transfer_learning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Transfer learning \n",
    "\n",
    "Multiclass classification with transfer learning, images features are extracted through Xception network pre-trained on image net dataset. \n",
    "Hyperparameter tunning by ```model_run_config``` dictionary parameter definition.\n",
    "Experiment results and model weights are save with **mlflow** framework. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "import os \n",
    "import numpy as np \n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf \n",
    "import mlflow \n",
    "\n",
    "from datetime import datetime\n",
    "from tensorflow.keras.utils import img_to_array, array_to_img, load_img\n",
    "from tensorflow.keras.layers import Input, Flatten, Dense, Dropout, Rescaling\n",
    "from tensorflow.python.keras.models import Sequential\n",
    "from sklearn.model_selection import train_test_split\n",
    "from tensorflow.keras.applications import ResNet152\n",
    "from tensorflow.keras.applications import Xception\n",
    "from tensorflow.keras.layers import Input\n",
    "\n",
    "from common.common_params import data_dir, img_dir\n",
    "from common.common_plots import plot_img, plot_train_val_accuracy_and_loss\n",
    "from data_load import get_train_val_set, prepare_subfolder_extract\n",
    "from common.common_params import data_dir, img_dir, npy_dir, model_dir, MAX_NUMBER_CLASS, input_shape, image_size, IMG_SIZE\n",
    "from common.common_img_functions import read_img\n",
    "from common.common_mlflow import save_experiment_mlflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.8.3\n"
     ]
    }
   ],
   "source": [
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_run_config = {\n",
    "    \"experiment_name\" : \"xception-transfer-learning-dog-breed-classifier\",\n",
    "    \"batch_size\" : 32,\n",
    "    \"NB_IMG_PER_CLASS\" : 150 ,\n",
    "    \"NB_CLASS\" : 30 ,\n",
    "    \"num_epochs\" : 30 ,\n",
    "    \"NB_NEURONS_DENSE_LAYER\": 512,\n",
    "    \"droupout\" : 0.8,\n",
    "    # scale = 1\n",
    "    \"scale\" : 1./255. ,\n",
    "    \"input_shape\" : input_shape\n",
    "}\n",
    "\n",
    "batch_size              =  model_run_config['batch_size']\n",
    "NB_IMG_PER_CLASS        =  model_run_config[\"NB_IMG_PER_CLASS\"]\n",
    "NB_CLASS                =  model_run_config[\"NB_CLASS\"]\n",
    "num_epochs              =  model_run_config[\"num_epochs\"]\n",
    "NB_NEURONS_DENSE_LAYER  =  model_run_config[\"NB_NEURONS_DENSE_LAYER\"]\n",
    "droupout                =  model_run_config[\"droupout\"]\n",
    "# scale = 1\n",
    "scale                   =  model_run_config[\"scale\"]\n",
    "experiment_name         =  model_run_config[\"experiment_name\"]\n",
    "input_shape             =  model_run_config[\"input_shape\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "list_of_folder = np.load(os.path.join(npy_dir,\"top_breed_folder_list_30.npy\"),  allow_pickle=True)\n",
    "train_dir, test_dir = prepare_subfolder_extract(NB_IMG_PER_CLASS, NB_CLASS, list_of_folder) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Transfer learning using Xception "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 6057 files belonging to 30 classes.\n",
      "Using 4846 files for training.\n",
      "Found 6057 files belonging to 30 classes.\n",
      "Using 1211 files for validation.\n",
      "WARNING:tensorflow:AutoGraph could not transform <function get_train_val_set.<locals>.<lambda> at 0x00000224852614C8> and will run it as-is.\n",
      "Cause: could not parse the source code of <function get_train_val_set.<locals>.<lambda> at 0x00000224852614C8>: no matching AST found among candidates:\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
      "WARNING: AutoGraph could not transform <function get_train_val_set.<locals>.<lambda> at 0x00000224852614C8> and will run it as-is.\n",
      "Cause: could not parse the source code of <function get_train_val_set.<locals>.<lambda> at 0x00000224852614C8>: no matching AST found among candidates:\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
      "WARNING:tensorflow:AutoGraph could not transform <function get_train_val_set.<locals>.<lambda> at 0x00000224851C2288> and will run it as-is.\n",
      "Cause: could not parse the source code of <function get_train_val_set.<locals>.<lambda> at 0x00000224851C2288>: no matching AST found among candidates:\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n",
      "WARNING: AutoGraph could not transform <function get_train_val_set.<locals>.<lambda> at 0x00000224851C2288> and will run it as-is.\n",
      "Cause: could not parse the source code of <function get_train_val_set.<locals>.<lambda> at 0x00000224851C2288>: no matching AST found among candidates:\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "# coding=utf-8\n",
      "(lambda x, y: (x, tf.one_hot(y, depth=num_classes)))\n",
      "To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert\n"
     ]
    }
   ],
   "source": [
    "train_ds, validation_ds, class_names = get_train_val_set(train_dir, image_size, batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['n02085936-Maltese_dog',\n",
       " 'n02086240-Shih-Tzu',\n",
       " 'n02086646-Blenheim_spaniel',\n",
       " 'n02086910-papillon',\n",
       " 'n02088094-Afghan_hound',\n",
       " 'n02088364-beagle',\n",
       " 'n02088466-bloodhound',\n",
       " 'n02090721-Irish_wolfhound',\n",
       " 'n02091134-whippet',\n",
       " 'n02091244-Ibizan_hound',\n",
       " 'n02091467-Norwegian_elkhound',\n",
       " 'n02091831-Saluki',\n",
       " 'n02092002-Scottish_deerhound',\n",
       " 'n02095570-Lakeland_terrier',\n",
       " 'n02095889-Sealyham_terrier',\n",
       " 'n02096051-Airedale',\n",
       " 'n02096177-cairn',\n",
       " 'n02096294-Australian_terrier',\n",
       " 'n02097474-Tibetan_terrier',\n",
       " 'n02107683-Bernese_mountain_dog',\n",
       " 'n02108000-EntleBucher',\n",
       " 'n02110185-Siberian_husky',\n",
       " 'n02110806-basenji',\n",
       " 'n02110958-pug',\n",
       " 'n02111129-Leonberg',\n",
       " 'n02111277-Newfoundland',\n",
       " 'n02111500-Great_Pyrenees',\n",
       " 'n02111889-Samoyed',\n",
       " 'n02112018-Pomeranian',\n",
       " 'n02112137-chow']"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_xception_model(scale=1):\n",
    "    model = Sequential()\n",
    "    if scale != 1 : \n",
    "        model.add(Rescaling(scale))\n",
    "    model.add(Xception(include_top=False, pooling='avg', weights=\"imagenet\"))\n",
    "    model.add(Dense(NB_NEURONS_DENSE_LAYER, activation='relu'))\n",
    "    model.add(Dropout(droupout))\n",
    "    model.add(Dense(NB_CLASS, activation='softmax'))\n",
    "    model.layers[0].trainable = False\n",
    "    model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [],
   "source": [
    "xmodel = get_xception_model(scale=scale)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Model failed to serialize as JSON. Ignoring... Layer ModuleWrapper has arguments in `__init__` and therefore must override `get_config`.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Model failed to serialize as JSON. Ignoring... Layer ModuleWrapper has arguments in `__init__` and therefore must override `get_config`.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "152/152 [==============================] - 1174s 8s/step - loss: 1.8550 - accuracy: 0.4897 - val_loss: 1.7937 - val_accuracy: 0.5211\n",
      "Epoch 2/30\n",
      "152/152 [==============================] - 1111s 7s/step - loss: 1.0309 - accuracy: 0.7185 - val_loss: 2.1319 - val_accuracy: 0.5797\n",
      "Epoch 3/30\n",
      "152/152 [==============================] - 1111s 7s/step - loss: 0.7441 - accuracy: 0.8000 - val_loss: 1.3891 - val_accuracy: 0.6424\n",
      "Epoch 4/30\n",
      "152/152 [==============================] - 1118s 7s/step - loss: 0.5940 - accuracy: 0.8465 - val_loss: 1.6135 - val_accuracy: 0.6416\n",
      "Epoch 5/30\n",
      "152/152 [==============================] - 1125s 7s/step - loss: 0.5629 - accuracy: 0.8496 - val_loss: 1.1792 - val_accuracy: 0.6862\n",
      "Epoch 6/30\n",
      "152/152 [==============================] - 1112s 7s/step - loss: 0.4047 - accuracy: 0.8954 - val_loss: 0.9435 - val_accuracy: 0.7572\n",
      "Epoch 7/30\n",
      "152/152 [==============================] - 1109s 7s/step - loss: 0.3630 - accuracy: 0.9032 - val_loss: 1.6378 - val_accuracy: 0.6623\n",
      "Epoch 8/30\n",
      "152/152 [==============================] - 1100s 7s/step - loss: 0.3607 - accuracy: 0.9069 - val_loss: 1.1522 - val_accuracy: 0.7085\n",
      "Epoch 9/30\n",
      "152/152 [==============================] - 1100s 7s/step - loss: 0.2802 - accuracy: 0.9298 - val_loss: 1.3130 - val_accuracy: 0.7283\n",
      "Epoch 10/30\n",
      "152/152 [==============================] - 1118s 7s/step - loss: 0.2618 - accuracy: 0.9352 - val_loss: 0.9527 - val_accuracy: 0.7779\n",
      "Epoch 11/30\n",
      "152/152 [==============================] - 1111s 7s/step - loss: 0.2589 - accuracy: 0.9362 - val_loss: 1.1102 - val_accuracy: 0.7432\n",
      "Epoch 12/30\n",
      "152/152 [==============================] - 1103s 7s/step - loss: 0.2171 - accuracy: 0.9470 - val_loss: 1.2459 - val_accuracy: 0.7614\n",
      "Epoch 13/30\n",
      "152/152 [==============================] - 1106s 7s/step - loss: 0.2066 - accuracy: 0.9505 - val_loss: 1.4672 - val_accuracy: 0.7201\n",
      "Epoch 14/30\n",
      "152/152 [==============================] - 1092s 7s/step - loss: 0.1745 - accuracy: 0.9556 - val_loss: 1.5290 - val_accuracy: 0.7135\n",
      "Epoch 15/30\n",
      "152/152 [==============================] - 1111s 7s/step - loss: 0.2669 - accuracy: 0.9354 - val_loss: 1.0289 - val_accuracy: 0.7655\n",
      "Epoch 16/30\n",
      "152/152 [==============================] - 1099s 7s/step - loss: 0.2016 - accuracy: 0.9503 - val_loss: 1.3106 - val_accuracy: 0.7316\n",
      "Epoch 17/30\n",
      "152/152 [==============================] - 1095s 7s/step - loss: 0.2022 - accuracy: 0.9519 - val_loss: 1.2453 - val_accuracy: 0.7399\n",
      "Epoch 18/30\n",
      "152/152 [==============================] - 1087s 7s/step - loss: 0.1618 - accuracy: 0.9610 - val_loss: 1.1338 - val_accuracy: 0.7514\n",
      "Epoch 19/30\n",
      "152/152 [==============================] - 1102s 7s/step - loss: 0.1411 - accuracy: 0.9662 - val_loss: 1.4275 - val_accuracy: 0.7308\n",
      "Epoch 20/30\n",
      "152/152 [==============================] - 1092s 7s/step - loss: 0.1009 - accuracy: 0.9761 - val_loss: 1.1072 - val_accuracy: 0.7770\n",
      "Epoch 21/30\n",
      "152/152 [==============================] - 1084s 7s/step - loss: 0.0805 - accuracy: 0.9823 - val_loss: 1.3868 - val_accuracy: 0.7399\n",
      "Epoch 22/30\n",
      "152/152 [==============================] - 1092s 7s/step - loss: 0.1505 - accuracy: 0.9655 - val_loss: 1.2335 - val_accuracy: 0.7432\n",
      "Epoch 23/30\n",
      "152/152 [==============================] - 1091s 7s/step - loss: 0.1884 - accuracy: 0.9577 - val_loss: 2.2093 - val_accuracy: 0.6854\n",
      "Epoch 24/30\n",
      "152/152 [==============================] - 1091s 7s/step - loss: 0.1666 - accuracy: 0.9641 - val_loss: 1.5777 - val_accuracy: 0.7242\n",
      "Epoch 25/30\n",
      "152/152 [==============================] - 1079s 7s/step - loss: 0.0907 - accuracy: 0.9767 - val_loss: 1.5296 - val_accuracy: 0.7275\n",
      "Epoch 26/30\n",
      "152/152 [==============================] - 1090s 7s/step - loss: 0.1143 - accuracy: 0.9777 - val_loss: 1.3123 - val_accuracy: 0.7316\n",
      "Epoch 27/30\n",
      "152/152 [==============================] - 1085s 7s/step - loss: 0.1389 - accuracy: 0.9678 - val_loss: 1.1543 - val_accuracy: 0.7647\n",
      "Epoch 28/30\n",
      "152/152 [==============================] - 1093s 7s/step - loss: 0.1145 - accuracy: 0.9748 - val_loss: 1.3463 - val_accuracy: 0.7556\n",
      "Epoch 29/30\n",
      "152/152 [==============================] - 1080s 7s/step - loss: 0.1070 - accuracy: 0.9721 - val_loss: 1.5126 - val_accuracy: 0.7440\n",
      "Epoch 30/30\n",
      "152/152 [==============================] - 1082s 7s/step - loss: 0.1367 - accuracy: 0.9670 - val_loss: 2.0080 - val_accuracy: 0.6986\n"
     ]
    }
   ],
   "source": [
    "run_name = datetime.now().strftime(\"%Y%m%d_%H%M%S\")\n",
    "logdir = os.path.join(data_dir,\"logs\", experiment_name, run_name)\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=logdir,  write_graph=True, histogram_freq=1)\n",
    "\n",
    "history = xmodel.fit(x=train_ds, \n",
    "        epochs=num_epochs, \n",
    "        validation_data=validation_ds,\n",
    "        callbacks=[tensorboard_callback])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "MLflow’s mlflow.sklearn library allows loading models back as a scikit-learn Pipeline object."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MLFlow Run ID:  20acd9cd81654a85bc287a9be718debf\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:Found untraced functions such as rescaling_3_layer_call_fn, rescaling_3_layer_call_and_return_conditional_losses, dense_6_layer_call_fn, dense_6_layer_call_and_return_conditional_losses, dropout_3_layer_call_fn while saving (showing 5 of 48). These functions will not be directly callable after loading.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: C:\\Users\\clara\\AppData\\Local\\Temp\\tmp4ws6g5oq\\model\\data\\model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: C:\\Users\\clara\\AppData\\Local\\Temp\\tmp4ws6g5oq\\model\\data\\model\\assets\n"
     ]
    }
   ],
   "source": [
    "save_experiment_mlflow(xmodel, model_run_config, run_name, history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_train_val_accuracy_and_loss(history)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "IMG2",
   "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": "3.7.9"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}