag5Nw6vJ7rFnZmZWmiLnoPaMiFdLuhkgIlZJ2qzkuMzMbIgrcgT1RLozeQBIej7wZKlRmZnZkFekQH0LuAx4gaQvkX1J98ulRmVmZkNekav4fiLpJmAfspvFvj0ifEsiMzMrVaF78UXE7cDtJcdiZma2QZEuPjMzsz7XbYGS9Oy+DMTMzCyv0RHUDQCSzu+jWMzMzDZodA5qM0mHAa+T9L+1EyPi5+WFZWZmQ12jAnU0cDgwAjigZloALlBmZlaabgtURFwPXC9pQUT8sA9jMjMzK3SZ+fmSjgXekIavA86KiCfKC8vMzIa6IgXqu8Cm6S/Au4DvAe8rKygzM7Mi34PaIyKOjIhr0+MoYI8iC5c0QdIdkpZIOqFBu3dICkntRQM3M7PBrUiBWi/pJV0Dkl4MrO9ppnSD2RnA/sA44FBJ4+q0Gw4cB/yhaNBmZjb4Feni+wTwG0lLye7FtyNwVIH5xgNLImIpgKRZwCRgcU27LwCnpfWYmZkBxW4We42kscAuadQdEbG2wLJHAstyw8uBPfMNJL0aGB0Rv5LUbYGSNBWYCtDW1kZnZ2eB1TcIbE2R8Ieuzs4VTS/DOW6sFTkG57kR57hvtCrP9RS9Wexa4NZWrljSs4DTgSkF1j8TmAnQ3t4eHR0dTa37jHl3NjX/YHdwx85NL8M5bqwVOQbnuRHnuG+0Ks/1lHmz2HuB0bnhUWlcl+HArkCnpLuB1wKzfaGEmZlBuQVqPjBW0k7pJ+InA7O7JkbEfyJiu4gYExFjgBuBiRGxoMSYzMxsgOixQEn6uaS3pi65wiJiHTANmAvcBlwcEYsknSJp4saFa2ZmQ0XRL+oeBXxL0iXAORFxR5GFR8QcYE7NuJO6adtRZJlmZjY09HhUFBFXR8ThwKuBu4GrJf1e0lGSNi07QDMzG5oKddtJeh7Z1XbvA24GziQrWPNKi8zMzIa0Hrv4JF1G9h2o84EDIuIfadJFknxBg5mZlaLIOahvRcRv6k2ICF8SbmZmpSjSxTdO0oiuAUnbSvpQeSGZmZkVK1Dvj4gHuwYiYhXw/tIiMjMzo1iBGiZJXQPpLuWblReSmZlZsXNQV5JdEPH9NPyBNM7MzKw0RQrUp8iK0gfT8Dzg7NIiMjMzo9jPbTxJ9hPv3ys/HDMzs0yR70GNBb5C9qu4m3eNj4gXlxiXmZkNcUUukjiH7OhpHfBG4MfABWUGZWZmVqRAbRER1wCKiHsiYjrw1nLDMjOzoa7IRRJr009t/FXSNLIfHdy63LDMzGyoK3IEdRywJXAs8BrgCODIMoMyMzNreASVvpR7SER8HHiY7HehzMzMStfwCCoi1gP/1UexmJmZbVDkHNTNkmYDlwCPdI2MiJ+XFpWZmQ15Rc5BbQ6sBN4EHJAebyuycEkTJN0haYmkE+pMP1rSnyUtlHS9pHG9Cd7MzAavIneS2KjzTun81QzgzcByYL6k2RGxONfspxFxVmo/ETgdmLAx6zMzs8GlyJ0kzgGidnxEvKeHWccDSyJiaVrOLGASsKFARcRDufZb1VuPmZkNTUXOQV2Re745cCCwosB8I4FlueHlwJ61jSR9GDie7Cc83lRvQZKmAlMB2tra6OzsLLD6BoGtWdvU/INdZ2eRf29jznFjrcgxOM+NOMd9o1V5rqdIF9/P8sOSLgSub1UAETEDmCHpMOCz1PmOVUTMBGYCtLe3R0dHR1PrPGPenU3NP9gd3LFz08twjhtrRY7BeW7EOe4brcpzPUUukqg1FnhBgXb3AqNzw6PSuO7MAt6+EfGYmdkgVOQc1Gqefm7oPrLfiOrJfGCspJ3ICtNk4LCaZY+NiL+mwbcCf8XMzIxiXXzDN2bBEbEu3btvLjAM+FFELJJ0CrAgImYD0yTtCzwBrMK3UDIzs6TIEdSBwLUR8Z80PALoiIjLe5o3IuYAc2rGnZR7flwv4zUzsyGiyDmok7uKE0BEPAicXFpEZmZmFCtQ9doUuTzdzMxsoxUpUAsknS7pJelxOnBT2YGZmdnQVqRAHQM8DlxEdin4GuDDZQZlZmZW5Cq+R4Bn3OjVzMysTD0eQUmal67c6xreVtLcUqMyM7Mhr0gX33bpyj0AImIVxe4kYWZmttGKFKgnJe3QNSBpR3zXcTMzK1mRy8U/A1wv6TpAwH+T7ixuZmZWliIXSVwp6dXAa9Ooj0TEA+WGZWZmQ13RL9yuB+4n+z2ocZKIiN+WF5aZmQ11Re7F9z7gOLKfy1hIdiR1A938uKCZmVkrFLlI4jhgD+CeiHgjsDvwYJlBmZmZFSlQayJiDYCkZ0fE7cAu5YZlZmZDXZFzUMvTF3UvB+ZJWgXcU2ZQZmZmRa7iOzA9nS7pN8BzgCtLjcrMzIa8Xv1sRkRcV1YgZmZmeUXOQZmZmfW5UguUpAmS7pC0RNIz7ogu6XhJiyXdKumadBslMzOz8gqUpGHADGB/YBxwqKRxNc1uBtoj4v8BlwJfLSseMzMbWMo8ghoPLImIpRHxONmPHU7KN4iI30TEo2nwRrIvA5uZmfXuIoleGgksyw0vB/Zs0P69wK/rTZA0lXSD2ra2Njo7O5sLbM3apuYf7Do7VzS9DOe4sVbkGJznRpzjvtGqPNdTZoEqTNIRQDuwd73pETETmAnQ3t4eHR0dTa3vjHl3NjX/YHdwx85NL8M5bqwVOQbnuRHnuG+0Ks/1lFmg7gVG54ZHpXFPI2lfsp/02DsivKtiZmZAueeg5gNjJe0kaTNgMjA730DS7sD3gYkRcX+JsZiZ2QBTWoGKiHXANGAucBtwcUQsknSKpImp2deArYFLJC2UNLubxZmZ2RBT6jmoiJgDzKkZd1Lu+b5lrt/MzAYu30nCzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqyQXKzMwqqdQCJWmCpDskLZF0Qp3pb5D0J0nrJB1UZixmZjawlFagJA0DZgD7A+OAQyWNq2n2d2AK8NOy4jAzs4FpkxKXPR5YEhFLASTNAiYBi7saRMTdadqTJcZhZmYDUJkFaiSwLDe8HNhzYxYkaSowFaCtrY3Ozs7mAluztqn5B7vOzhVNL8M5bqwVOQbnuRHnuG+0Ks/1lFmgWiYiZgIzAdrb26Ojo6Op5Z0x784WRDV4Hdyxc9PLcI4ba0WOwXluxDnuG63Kcz1lXiRxLzA6NzwqjTMzM+tRmQVqPjBW0k6SNgMmA7NLXJ+ZmQ0ipRWoiFgHTAPmArcBF0fEIkmnSJoIIGkPScuBdwLfl7SorHjMzGxgKfUcVETMAebUjDsp93w+WdefmZnZ0/hOEmZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkkuUGZmVkmlFihJEyTdIWmJpBPqTH+2pIvS9D9IGlNmPGZmNnCUVqAkDQNmAPsD44BDJY2rafZeYFVEvBQ4AzitrHjMzGxgKfMIajywJCKWRsTjwCxgUk2bScB56fmlwD6SVGJMZmY2QGxS4rJHAstyw8uBPbtrExHrJP0HeB7wQL6RpKnA1DT4sKQ7Som4/2xHzTb3p+P7O4ByOMd9ozJ5do77RovyvGO9kWUWqJaJiJnAzP6OoyySFkREe3/HMZg5x33DeS7fUMpxmV189wKjc8Oj0ri6bSRtAjwHWFliTGZmNkCUWaDmA2Ml7SRpM2AyMLumzWzgyPT8IODaiIgSYzIzswGitC6+dE5pGjAXGAb8KCIWSToFWBARs4EfAudLWgL8m6yIDUWDtvuyQpzjvuE8l2/I5Fg+YDEzsyrynSTMzKySXKDMzKySXKC6IWm9pIW5xxhJv2/RskdI+lArljWQDZYcS3qnpEWSnpRUqct/B1GOvybpdkm3SrpM0oi+WG9RgyjPX0g5XijpKknb98V6u43H56Dqk/RwRGxd0rLHAFdExK5lLH+gGCw5lvRy4Eng+8DHI2JB2essahDleD+yq3zXSToNICI+VfZ6ixpEed4mIh5Kz48FxkXE0WWvtzs+guoFSQ+nvx2SOiVdmvbqftJ1iyZJr5F0naSbJM2V9KI6izoVeEnaS/laWt4VufV8R9KU9PxuSZ+X9CdJf5b0sjR+K0k/kvRHSTdLqr2NVNeyxki6Nu0VXSNphxanpaUGYo4j4raIGDB3NxmgOb4qItalwRvJvldZaQM0zw/lBrcC+vcIJiL8qPMA1gML0+OyNO7h9LcD+A/Zm+RZwA3AfwGbAr8Hnp/aHUJ2eX3tsscAf8kNd5DtIXUNfweYkp7fDRyTnn8IODs9/zJwRHo+ArgT2KrOun4JHJmevwe4vL9zO9hynFtmJ9De33kdzDnOvaaP6O/cDtY8A18iuwXdX7pi66/HgLjVUT95LCJ2azD9jxGxHEDSQrIX0YPArsC8tIM0DPhHC2L5efp7E/C/6fl+wERJH0/DmwM7ALfVzLtXbp7zga+2IJ5WGSw5rrJBlWNJnwHWAT9pQTytNGjyHBGfAT4j6URgGnByC2LaKC5QG29t7vl6slwKWBQRe+UbShpNttcHcBZwZc2y1vH07tbNu1lX13pI63pH1HQtSToH2B1YERFvKbw11eQcl2/A5Dh1Y70N2CfSrv4AMmDynPMTYA79WKB8Dqq17gCeL2kvAEmbSnpFRCyLiN3S4yxgNTA8N989wDhlP+A4AtinwLrmAsfk+rJ3B4iIo9J6ul5sv+epO3QcDvxfk9vY36qY48GmcjmWNAH4JDAxIh5t0Xb2tyrmeWxunknA7U1uY1NcoFoost+9Ogg4TdItZP3Rr6vTbiXwO0l/kfS1iFgGXEzW53sxcHOB1X2BrA/7VkmL0nA9xwBHSboVeBdwXO+2qlqqmGNJB0paTtad+itJc3u/ZdVRxRyTnWcZTtYdtlDSWb3crMqpaJ5PTeu5laxbsF8/L3yZuZmZVZKPoMzMrJJcoMzMrJJcoMzMrJJcoMzMrJJcoMzMrJJcoMzMrJJcoMzMrJL+P4MycsbaQrg7AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Build the plot\n",
    "fig, ax = plt.subplots()\n",
    "ax.bar(x_positon, b_height,\n",
    "       yerr=SE,\n",
    "       #yerr=error,\n",
    "       align='center',\n",
    "       alpha=0.5,\n",
    "       ecolor='black',\n",
    "       capsize=10)\n",
    "ax.set_ylabel('accuracy of each label')\n",
    "ax.set_xticks(x_positon)\n",
    "ax.set_xticklabels(labels)\n",
    "ax.set_title('bar plot of  4 modle finetuning various layer-upt 4 layers')\n",
    "ax.yaxis.grid(True)\n",
    "\n",
    "# Save the figure and show\n",
    "plt.tight_layout()\n",
    "plt.savefig('bar_plot_with_error_bars.png')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}