DL_Lab10c_J050.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "DL_Lab10c_J050.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yufkR8oGJU-d",
        "colab_type": "text"
      },
      "source": [
        "#Fine-tuning the top layers of a a pre-trained network\n",
        "To further improve our previous result, we can try to \"fine-tune\" the last convolutional block of the VGG16 model alongside the top-level classifier. Fine-tuning consist in starting from a trained network, then re-training it on a new dataset using very small weight updates. In our case, this can be done in 3 steps:\n",
        "\n",
        "1. instantiate the convolutional base of VGG16 and load its weights\n",
        "2. add our previously defined fully-connected model on top, and load its weights\n",
        "3. freeze the layers of the VGG16 model up to the last convolutional block\n",
        "\n",
        "![alt text](http://upscfever.com/upsc-fever/images/fine-tuning.png)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M-iIjVOPJM47",
        "colab_type": "code",
        "outputId": "d3729c35-0518-471e-8921-54ebd0e6cec1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#code part 1\n",
        "from bs4 import BeautifulSoup\n",
        "import numpy as np\n",
        "import requests\n",
        "import cv2\n",
        "import PIL.Image\n",
        "import urllib\n",
        "\n",
        "page = requests.get(\"http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n04194289\")#ship synset\n",
        "print(page.content)\n",
        "\n",
        "# BeautifulSoup is an HTML parsing library\n",
        "\n",
        "soup = BeautifulSoup(page.content, 'html.parser')#puts the content of the website into the soup variable, each url on a different line\n",
        "#print(soup)\n",
        "#print(soup.prettify())\n",
        "\n",
        "#code part 1.1\n",
        "bikes_page = requests.get(\"http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n02834778\")#bicycle synset\n",
        "print(bikes_page.content)\n",
        "\n",
        "# BeautifulSoup is an HTML parsing library\n",
        "from bs4 import BeautifulSoup\n",
        "bikes_soup = BeautifulSoup(bikes_page.content, 'html.parser')#puts the content of the website into the soup variable, each url on a different line\n",
        "\n",
        "#code part 2\n",
        "str_soup=str(soup)#convert soup to string so it can be split\n",
        "type(str_soup)\n",
        "split_urls=str_soup.split('\\r\\n')#split so each url is a different possition on a list\n",
        "print(len(split_urls))#print the length of the list so you know how many urls you have\n",
        "\n",
        "#code part 2.2\n",
        "bikes_str_soup=str(bikes_soup)#convert soup to string so it can be split\n",
        "type(bikes_str_soup)\n",
        "bikes_split_urls=bikes_str_soup.split('\\r\\n')#split so each url is a different possition on a list\n",
        "print(len(bikes_split_urls))\n",
        "\n",
        "#code part 3\n",
        "#check if all the images where stored on the files system\n",
        "!mkdir /content/train #create the Train folder\n",
        "!mkdir /content/train/ships #create the ships folder\n",
        "!mkdir /content/train/bikes #create the bikes folder\n",
        "!mkdir /content/validation\n",
        "!mkdir /content/validation/ships #create the ships folder\n",
        "!mkdir /content/validation/bikes #create the bikes folder\n",
        "#!ls /content/train/ships #list the files inside ships\n",
        "!mkdir /content/test/ \n",
        "!mkdir /content/test/test #list the files inside test\n",
        "\n",
        "#code part 4\n",
        "img_rows, img_cols = 32, 32 #number of rows and columns to convert the images to\n",
        "input_shape = (img_rows, img_cols, 3)#format to store the images (rows, columns,channels) called channels last\n",
        "\n",
        "def url_to_image(url):\n",
        "\t# download the image, convert it to a NumPy array, and then read\n",
        "\t# it into OpenCV format\n",
        "\tresp = urllib.request.urlopen(url)\n",
        "\timage = np.asarray(bytearray(resp.read()), dtype=\"uint8\")\n",
        "\timage = cv2.imdecode(image, cv2.IMREAD_COLOR)\n",
        " \n",
        "\t# return the image\n",
        "\treturn image\n",
        "\n",
        "n_of_training_images=150#the number of training images to use\n",
        "for progress in range(n_of_training_images):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(split_urls[progress])\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/train/ships/img'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "\n",
        "#do the same for bikes:\n",
        "for progress in range(n_of_training_images):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not bikes_split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(bikes_split_urls[progress])\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/train/bikes/img'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "        \n",
        "        \n",
        "#Validation data:\n",
        "\n",
        "for progress in range(50):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(split_urls[n_of_training_images+progress])#get images that are different from the ones used for training\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/validation/ships/img'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "\n",
        "#do the same for bikes:\n",
        "for progress in range(50):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not bikes_split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(bikes_split_urls[n_of_training_images+progress])#get images that are different from the ones used for training\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/validation/bikes/img'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "        \n",
        "print(\"\\nTRAIN:\\n\")          \n",
        "print(\"\\nlist the files inside ships directory:\\n\")        \n",
        "!ls /content/train/ships #list the files inside ships\n",
        "print(\"\\nlist the files inside bikes directory:\\n\")\n",
        "!ls /content/train/bikes #list the files inside bikes\n",
        "print(\"\\nVALIDATION:\\n\")\n",
        "print(\"\\nlist the files inside ships directory:\\n\")        \n",
        "!ls /content/validation/ships #list the files inside ships\n",
        "print(\"\\nlist the files inside bikes directory:\\n\")\n",
        "!ls /content/validation/bikes #list the files inside bikes   \n",
        "\n",
        "for progress in range(50):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(split_urls[n_of_training_images+progress+50])#get images that are different from the ones used for training\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/test/test/ships_'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "\n",
        "#do the same for bikes:\n",
        "for progress in range(50):#store all the images on a directory\n",
        "    # Print out progress whenever progress is a multiple of 20 so we can follow the\n",
        "    # (relatively slow) progress\n",
        "    if(progress%20==0):\n",
        "        print(progress)\n",
        "    if not bikes_split_urls[progress] == None:\n",
        "      try:\n",
        "        I = url_to_image(bikes_split_urls[n_of_training_images+progress+50])#get images that are different from the ones used for training\n",
        "        if (len(I.shape))==3: #check if the image has width, length and channels\n",
        "          save_path = '/content/test/test/bikes_'+str(progress)+'.jpg'#create a name of each image\n",
        "          cv2.imwrite(save_path,I)\n",
        "\n",
        "      except:\n",
        "        None\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "b\"http://static.flickr.com/85/233578996_8f74463083.jpg\\r\\nhttp://farm3.static.flickr.com/2317/2472759369_f4cc1ff8a0.jpg\\r\\nhttp://static.flickr.com/2173/2071797511_200e3e4d2e.jpg\\r\\nhttp://farm2.static.flickr.com/1019/534041866_3aa182311c.jpg\\r\\nhttp://static.flickr.com/88/231940102_7beb5ecf3e.jpg\\r\\nhttp://farm4.static.flickr.com/3138/2724596903_d74dd64821.jpg\\r\\nhttp://static.flickr.com/174/453920264_9406e9972c.jpg\\r\\nhttp://farm2.static.flickr.com/1419/1408865234_caf87d02e3.jpg\\r\\nhttp://farm1.static.flickr.com/147/423542865_c091691a6b.jpg\\r\\nhttp://farm3.static.flickr.com/2074/2367492616_272bf6af6a.jpg\\r\\nhttp://farm2.static.flickr.com/1081/1365111200_8d105d9fb1.jpg\\r\\nhttp://farm3.static.flickr.com/2269/2348294071_94fc609f4a.jpg\\r\\nhttp://static.flickr.com/1016/1443997081_b5eb875055.jpg\\r\\nhttp://www.walleniuslines.com/img_upload/maersk_teal.jpg\\r\\nhttp://www.fondear.org/infonautic/Barco/Barco_Navegando/Contenedores_Peligro/sea_container_ship2.jpg\\r\\nhttp://static.flickr.com/198/491737075_0100196a8f.jpg\\r\\nhttp://farm4.static.flickr.com/3132/2640561254_b114ac2788.jpg\\r\\nhttp://farm1.static.flickr.com/12/14188200_7c98ba44cc.jpg\\r\\nhttp://farm4.static.flickr.com/3226/2350689291_dbcebb71ee.jpg\\r\\nhttp://farm1.static.flickr.com/181/425345889_cb12e533d0.jpg\\r\\nhttp://farm1.static.flickr.com/99/306190074_b992566897.jpg\\r\\nhttp://farm2.static.flickr.com/1087/986936649_ee4fb54d34.jpg\\r\\nhttp://209.191.55.51/images/tahiti/ship.jpg\\r\\nhttp://farm4.static.flickr.com/3282/2576278846_63f669f036.jpg\\r\\nhttp://farm4.static.flickr.com/3171/2640549706_a942ef34dc.jpg\\r\\nhttp://farm4.static.flickr.com/3174/2596564567_44576822ca.jpg\\r\\nhttp://farm2.static.flickr.com/1012/536894323_84d889253a.jpg\\r\\nhttp://farm3.static.flickr.com/2322/2492780954_38cd3f3486.jpg\\r\\nhttp://farm4.static.flickr.com/3260/2639707723_2af19b3414.jpg\\r\\nhttp://farm2.static.flickr.com/1005/1382829231_d16ebb046f.jpg\\r\\nhttp://farm2.static.flickr.com/1266/986998821_5f149dc334.jpg\\r\\nhttp://farm1.static.flickr.com/82/278168799_4b010387aa.jpg\\r\\nhttp://farm1.static.flickr.com/137/391692335_15b86a4963.jpg\\r\\nhttp://farm4.static.flickr.com/3163/2639657427_bde06d1a66.jpg\\r\\nhttp://www.simplonpc.co.uk/Waverley2005/Waverley-Cruise_20051010-08_mv-Geestdijk.jpg\\r\\nhttp://www.c-map.no/img/news/091106.jpg\\r\\nhttp://www.corlaan.dds.nl/pic/ijsse3a.jpg\\r\\nhttp://farm4.static.flickr.com/3100/2476988884_fc1d868731.jpg\\r\\nhttp://www.wright.edu/news/cruise_ship_lg.jpg\\r\\nhttp://www.ships-for-sale.com/ships%20for%20sale%20pics/3_masted_tall_ship.jpg\\r\\nhttp://farm4.static.flickr.com/3058/2641276110_19648b062f.jpg\\r\\nhttp://www.warshipsifr.com/media/burningShip3.jpg\\r\\nhttp://farm4.static.flickr.com/3254/2353684162_4221ee78c2.jpg\\r\\nhttp://farm3.static.flickr.com/2197/2713791465_810b0feb85.jpg\\r\\nhttp://farm1.static.flickr.com/35/97785017_4f68357f1b.jpg\\r\\nhttp://farm2.static.flickr.com/1344/1408870240_4573ace986.jpg\\r\\nhttp://static.flickr.com/2363/2088993690_f0afa46ddf.jpg\\r\\nhttp://farm4.static.flickr.com/3294/2525068398_407d6d9fb8.jpg\\r\\nhttp://www.germes-online.com/direct/dbimage/50091130/General_Cargo_Ship_DWT_3300.jpg\\r\\nhttp://childsupportfoundations.org/Images/hospital_ship_box.jpg\\r\\nhttp://farm3.static.flickr.com/2077/2111214220_8dba8ca42f.jpg\\r\\nhttp://static.flickr.com/47/145275698_ee5670710a.jpg\\r\\nhttp://farm2.static.flickr.com/1325/1448780891_83d8fe4a44.jpg\\r\\nhttp://farm4.static.flickr.com/3100/2350638583_2f2db30ce0.jpg\\r\\nhttp://farm3.static.flickr.com/2400/2277555953_d421a9fec5.jpg\\r\\nhttp://farm3.static.flickr.com/2104/2310067920_621f883dd1.jpg\\r\\nhttp://farm3.static.flickr.com/2386/1554133602_8258a62709.jpg\\r\\nhttp://farm2.static.flickr.com/1357/1407989677_ab300aa696.jpg\\r\\nhttp://farm4.static.flickr.com/3110/2627593753_df2db49f87.jpg\\r\\nhttp://farm4.static.flickr.com/3217/2345819329_ab90c25dda.jpg\\r\\nhttp://farm1.static.flickr.com/197/447349675_00e6cdc6cf.jpg\\r\\nhttp://www.borashipping.com/uploads/shipbuildings/sb3.jpg\\r\\nhttp://farm4.static.flickr.com/3279/2493549904_03c29028c3.jpg\\r\\nhttp://www.indybay.org/uploads/2008/01/22/soo08-287.jpg\\r\\nhttp://farm4.static.flickr.com/3110/2295789714_a86d56eb08.jpg\\r\\nhttp://farm4.static.flickr.com/3020/2584711173_af9d536da0.jpg\\r\\nhttp://static.flickr.com/42/124387718_63f426774e.jpg\\r\\nhttp://farm2.static.flickr.com/1263/559973221_6a2e1920f5.jpg\\r\\nhttp://farm3.static.flickr.com/2022/2116299410_05fce403d0.jpg\\r\\nhttp://www.boatnerd.com/news/newpictures/legrandehermine7-07-01-mn.jpg\\r\\nhttp://www.artemotore.com/public/club/foto_news/OTE0V9A6AJAY.jpg\\r\\nhttp://www.fimr.fi/en/aranda/laiva/en_GB/laiva/_files/12076504320080273/default/1_arandaTakaaWeb.jpg\\r\\nhttp://farm3.static.flickr.com/2100/2240181304_05152b5ec1.jpg\\r\\nhttp://static.flickr.com/2202/1809534307_f0698af93a.jpg\\r\\nhttp://farm3.static.flickr.com/2066/2519134597_c290a5ff42.jpg\\r\\nhttp://farm1.static.flickr.com/103/285559622_4c02e567cf.jpg\\r\\nhttp://farm4.static.flickr.com/3012/2492751811_6bcb89c181.jpg\\r\\nhttp://farm1.static.flickr.com/128/423544132_d54428ed36.jpg\\r\\nhttp://farm2.static.flickr.com/1013/907987911_4b91dc732e.jpg\\r\\nhttp://www.seascout.org/about/history/photos/1973-Trident-11.jpg\\r\\nhttp://static.flickr.com/1088/528148193_875873e338.jpg\\r\\nhttp://www.victorialodging.com/files/bounty_tall_ship.jpg\\r\\nhttp://www.adventureassociates.com/media/vessels/SKO2_Ship.jpg\\r\\nhttp://static.flickr.com/85/247715192_ce2b0dc28f.jpg\\r\\nhttp://static.flickr.com/1062/1467640552_a65e946b29.jpg\\r\\nhttp://farm2.static.flickr.com/1013/528148183_306190efcb.jpg\\r\\nhttp://farm1.static.flickr.com/32/98158197_adcd005b00.jpg\\r\\nhttp://static.flickr.com/148/416408425_1dbe2cd7fe.jpg\\r\\nhttp://photos.msts-history.org/images/mm-patrick.jpg\\r\\nhttp://farm2.static.flickr.com/1434/987652394_18b58fe11c.jpg\\r\\nhttp://www.spanopoulos-group.com/images/orca_research.jpg\\r\\nhttp://farm2.static.flickr.com/1182/909081796_2e2f4bff2b.jpg\\r\\nhttp://farm4.static.flickr.com/3015/2370814369_f018acc79b.jpg\\r\\nhttp://static.flickr.com/1386/1405446605_37cfea5cdd.jpg\\r\\nhttp://farm4.static.flickr.com/3163/2525263117_1c4e136aec.jpg\\r\\nhttp://farm3.static.flickr.com/2220/2337692322_55ab0952d1.jpg\\r\\nhttp://dlund.20m.com/images_new/Ship0428b.JPG\\r\\nhttp://farm3.static.flickr.com/2091/2348301179_144279a917.jpg\\r\\nhttp://farm3.static.flickr.com/2267/2478862000_3e8c4a2e5d.jpg\\r\\nhttp://www.mastermariners.org.nz/wp-content/uploads/2008/05/vessel_near_miss_3.jpg\\r\\nhttp://farm2.static.flickr.com/1354/1408870744_8122e7bd30.jpg\\r\\nhttp://farm4.static.flickr.com/3283/2680852912_07a7024aa4.jpg\\r\\nhttp://farm1.static.flickr.com/184/423544305_919c33018f.jpg\\r\\nhttp://static.flickr.com/94/231693330_650503f3d4.jpg\\r\\nhttp://www.macduffshipdesign.com/images/clip_image008_001.jpg\\r\\nhttp://farm4.static.flickr.com/3062/2492752927_d0bb10a4db.jpg\\r\\nhttp://farm1.static.flickr.com/146/425253720_22da63045f.jpg\\r\\nhttp://farm1.static.flickr.com/91/274979078_be8e5c9214.jpg\\r\\nhttp://farm3.static.flickr.com/2114/1595040044_5bcedd4cb7.jpg\\r\\nhttp://military.technomad.com/images/ship_system4.jpg\\r\\nhttp://farm4.static.flickr.com/3258/2468476494_0902a809a0.jpg\\r\\nhttp://static.flickr.com/230/453925740_33c5c67738.jpg\\r\\nhttp://farm3.static.flickr.com/2287/2072592222_1e9537acc1.jpg\\r\\nhttp://farm4.static.flickr.com/3279/2296054291_4674151a66.jpg\\r\\nhttp://web.mac.com/derm22/iWeb/Ivanhoe_Field_School/Project%20_files/Ivanhoe_pic.png\\r\\nhttp://static.flickr.com/1372/600598136_bd01ef940e.jpg\\r\\nhttp://farm3.static.flickr.com/2187/2210887412_8518ed5961.jpg\\r\\nhttp://static.flickr.com/158/349417687_08bf4e538c.jpg\\r\\nhttp://farm4.static.flickr.com/3027/2493578484_f8b18ebc1e.jpg\\r\\nhttp://imgsrv.kyw1060.com/image/kyw/UserFiles/Image/ship1.jpg\\r\\nhttp://farm4.static.flickr.com/3168/2568844735_5f1ac75233.jpg\\r\\nhttp://www.orcadian.co.uk/archive/2002/russianboat.jpg\\r\\nhttp://farm1.static.flickr.com/128/416410398_8def632b17.jpg\\r\\nhttp://farm3.static.flickr.com/2017/2467849087_6e79fd617e.jpg\\r\\nhttp://www.admiralshipping.org/images/ship_freight.gif\\r\\nhttp://farm3.static.flickr.com/2353/2217570196_503c5fffa9.jpg\\r\\nhttp://farm2.static.flickr.com/1115/1463757661_b0c653912f.jpg\\r\\nhttp://farm1.static.flickr.com/38/123127357_b399d76384.jpg\\r\\nhttp://farm3.static.flickr.com/2084/2482610056_75e5c94f12.jpg\\r\\nhttp://static.flickr.com/1181/1327210126_6e775c0fce.jpg\\r\\nhttp://farm4.static.flickr.com/3106/2567130881_2a18899e2e.jpg\\r\\nhttp://farm3.static.flickr.com/2327/2300313948_2cc2b9483c.jpg\\r\\nhttp://farm4.static.flickr.com/3126/2578532726_faf0b834a9.jpg\\r\\nhttp://farm1.static.flickr.com/163/436369534_c7cf28991e.jpg\\r\\nhttp://www.stemnet.nf.ca/CITE/sealingvessel_small.jpg\\r\\nhttp://farm1.static.flickr.com/47/145283840_0123484298.jpg\\r\\nhttp://farm1.static.flickr.com/30/57020037_c821b147d6.jpg\\r\\nhttp://krapp.org/rupert/blog-uploads/explorer.jpg\\r\\nhttp://farm3.static.flickr.com/2243/2147913518_de1442e28a.jpg\\r\\nhttp://www.kepcorp.com/upload/PressRelease/artist%20impression%20of%20icebreakers(1).jpg\\r\\nhttp://farm1.static.flickr.com/189/486654509_8abbca6bd8.jpg\\r\\nhttp://farm3.static.flickr.com/2153/2153066783_1a053e1228.jpg\\r\\nhttp://farm3.static.flickr.com/2321/2264736201_2ed4a1844b.jpg\\r\\nhttp://commodoreinc.com/ships_files/RORO_L.jpg\\r\\nhttp://farm4.static.flickr.com/3083/2388584668_b360746926.jpg\\r\\nhttp://farm2.static.flickr.com/1382/1061511728_1b45a55602.jpg\\r\\nhttp://farm3.static.flickr.com/2278/2358613750_4aa8179f35.jpg\\r\\nhttp://static.flickr.com/90/235765332_69d6348d9b.jpg\\r\\nhttp://farm2.static.flickr.com/1228/843233478_d4f0e37d80.jpg\\r\\nhttp://static.flickr.com/2225/2118893208_6fe69c9574.jpg\\r\\nhttp://www.nuclearfiles.org/images/key-issues/missile-defense/sea-based-radar.jpg\\r\\nhttp://static.flickr.com/146/376379013_1b7af0e668.jpg\\r\\nhttp://farm3.static.flickr.com/2014/2170677672_6397b769c6.jpg\\r\\nhttp://farm4.static.flickr.com/3120/2729349407_cffc4d26cc.jpg\\r\\nhttp://farm3.static.flickr.com/2139/2427471715_644039a1d0.jpg\\r\\nhttp://farm3.static.flickr.com/2240/2507549817_ac355c7538.jpg\\r\\nhttp://farm4.static.flickr.com/3205/2318410514_e855eaa046.jpg\\r\\nhttp://farm1.static.flickr.com/49/145283849_b8ea58b7cf.jpg\\r\\nhttp://www.worldcruise-network.com/contractor_images/resolve-marine/5-cargo-ship.jpg\\r\\nhttp://farm3.static.flickr.com/2066/2531273579_8a1b4c9ab7.jpg\\r\\nhttp://www.shipconstructor.com/images/projects/commercial/taysongeneralcargovessel_large01.jpg\\r\\nhttp://farm1.static.flickr.com/81/246444203_e260977a19.jpg\\r\\nhttp://static.flickr.com/115/279850413_8b2ab78cb9.jpg\\r\\nhttp://farm1.static.flickr.com/125/360485065_9f45ffaa29.jpg\\r\\nhttp://lh3.ggpht.com/_6wOVgaRrTL8/R4B4xmfEy4I/AAAAAAAAAQU/pIKKdag_SyY/bdaybash83.jpg\\r\\nhttp://www.bridgehands.com/Reviews/Bridge_Cruises/Marco_Polo.jpg\\r\\nhttp://farm3.static.flickr.com/2034/2497655824_b5afd1a48e.jpg\\r\\nhttp://farm1.static.flickr.com/208/514507931_7b1328abf4.jpg\\r\\nhttp://www.apparelsearch.com/Vessel_customs_photo.jpg\\r\\nhttp://www.walleniuslines.com/img_upload/tosca.jpg\\r\\nhttp://farm3.static.flickr.com/2361/2050999339_b768fffdeb.jpg\\r\\nhttp://farm3.static.flickr.com/2183/2060979959_37c7eeb189.jpg\\r\\nhttp://www.cargolaw.com/images/disaster2003.Tricolor.3.GIF\\r\\nhttp://farm2.static.flickr.com/1047/1408868030_a046eb4242.jpg\\r\\nhttp://static.flickr.com/136/318437233_a4fbb0d15e.jpg\\r\\nhttp://farm2.static.flickr.com/1433/611958985_fbc3c88d97.jpg\\r\\nhttp://www.yellowbirdcharters.com/images/ship2.jpg\\r\\nhttp://farm4.static.flickr.com/3117/2608178893_75fa1eb88c.jpg\\r\\nhttp://farm2.static.flickr.com/1215/538089335_2b1be03d58.jpg\\r\\nhttp://www.indonesia-ship.com/pic/1755_1.jpg\\r\\nhttp://farm4.static.flickr.com/3123/2404649881_7d88268d5f.jpg\\r\\nhttp://farm1.static.flickr.com/185/375688474_acfb9e3ccb.jpg\\r\\nhttp://www.lennissite.com/My%20Ships/Corner.JPG\\r\\nhttp://research.usm.maine.edu/gulfofmaine-census/wp-content/images/Technology/platforms/Fig05NEFSCDelawareII.jpg\\r\\nhttp://farm1.static.flickr.com/154/381108004_b3d876590f.jpg\\r\\nhttp://news.mainemaritime.edu/index_files/stateofmaine.jpg\\r\\nhttp://farm3.static.flickr.com/2282/2249306393_8071a30676.jpg\\r\\nhttp://lh4.ggpht.com/_tvDJjOVmIR4/R4C9WFNHeUI/AAAAAAAACfc/YjMaKt0sXbo/Caribbean+Cruise+February+06+161.jpg\\r\\nhttp://farm3.static.flickr.com/2399/1898279385_bc2ad7e999.jpg\\r\\nhttp://mwcnews.net/images/stories/Asia/A/1/2/3/4/5/Taiwanese-ship.jpg\\r\\nhttp://www.cruisecompany.net/elegant/images/CezanneShip.jpg\\r\\nhttp://farm1.static.flickr.com/23/25475789_acb73d8ef1.jpg\\r\\nhttp://www.emsa.europa.eu/Pics/ships/mistra01.jpg\\r\\nhttp://www.oldendorff.com/news/newspages/images/Black%20Marlin.jpg\\r\\nhttp://static.flickr.com/107/290645859_ca3b4ccc4b.jpg\\r\\nhttp://farm1.static.flickr.com/69/188033321_0b30fbe540.jpg\\r\\nhttp://static.flickr.com/26/99902131_1292a982a8.jpg\\r\\nhttp://farm3.static.flickr.com/2304/2186733811_a09117de33.jpg\\r\\nhttp://farm1.static.flickr.com/232/512075538_b2eca55e6b.jpg\\r\\nhttp://farm4.static.flickr.com/3094/2640492724_9fd6894f06.jpg\\r\\nhttp://static.flickr.com/90/235757071_9d6a7c8e19.jpg\\r\\nhttp://www.ship-technology.com/projects/fesco/images/2-snow_boat.jpg\\r\\nhttp://farm3.static.flickr.com/2316/2529739994_dc61578245.jpg\\r\\nhttp://www.yuexinship.com/uploadfile/product/20080112110206.jpg\\r\\nhttp://farm3.static.flickr.com/2242/2487603455_37890e1476.jpg\\r\\nhttp://static.flickr.com/2315/2055806650_b2381bdfc9.jpg\\r\\nhttp://farm2.static.flickr.com/1234/908996902_720a2f9927.jpg\\r\\nhttp://english.kinca.cn/admin/edit/UploadFile/2007531211048183.jpg\\r\\nhttp://farm4.static.flickr.com/3214/2453600559_11340b6af7.jpg\\r\\nhttp://farm1.static.flickr.com/142/355225321_c26bc18658.jpg\\r\\nhttp://farm2.static.flickr.com/1102/908072033_dff9574b1d.jpg\\r\\nhttp://farm1.static.flickr.com/47/120052318_c8ddf56376.jpg\\r\\nhttp://farm4.static.flickr.com/3229/2724029367_9f0c80bb3e.jpg\\r\\nhttp://static.flickr.com/1249/1441702520_e32b26a567.jpg\\r\\nhttp://farm1.static.flickr.com/201/441935553_ca84fbdc42.jpg\\r\\nhttp://farm1.static.flickr.com/183/465267271_e4952872b0.jpg\\r\\nhttp://static.flickr.com/2390/1538449116_85af371e31.jpg\\r\\nhttp://farm4.static.flickr.com/3025/2627593553_d45c80f31d.jpg\\r\\nhttp://cache.daylife.com/imageserve/01538mGeRwbdI/610x.jpg\\r\\nhttp://static.flickr.com/2315/1507421325_a76d8170e9.jpg\\r\\nhttp://farm1.static.flickr.com/118/297100475_d1876e14dc.jpg\\r\\nhttp://farm2.static.flickr.com/1172/1458374585_5a5d4fe2e0.jpg\\r\\nhttp://farm3.static.flickr.com/2114/2483032347_1798cebfae.jpg\\r\\nhttp://farm3.static.flickr.com/2386/1754824661_0d0783ec25.jpg\\r\\nhttp://farm4.static.flickr.com/3047/2712774805_17cf864358.jpg\\r\\nhttp://farm4.static.flickr.com/3166/2460417270_1bb773d311.jpg\\r\\nhttp://farm4.static.flickr.com/3250/2693741876_9ab2b1f292.jpg\\r\\nhttp://farm4.static.flickr.com/3012/2664460940_def095b01d.jpg\\r\\nhttp://www.alexsurvey.com/images/vessel_2.gif\\r\\nhttp://bourbon-online.com/images/rescue-events/rokia-delmas.jpg\\r\\nhttp://farm1.static.flickr.com/154/355474551_ecdbd37182.jpg\\r\\nhttp://farm2.static.flickr.com/1029/798322881_7bfb0885a1.jpg\\r\\nhttp://farm2.static.flickr.com/1322/698952815_153bad0953.jpg\\r\\nhttp://www.novanewsnow.com/imgs/dynamique/articles/gros/Sealingboatsinking.jpg\\r\\nhttp://farm3.static.flickr.com/2344/2452035387_803f49e834.jpg\\r\\nhttp://farm4.static.flickr.com/3107/2640449613_bc8bd43b10.jpg\\r\\nhttp://farm3.static.flickr.com/2210/2180119269_f95fde17d2.jpg\\r\\nhttp://farm4.static.flickr.com/3260/2492729079_2aed94514b.jpg\\r\\nhttp://farm4.static.flickr.com/3049/2602123185_254749735f.jpg\\r\\nhttp://farm2.static.flickr.com/1322/1448111159_6697cad2a8.jpg\\r\\nhttp://farm3.static.flickr.com/2234/2167268318_f97cd11159.jpg\\r\\nhttp://farm3.static.flickr.com/2340/2478837754_bd1ff13944.jpg\\r\\nhttp://static.flickr.com/87/231698153_b3e27fc211.jpg\\r\\nhttp://farm2.static.flickr.com/1086/1407988275_5bec48ddfc.jpg\\r\\nhttp://farm3.static.flickr.com/2035/2359142241_1374c4cab1.jpg\\r\\nhttp://farm1.static.flickr.com/46/145308544_5bfb02f5ac.jpg\\r\\nhttp://farm3.static.flickr.com/2382/2444440896_e7b1877c3d.jpg\\r\\nhttp://farm4.static.flickr.com/3247/2600672325_defc777e1c.jpg\\r\\nhttp://farm1.static.flickr.com/252/453920158_06432cbf6c.jpg\\r\\nhttp://farm3.static.flickr.com/2190/2493576070_cb14aec71c.jpg\\r\\nhttp://farm4.static.flickr.com/3289/2675229236_b5b92e8cab.jpg\\r\\nhttp://farm2.static.flickr.com/1150/1350872014_8ed8e8d0de.jpg\\r\\nhttp://farm4.static.flickr.com/3081/2308114311_96a29f2352.jpg\\r\\nhttp://www.marinelawyers.net/images/Ship.jpg\\r\\nhttp://farm4.static.flickr.com/3253/2732608587_e86002f169.jpg\\r\\nhttp://static.flickr.com/2182/2081475080_2464ddb8a8.jpg\\r\\nhttp://farm4.static.flickr.com/3290/2521011534_fd41f4c824.jpg\\r\\nhttp://farm4.static.flickr.com/3230/2297949516_8c2bf8e5bf.jpg\\r\\nhttp://ab.mec.edu/jamestown/images/ship2.gif\\r\\nhttp://farm1.static.flickr.com/42/76710748_7e93e03d10.jpg\\r\\nhttp://farm4.static.flickr.com/3047/2391507499_945c8ea9b9.jpg\\r\\nhttp://static.flickr.com/1343/1466791763_4e9dc25c8f.jpg\\r\\nhttp://static.flickr.com/119/275720525_0d8c49c7f1.jpg\\r\\nhttp://farm3.static.flickr.com/2065/2069817534_2d12ddbe7c.jpg\\r\\nhttp://farm1.static.flickr.com/163/423544672_eace61952c.jpg\\r\\nhttp://farm1.static.flickr.com/88/244287747_ac321b6155.jpg?v=0\\r\\nhttp://farm3.static.flickr.com/2037/1983086479_6051919d14.jpg\\r\\nhttp://farm1.static.flickr.com/77/158353631_f1b1cbc5d2.jpg\\r\\nhttp://farm1.static.flickr.com/127/417205399_6b8d86aaee.jpg\\r\\nhttp://farm3.static.flickr.com/2011/2069826334_53629e74b8.jpg\\r\\nhttp://farm3.static.flickr.com/2070/2493577876_c999bde6cb.jpg\\r\\nhttp://farm3.static.flickr.com/2248/2337692254_aefc023fa3.jpg\\r\\nhttp://farm1.static.flickr.com/177/414673091_6ae879d7b8.jpg\\r\\nhttp://farm3.static.flickr.com/2193/2167812596_a33080a5f7.jpg\\r\\nhttp://farm1.static.flickr.com/55/145309878_f7145aa77a.jpg\\r\\nhttp://farm4.static.flickr.com/3118/2602311370_af7fed4d8b.jpg\\r\\nhttp://farm2.static.flickr.com/1357/612566352_c05a06ff57.jpg\\r\\nhttp://static.flickr.com/212/447330114_ca487596f2.jpg\\r\\nhttp://www.wilhelmsen.com/services/maritime/companies/barbership/AboutBarber/PressRoom/PressNews/PublishingImages/EukorVessel_1.jpg\\r\\nhttp://farm2.static.flickr.com/1138/917205220_5e6ae8b550.jpg\\r\\nhttp://farm3.static.flickr.com/2102/2391507831_ce7c36a7a7.jpg\\r\\nhttp://farm1.static.flickr.com/136/358352461_89b0e3d851.jpg\\r\\nhttp://farm3.static.flickr.com/2405/2308916468_2c50168f02.jpg\\r\\nhttp://farm3.static.flickr.com/2273/1803991460_a011438ed3.jpg\\r\\nhttp://farm1.static.flickr.com/146/422950227_9164ef7366.jpg\\r\\nhttp://farm3.static.flickr.com/2238/2451871197_f057594147.jpg\\r\\nhttp://farm3.static.flickr.com/2002/2435328930_c4308db337.jpg\\r\\nhttp://farm1.static.flickr.com/92/215737210_8920b391af.jpg\\r\\nhttp://farm1.static.flickr.com/121/311746536_6d052521df.jpg\\r\\nhttp://farm3.static.flickr.com/2044/2536989557_66ced7b53c.jpg\\r\\nhttp://farm1.static.flickr.com/183/442859874_8dba01a9ed.jpg\\r\\nhttp://farm1.static.flickr.com/232/500020024_1fcfbab8de.jpg\\r\\nhttp://www.photopumpkin.com/wp-content/uploads/1197.jpg\\r\\nhttp://farm1.static.flickr.com/171/423543663_8d24a2afe2.jpg\\r\\nhttp://farm4.static.flickr.com/3078/2493557338_8e500c6d7f.jpg\\r\\nhttp://farm1.static.flickr.com/49/120056567_78e9a39eff.jpg\\r\\nhttp://farm2.static.flickr.com/1132/1408870648_ab2981c682.jpg\\r\\nhttp://farm3.static.flickr.com/2030/2441420241_67ee7fd737.jpg\\r\\nhttp://farm4.static.flickr.com/3279/2366371180_269f9e7cf3.jpg\\r\\nhttp://farm3.static.flickr.com/2242/1562973902_f935eaa6d9.jpg\\r\\nhttp://farm4.static.flickr.com/3136/2602356250_9a0a85233c.jpg\\r\\nhttp://farm1.static.flickr.com/42/120060194_021fccad1f.jpg\\r\\nhttp://farm4.static.flickr.com/3076/2492733199_9c275b1c5f.jpg\\r\\nhttp://k53.pbase.com/v3/44/486444/2/49268314.DSC_0230resized.jpg\\r\\nhttp://bridgehands.com/Reviews/Bridge_Cruises/Rotterdam.jpg\\r\\nhttp://farm2.static.flickr.com/1291/836220883_eb1b965e29.jpg\\r\\nhttp://www.tollshipping.com.au/images/ship.jpg\\r\\nhttp://static.flickr.com/30/35688113_89a39086f0.jpg\\r\\nhttp://farm3.static.flickr.com/2076/2459581653_972e48b1ba.jpg\\r\\nhttp://www.ichikawa-design.com/ship_pic/fishing_vessel.jpg\\r\\nhttp://farm3.static.flickr.com/2031/2135499176_0c74616e52.jpg\\r\\nhttp://www.lkl.nl/images/gazprom.JPG\\r\\nhttp://www.energy-enviro.fi/images/newsletter/2_2005/8Ship%20ropeA4.jpg\\r\\nhttp://static.flickr.com/21/31892323_7f43bd422b.jpg\\r\\nhttp://newsimg.bbc.co.uk/media/images/44359000/jpg/_44359038_iceprince9_marinenat.jpg\\r\\nhttp://farm1.static.flickr.com/207/447347327_bcb5206ace.jpg\\r\\nhttp://farm1.static.flickr.com/56/145572316_ab302d1906.jpg\\r\\nhttp://farm3.static.flickr.com/2343/2287495341_6f4e320921.jpg\\r\\nhttp://static.flickr.com/80/232550229_0b478ee667.jpg\\r\\nhttp://www.destinationoceans.com/images/base/ship-Spirit-w630x300.jpg\\r\\nhttp://images.usatoday.com/money/_photos/2006/08/04/inside-cargo-ship.jpg\\r\\nhttp://farm1.static.flickr.com/43/120051489_23d072a468.jpg\\r\\nhttp://www.lcaships.com/ohio/6.jpg\\r\\nhttp://www.seamar-srl.com/container_ship2.jpg\\r\\nhttp://farm1.static.flickr.com/206/504238966_7147a78ca7.jpg\\r\\nhttp://farm1.static.flickr.com/40/74907375_4f8b77a4d9.jpg\\r\\nhttp://farm3.static.flickr.com/2322/2147118747_87856ed379.jpg\\r\\nhttp://www.italthaimarine.com/images/ship/images/HULL_NO._154_TRAINING_VESSEL-29.jpg\\r\\nhttp://farm3.static.flickr.com/2123/2336857553_754ca8205d.jpg\\r\\nhttp://www.customs.gov.au/webdata/resources/images/SOMPRU04.jpg\\r\\nhttp://farm3.static.flickr.com/2204/2061373476_459a008301.jpg\\r\\nhttp://farm3.static.flickr.com/2175/2539148552_a8f5d5658f.jpg\\r\\nhttp://static.flickr.com/192/508536506_bf3533a80b.jpg\\r\\nhttp://i62.photobucket.com/albums/h104/captainsvoyage/CValbum11/119-1943_IMG-1.jpg\\r\\nhttp://farm4.static.flickr.com/3233/2493566382_a08d370a28.jpg\\r\\nhttp://farm1.static.flickr.com/164/425314433_674a34a2ec.jpg\\r\\nhttp://farm1.static.flickr.com/24/37183286_437e0a0bd7.jpg\\r\\nhttp://farm2.static.flickr.com/1437/909155252_93b182be16.jpg\\r\\nhttp://farm3.static.flickr.com/2077/2518805127_d05d2bca46.jpg\\r\\nhttp://farm3.static.flickr.com/2024/2069023773_af5c62c963.jpg\\r\\nhttp://farm2.static.flickr.com/1026/1423403327_e1502f336b.jpg\\r\\nhttp://www.netmarine.net/bat/remorque/tenace/photo03.jpg\\r\\nhttp://farm2.static.flickr.com/1200/1438674381_f6b7eb9af5.jpg\\r\\nhttp://farm2.static.flickr.com/1326/864173524_8c1fd858f1.jpg\\r\\nhttp://farm1.static.flickr.com/63/201881209_76466ce99f.jpg\\r\\nhttp://farm3.static.flickr.com/2137/1899124460_62e2576cc7.jpg\\r\\nhttp://www.smit.com/sitefactor/public/downloads/images/GenmarKestrel.jpg\\r\\nhttp://farm2.static.flickr.com/1186/1462630841_2f6e53dec0.jpg\\r\\nhttp://farm3.static.flickr.com/2230/2407255749_902afd737f.jpg\\r\\nhttp://static.flickr.com/65/195932745_6b7f76d6e6.jpg\\r\\nhttp://farm2.static.flickr.com/1361/1344783378_2c93cc7acf.jpg\\r\\nhttp://static.flickr.com/48/188157378_f70174edb6.jpg\\r\\nhttp://farm4.static.flickr.com/3083/2409348875_b0d32daa1f.jpg\\r\\nhttp://www.svithoidtankers.com/system/inneh\\xc3\\xa5llsbilder/Bilder%20fartyg/060831_ET_English_Channel%5B1%5D.JPG\\r\\nhttp://farm3.static.flickr.com/2269/2369660909_cb9ea22624.jpg\\r\\nhttp://farm1.static.flickr.com/179/423544173_e32a85172e.jpg\\r\\nhttp://static.flickr.com/211/524458531_4615eb5bcb.jpg\\r\\nhttp://farm3.static.flickr.com/2126/2478049247_6dd0597095.jpg\\r\\nhttp://farm3.static.flickr.com/2306/2493570646_53d584e4e1.jpg\\r\\nhttp://farm2.static.flickr.com/1069/1028273416_babe2179cf.jpg\\r\\nhttp://farm3.static.flickr.com/2277/2265248577_754a531a2c.jpg\\r\\nhttp://farm2.static.flickr.com/1369/674924031_8f59611e80.jpg\\r\\nhttp://static.flickr.com/1392/1403169270_edf98c62d1.jpg\\r\\nhttp://www.automotoportal.com/media/images/vijesti/061216001.jpg\\r\\nhttp://farm4.static.flickr.com/3037/2692696816_4b3843fdb1.jpg\\r\\nhttp://farm1.static.flickr.com/172/447345911_a40f07753c.jpg\\r\\nhttp://www.painetworks.com/photos/co/co0251.JPG\\r\\nhttp://farm2.static.flickr.com/1044/799286518_e8333d403a.jpg\\r\\nhttp://farm1.static.flickr.com/146/360485150_821185ec2c.jpg\\r\\nhttp://www.marinetimes.ie/Assets/_archive_2004/0504_news_08_01.jpg\\r\\nhttp://www.zr88.com/images/ship.jpg\\r\\nhttp://www.mak-global.com/news/img/Manulani%20Container%20Vessel%202x8M20%202x9M20%20Genset.jpg\\r\\nhttp://farm4.static.flickr.com/3104/2493557612_08a5183ac3.jpg\\r\\nhttp://farm3.static.flickr.com/2208/2659167251_4e6e031d60.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/b/b8/Lash_vessel.JPG/300px-Lash_vessel.JPG\\r\\nhttp://farm2.static.flickr.com/1177/1062207942_a5dae71948.jpg\\r\\nhttp://farm3.static.flickr.com/2410/2348331155_7a6d7a775a.jpg\\r\\nhttp://koopvaardij.web-log.nl/koopvaardij/images/2008/02/23/naamloos.jpg\\r\\nhttp://farm3.static.flickr.com/2232/2189988373_99e3bc2c0a.jpg\\r\\nhttp://farm2.static.flickr.com/1089/751242405_28b6833d12.jpg\\r\\nhttp://farm4.static.flickr.com/3030/2461666131_d386b4ec63.jpg\\r\\nhttp://farm3.static.flickr.com/2344/2283483220_8384501280.jpg\\r\\nhttp://farm2.static.flickr.com/1379/1409205167_8c424a6e81.jpg\\r\\nhttp://farm1.static.flickr.com/145/388462375_5e59078e6c.jpg\\r\\nhttp://farm3.static.flickr.com/2226/2172054933_76a8ac0595.jpg\\r\\nhttp://farm4.static.flickr.com/3023/2636163779_41e0030854.jpg\\r\\nhttp://static.flickr.com/211/486298785_0802ca269d.jpg\\r\\nhttp://farm2.static.flickr.com/1414/1409823396_02f1647e33.jpg\\r\\nhttp://farm3.static.flickr.com/2351/2240952811_0087cf9ec3.jpg\\r\\nhttp://static.flickr.com/2022/2056798820_e9b41d08b0.jpg\\r\\nhttp://farm2.static.flickr.com/1305/536037680_0689f0a65e.jpg\\r\\nhttp://farm1.static.flickr.com/104/267378455_c156ca731f.jpg\\r\\nhttp://static.flickr.com/102/307176121_cebab47724.jpg\\r\\nhttp://farm4.static.flickr.com/3239/2308613274_c06fa6f581.jpg\\r\\nhttp://farm1.static.flickr.com/62/226484836_f4451e1286.jpg\\r\\nhttp://farm4.static.flickr.com/3149/2483591978_2fddc43c2f.jpg\\r\\nhttp://farm1.static.flickr.com/94/233107999_588dbe5758.jpg\\r\\nhttp://farm1.static.flickr.com/203/449027701_689bacba2b.jpg\\r\\nhttp://www.imagevortex.com/stock/photos/thumbnail/8/0/2/ship_boat_vessel_transportation_travel-802055708.jpg\\r\\nhttp://static.flickr.com/93/231690099_1d2b5d024f.jpg\\r\\nhttp://farm2.static.flickr.com/1294/908811002_657d73e67e.jpg\\r\\nhttp://farm3.static.flickr.com/2069/2381885939_2c63b76a09.jpg\\r\\nhttp://www.ci.homer.ak.us/Mvc-022f.jpg\\r\\nhttp://farm2.static.flickr.com/1315/987669184_4033b71cb9.jpg\\r\\nhttp://farm2.static.flickr.com/1127/908352589_f73c9c791a.jpg\\r\\nhttp://farm1.static.flickr.com/123/425281685_c607a1d262.jpg\\r\\nhttp://farm3.static.flickr.com/2281/2486678998_11279f03f1.jpg\\r\\nhttp://farm3.static.flickr.com/2415/2311732245_fdbe730091.jpg\\r\\nhttp://farm1.static.flickr.com/102/289992402_2938323b92.jpg\\r\\nhttp://farm1.static.flickr.com/126/369351508_faf08c6375.jpg\\r\\nhttp://farm4.static.flickr.com/3099/2554287871_8bae1fb9b1.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2584629639_ce1742caaf.jpg\\r\\nhttp://farm2.static.flickr.com/1229/1407985335_f9d92373e0.jpg\\r\\nhttp://farm4.static.flickr.com/3075/2627593683_a1044d374f.jpg\\r\\nhttp://farm1.static.flickr.com/46/120060553_ab3adc0e7a.jpg\\r\\nhttp://farm1.static.flickr.com/132/423544011_b2aad403bd.jpg\\r\\nhttp://farm1.static.flickr.com/55/145565017_34b50c96de.jpg\\r\\nhttp://farm1.static.flickr.com/31/35386449_9a5b63d350.jpg\\r\\nhttp://farm4.static.flickr.com/3259/2633994995_bbe61df9b6.jpg\\r\\nhttp://farm4.static.flickr.com/3289/2308540682_74817b044d.jpg\\r\\nhttp://www.worldofwallpapers.nuche.org/content/boating/1024/fishing-vessel-in-storm-desktops.jpg\\r\\nhttp://farm3.static.flickr.com/2285/1560692198_4737e7d584.jpg\\r\\nhttp://farm1.static.flickr.com/55/145302518_bfcf82e8ca.jpg\\r\\nhttp://farm1.static.flickr.com/48/138503577_184f623586.jpg\\r\\nhttp://www.greenpeace.org.uk/files/images/migrated/MultimediaFiles/Live/Image/7358.jpg\\r\\nhttp://www.indymedia.org.uk/images/2006/01/331045.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2478005937_9a2559eeb7.jpg\\r\\nhttp://bourbon-online.com/images/rescue-events/baiedeseine_happy.jpg\\r\\nhttp://farm4.static.flickr.com/3171/2352879683_c38e45ab6e.jpg\\r\\nhttp://www.highcountrypassage.com/web-bin/tripPage/230/Nez_Ship.jpg\\r\\nhttp://farm3.static.flickr.com/2103/2525991416_8f56052681.jpg\\r\\nhttp://farm3.static.flickr.com/2190/2365945386_4e3455afdd.jpg\\r\\nhttp://farm1.static.flickr.com/23/25475793_4825868af1.jpg\\r\\nhttp://farm4.static.flickr.com/3151/2454426918_fe24073124.jpg\\r\\nhttp://farm1.static.flickr.com/184/382955436_d8d62477d3.jpg\\r\\nhttp://static.flickr.com/43/121335594_538788d9dc.jpg\\r\\nhttp://static.flickr.com/2417/2080689931_0d724ebae7.jpg\\r\\nhttp://www.saratogamuseum.org/images/cv-60a2.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2640494356_afd6568662.jpg\\r\\nhttp://farm3.static.flickr.com/2031/2051782278_5551856419.jpg\\r\\nhttp://static.flickr.com/98/232539594_03cf11f02f.jpg\\r\\nhttp://bridgehands.com/Reviews/Bridge_Cruises/Queen_Elizabeth_II.jpg\\r\\nhttp://farm1.static.flickr.com/119/306516984_b712a20ff9.jpg\\r\\nhttp://farm1.static.flickr.com/169/445326229_ba409b1f28.jpg\\r\\nhttp://www.remontowa.com.pl/en_db/img/pagehtml_129270985644bc93aee06d8.jpg\\r\\nhttp://farm1.static.flickr.com/79/233459944_d110b9c176.jpg\\r\\nhttp://farm1.static.flickr.com/47/138502952_cdf0c6262d.jpg\\r\\nhttp://farm1.static.flickr.com/39/118622081_b6c0b9ef55.jpg\\r\\nhttp://static.flickr.com/2250/2063451336_bd7906b1b1.jpg\\r\\nhttp://farm1.static.flickr.com/173/434905255_4939686fb3.jpg\\r\\nhttp://farm3.static.flickr.com/2014/2538645423_e1e8a667f6.jpg\\r\\nhttp://farm1.static.flickr.com/165/397310280_4f9d49caa3.jpg\\r\\nhttp://farm1.static.flickr.com/168/437999446_38281c84a0.jpg\\r\\nhttp://static.flickr.com/2368/2077116703_be08dcc33b.jpg\\r\\nhttp://farm1.static.flickr.com/161/423543687_edfda8acc2.jpg\\r\\nhttp://www.vos.noaa.gov/MWL/aug_06/Images/shipwreck.jpg\\r\\nhttp://farm4.static.flickr.com/3015/2729957586_b40b801f24.jpg\\r\\nhttp://static.flickr.com/2118/2041510746_3a83386b31.jpg\\r\\nhttp://www.soest.hawaii.edu/SOEST_News/PressReleases/AGU2003/KM%20Arrival%20by%20Bob%20Chin%20008.jpg\\r\\nhttp://farm3.static.flickr.com/2335/2536642893_7696022993.jpg\\r\\nhttp://farm1.static.flickr.com/218/512118311_8f56366a7d.jpg\\r\\nhttp://farm3.static.flickr.com/2007/2389399796_3a3aca0d5e.jpg\\r\\nhttp://farm3.static.flickr.com/2049/2069033679_7e4a4650f3.jpg\\r\\nhttp://farm4.static.flickr.com/3151/2493567070_f11316d79c.jpg\\r\\nhttp://www.msc.navy.mil/sealift/2002/November/graphics/0209193.jpg\\r\\nhttp://farm4.static.flickr.com/3265/2526082582_f6dbe9690c.jpg\\r\\nhttp://static.flickr.com/2417/1799352686_41953b144a.jpg\\r\\nhttp://shipbuilding.ru/images/docs/774.jpg\\r\\nhttp://www.studentsonice.com/whale2004/assets/images/Sedna_bow.jpg\\r\\nhttp://farm4.static.flickr.com/3112/2374767528_d451d470d0.jpg\\r\\nhttp://farm4.static.flickr.com/3087/2537460496_c98dd031d4.jpg\\r\\nhttp://farm1.static.flickr.com/70/169553834_b9aa48cf0b.jpg\\r\\nhttp://static.flickr.com/80/231643916_52533dadc2.jpg\\r\\nhttp://www.mnr.gov.on.ca/MNR/biodiversity/images/iStock_large_ship.jpg\\r\\nhttp://farm4.static.flickr.com/3045/2717710091_366fc44417.jpg\\r\\nhttp://farm4.static.flickr.com/3245/2536391519_36bde92db1.jpg\\r\\nhttp://farm1.static.flickr.com/59/192656788_86645dbfaf.jpg\\r\\nhttp://farm2.static.flickr.com/1436/751242349_9676f6c92d.jpg\\r\\nhttp://www.ships-for-sale.com/ships%20for%20sale%20pics/small_container_ship_for_sale2.jpg\\r\\nhttp://farm1.static.flickr.com/220/535932527_c7dcb0e9a9.jpg\\r\\nhttp://farm3.static.flickr.com/2161/2110904232_1d5acb4cc5.jpg\\r\\nhttp://www.ymcakl.com/images/youth/Doulos_Ship.JPG\\r\\nhttp://farm2.static.flickr.com/1280/865687805_248a76a36d.jpg\\r\\nhttp://farm1.static.flickr.com/194/457872408_28504dbb63.jpg\\r\\nhttp://farm3.static.flickr.com/2322/2069818590_4b151eab95.jpg\\r\\nhttp://static.flickr.com/116/277212315_de75e9089e.jpg\\r\\nhttp://farm2.static.flickr.com/1408/730833157_206df4e2e2.jpg\\r\\nhttp://static.flickr.com/2157/2115398002_ecc66dfc3b.jpg\\r\\nhttp://farm1.static.flickr.com/15/68921551_d4e2bdfa11.jpg\\r\\nhttp://farm1.static.flickr.com/215/512117783_6b5062dc78.jpg\\r\\nhttp://farm3.static.flickr.com/2001/2370470882_4958e2bafc.jpg\\r\\nhttp://static.flickr.com/95/232550226_eda2594653.jpg\\r\\nhttp://static.flickr.com/96/216557963_5422283a9c.jpg\\r\\nhttp://farm2.static.flickr.com/1161/534709835_23fcfd2d64.jpg\\r\\nhttp://farm3.static.flickr.com/2052/2337692288_b1fdf586ba.jpg\\r\\nhttp://farm3.static.flickr.com/2182/2487596531_a287c7ed41.jpg\\r\\nhttp://www.montserrat-newsletter.com/archives/ship.gif\\r\\nhttp://www.highnorth.no/bilder/solo.jpg\\r\\nhttp://www.expertcruiser.com/wp-content/uploads/ssgrenada2.jpg\\r\\nhttp://farm4.static.flickr.com/3203/2705956250_88b3581de6.jpg\\r\\nhttp://farm3.static.flickr.com/2360/2406238957_1f1b70fe1c.jpg\\r\\nhttp://farm1.static.flickr.com/169/423544699_7984e5ac70.jpg\\r\\nhttp://farm2.static.flickr.com/1123/1424283008_7af60c9e20.jpg\\r\\nhttp://www.defenselink.mil/photos/Feb1996/960224-N-6240R-005.jpg\\r\\nhttp://photos.signonsandiego.com/gallery1.5/albums/beachedship/JG_15StuckShip253571x004.jpg\\r\\nhttp://cache.daylife.com/imageserve/00XfcjFcMg5Xp/610x.jpg\\r\\nhttp://static.flickr.com/94/255484201_d16320984a.jpg\\r\\nhttp://www.history.com/marquees/mayflower/images/mayflower.jpg\\r\\nhttp://farm4.static.flickr.com/3100/2730188512_975ddfc113.jpg\\r\\nhttp://farm4.static.flickr.com/3178/2491959523_87280401d3.jpg\\r\\nhttp://schools-wikipedia.org/images/109/10924.jpg\\r\\nhttp://farm1.static.flickr.com/47/176214788_4541a08cdf.jpg\\r\\nhttp://farm3.static.flickr.com/2359/2050999419_8749253406.jpg\\r\\nhttp://farm4.static.flickr.com/3034/2627486310_7f3ebbf96a.jpg\\r\\nhttp://farm4.static.flickr.com/3156/2577581340_851d751a7b.jpg\\r\\nhttp://farm4.static.flickr.com/3061/2310080638_3927b595df.jpg\\r\\nhttp://farm3.static.flickr.com/2117/2337692390_f253430053.jpg\\r\\nhttp://static.flickr.com/95/231698155_ce360d360d.jpg\\r\\nhttp://static.flickr.com/87/232513315_de2a4c54f6.jpg\\r\\nhttp://farm3.static.flickr.com/2260/2147118555_ecd9001887.jpg\\r\\nhttp://farm2.static.flickr.com/1170/653843307_5de575e1d7.jpg\\r\\nhttp://www.cep.unep.org/pubs/cepnews/v17n2/ship.JPG\\r\\nhttp://farm3.static.flickr.com/2393/2522541785_e8a0064fdb.jpg\\r\\nhttp://farm3.static.flickr.com/2271/2499379248_198af9e98f.jpg\\r\\nhttp://farm3.static.flickr.com/2270/2371722691_ba656aceab.jpg\\r\\nhttp://farm2.static.flickr.com/1246/905716669_9966a7db01.jpg\\r\\nhttp://farm3.static.flickr.com/2414/2483408139_04d605f800.jpg\\r\\nhttp://farm4.static.flickr.com/3146/2609002518_788305f5a9.jpg\\r\\nhttp://www.worldwaterways.com/VesselImages/PageSize/1047_ship-1.jpg\\r\\nhttp://farm2.static.flickr.com/1004/526703899_d5aad1bf49.jpg\\r\\nhttp://farm4.static.flickr.com/3158/2595582547_e490715690.jpg\\r\\nhttp://farm2.static.flickr.com/1241/1130268732_8700bd0695.jpg\\r\\nhttp://farm2.static.flickr.com/1298/614777563_e61844428d.jpg\\r\\nhttp://www.port2port.com/_Uploads/1443Evergreen_Ship5.jpg\\r\\nhttp://www.akermarine.com/images/SA-85m.jpg\\r\\nhttp://static.flickr.com/94/231667284_13653bbd88.jpg\\r\\nhttp://farm3.static.flickr.com/2138/2253890952_acda6abe22.jpg\\r\\nhttp://farm1.static.flickr.com/53/145543192_503b7dcccd.jpg\\r\\nhttp://farm2.static.flickr.com/1120/1402287435_f8c6c1782d.jpg\\r\\nhttp://farm2.static.flickr.com/1193/1459240632_e0e768aab4.jpg\\r\\nhttp://farm4.static.flickr.com/3097/2491959683_fecb1d4b0f.jpg\\r\\nhttp://farm1.static.flickr.com/223/447349125_fd1eba17e7.jpg\\r\\nhttp://media.npr.org/news/images/2007/sept/06/arctic_ship540.jpg\\r\\nhttp://farm1.static.flickr.com/80/249544914_8c827db430.jpg\\r\\nhttp://farm4.static.flickr.com/3133/2609299277_3c085825b3.jpg\\r\\nhttp://farm4.static.flickr.com/3050/2492780994_1098c3c342.jpg\\r\\nhttp://farm4.static.flickr.com/3226/2656328874_04ea30e8e5.jpg\\r\\nhttp://www.providence.edu/polisci/students/megaport/images/marchen.jpg\\r\\nhttp://farm1.static.flickr.com/199/499435955_63a1c2e6e5.jpg\\r\\nhttp://farm2.static.flickr.com/1076/630396043_e36fb81c8e.jpg\\r\\nhttp://static.flickr.com/1194/1395050238_004a88eae7.jpg\\r\\nhttp://farm3.static.flickr.com/2264/2396913818_c64fa7ffe6.jpg\\r\\nhttp://farm4.static.flickr.com/3153/2684219508_5802d4819e.jpg\\r\\nhttp://farm3.static.flickr.com/2102/2233516551_99d5e5261f.jpg\\r\\nhttp://farm3.static.flickr.com/2388/2109386801_ff67f78142.jpg\\r\\nhttp://archive.greenpeace.org/~nuclear/bnfl/images/plutoship.jpg\\r\\nhttp://farm4.static.flickr.com/3279/2406236669_2a75e4b39a.jpg\\r\\nhttp://static.flickr.com/81/231940104_be24089ea7.jpg\\r\\nhttp://www.aniketsolutions.com/Images/marineship.PNG\\r\\nhttp://static.flickr.com/227/464960289_c737e11828.jpg\\r\\nhttp://www.simplonpc.co.uk/WhatShip/Lazarettschiff_Genua_b.jpg\\r\\nhttp://farm1.static.flickr.com/29/141148192_15afa780b9.jpg\\r\\nhttp://farm4.static.flickr.com/3210/2525215371_e5b6201c2c.jpg\\r\\nhttp://farm4.static.flickr.com/3170/2574923566_6a040e542a.jpg\\r\\nhttp://farm4.static.flickr.com/3113/2373795822_6d56b2a5c6.jpg\\r\\nhttp://farm4.static.flickr.com/3291/2595585303_84f900fb31.jpg\\r\\nhttp://farm4.static.flickr.com/3095/2314839831_79eae572eb.jpg\\r\\nhttp://farm3.static.flickr.com/2409/2408065104_7c7e41a66b.jpg\\r\\nhttp://www.globalsecurity.org/military/systems/ship/images/LARC-pic1.jpg\\r\\nhttp://farm3.static.flickr.com/2184/2234153096_d5572d1491.jpg\\r\\nhttp://farm2.static.flickr.com/1232/1458371127_7ca0837158.jpg\\r\\nhttp://soundwaves.usgs.gov/2006/01/KatrinaShipLG.jpg\\r\\nhttp://farm4.static.flickr.com/3218/2338554737_dca6af620f.jpg\\r\\nhttp://farm2.static.flickr.com/1144/611332836_b6e6576a33.jpg\\r\\nhttp://farm2.static.flickr.com/1268/674865675_1e80d35c23.jpg\\r\\nhttp://farm2.static.flickr.com/1286/537608262_2d58bff785.jpg\\r\\nhttp://static.flickr.com/1260/798386169_944e218e04.jpg\\r\\nhttp://farm3.static.flickr.com/2094/2239347565_690c346461.jpg\\r\\nhttp://farm3.static.flickr.com/2051/2498892120_409bf0e453.jpg\\r\\nhttp://farm3.static.flickr.com/2242/2265281065_ba2ecfa895.jpg\\r\\nhttp://www.nearfield.com/~dan/sports/bike/beach/longbch/D2852.jpg\\r\\nhttp://farm1.static.flickr.com/89/230929173_7b2f391438.jpg\\r\\nhttp://farm3.static.flickr.com/2089/2308913120_408f7b5dc8.jpg\\r\\nhttp://farm3.static.flickr.com/2092/2201758830_151fb28d4c.jpg\\r\\nhttp://farm4.static.flickr.com/3220/2382650149_ff939457e5.jpg\\r\\nhttp://farm3.static.flickr.com/2146/2167812468_a2bc191e08.jpg\\r\\nhttp://farm1.static.flickr.com/211/460583918_7d2c607f07.jpg\\r\\nhttp://farm1.static.flickr.com/52/138504031_32d5203f9f.jpg\\r\\nhttp://farm4.static.flickr.com/3105/2610786179_bf22ff9880.jpg\\r\\nhttp://farm1.static.flickr.com/29/35385709_0a43aa9223.jpg\\r\\nhttp://farm3.static.flickr.com/2418/2374250563_011ab1bf62.jpg\\r\\nhttp://farm4.static.flickr.com/3002/2656321234_fc62d0b3d5.jpg\\r\\nhttp://www.rvmtmitchell.com/graphics/ship_canal.jpg\\r\\nhttp://farm4.static.flickr.com/3257/2725430937_7038e9d1e8.jpg\\r\\nhttp://www.mec-ltd.co.uk/ship2.jpg\\r\\nhttp://www.heyn.co.uk/Shipping/images/macandrew-vessel.jpg\\r\\nhttp://www.valderhaug.no/offshore/Skandi%20Rona%201RQ-18.jpg\\r\\nhttp://farm3.static.flickr.com/2107/2557338819_2ef04fd9c0.jpg\\r\\nhttp://homepages.enterprise.net/hyland/maps/interest/05_jumbo_s.jpg\\r\\nhttp://farm1.static.flickr.com/64/183955439_7852146750.jpg\\r\\nhttp://farm1.static.flickr.com/118/289992181_d34cf61ca0.jpg\\r\\nhttp://farm3.static.flickr.com/2054/2201493832_3cd8d5103c.jpg\\r\\nhttp://farm3.static.flickr.com/2176/1978501533_4f9f53bdba.jpg\\r\\nhttp://farm3.static.flickr.com/2275/1979326248_38ce7d7d09.jpg\\r\\nhttp://farm4.static.flickr.com/3170/2724995433_304571a705.jpg\\r\\nhttp://lejla.tunapuff.net/wp-content/blogs.dir/349/files/2007/06/ship.jpg\\r\\nhttp://farm2.static.flickr.com/1403/909038264_79292b9fac.jpg\\r\\nhttp://www.directshipservices.com/images/Cargo%20vessel_0.jpg\\r\\nhttp://farm2.static.flickr.com/1188/760189412_752c3681d5.jpg\\r\\nhttp://farm3.static.flickr.com/2013/1802065164_498c82fab0.jpg\\r\\nhttp://farm3.static.flickr.com/2242/2350668003_1aedb2dbf4.jpg\\r\\nhttp://static.flickr.com/53/188977860_e7d22c9081.jpg\\r\\nhttp://www.chinadaily.com.cn/english/doc/2004-12/15/xin_37120115104623444004.jpg\\r\\nhttp://www.myc.org/images/2006/admiral's-launch.jpg\\r\\nhttp://farm3.static.flickr.com/2228/2217571026_1c040f5cb0.jpg\\r\\nhttp://www.warsailors.com/norships/sorlandet.jpg\\r\\nhttp://static.flickr.com/90/231690096_ea88e9ac72.jpg\\r\\nhttp://farm3.static.flickr.com/2053/2223986969_e216199275.jpg\\r\\nhttp://farm4.static.flickr.com/3189/2718530264_802a950f80.jpg\\r\\nhttp://www.netpublikationer.dk/DFFE/5951/images/rap_electronic_transport_documentation_2x37x1.jpg\\r\\nhttp://farm4.static.flickr.com/3121/2627482802_cbe5b8b88d.jpg\\r\\nhttp://farm3.static.flickr.com/2192/2118528314_5bb102f693.jpg\\r\\nhttp://farm3.static.flickr.com/2357/2309247702_34d77aa7ac.jpg\\r\\nhttp://static.flickr.com/2257/1977902317_a87289799a.jpg\\r\\nhttp://farm1.static.flickr.com/208/491737051_82d3d2e636.jpg\\r\\nhttp://www.oilpubs.com/_ni/ulstein.jpg\\r\\nhttp://static.flickr.com/2129/2121148190_d3d2d8a448.jpg\\r\\nhttp://farm1.static.flickr.com/62/179483980_da2441afd1.jpg\\r\\nhttp://www.seabird-marine.com/ship1.gif\\r\\nhttp://farm4.static.flickr.com/3148/2640491244_f5ce425d0e.jpg\\r\\nhttp://farm2.static.flickr.com/1067/1343241356_c39000950d.jpg\\r\\nhttp://static.flickr.com/1034/1061510564_2e1f2d9102.jpg\\r\\nhttp://farm3.static.flickr.com/2033/1978499537_9f54185a58.jpg\\r\\nhttp://farm3.static.flickr.com/2381/2103169707_48639b31b0.jpg\\r\\nhttp://i38.photobucket.com/albums/e109/Raivynns/jose_gasparilla_ship.png\\r\\nhttp://farm3.static.flickr.com/2278/2239408973_009474c917.jpg\\r\\nhttp://www.environment.gov.au/coasts/pollution/antifouling/images/ship_s.jpg\\r\\nhttp://www.worldwideseafoods.ca/images/fleet_boldperformance.jpg\\r\\nhttp://farm3.static.flickr.com/2060/2626674725_975cf7f773.jpg\\r\\nhttp://farm3.static.flickr.com/2283/1517153376_d086a3f55a.jpg\\r\\nhttp://farm4.static.flickr.com/3200/2481798543_f76c42fca0.jpg\\r\\nhttp://farm3.static.flickr.com/2349/2370921915_ddac030406.jpg\\r\\nhttp://farm1.static.flickr.com/44/138502844_764fffd900.jpg\\r\\nhttp://www.marcon.com/library/sales_reports/2007_2009/2007Sales/OceanVeritas.jpg\\r\\nhttp://farm1.static.flickr.com/73/207049524_8a7d511f25.jpg\\r\\nhttp://static.flickr.com/201/524458663_2721a72508.jpg\\r\\nhttp://lh3.ggpht.com/_Z4TXFpxoEuU/SGnGvSQsB9I/AAAAAAAAAg0/FzKjUPdPJKw/IMG_6486.JPG\\r\\nhttp://farm1.static.flickr.com/237/515550861_39a7a544e8.jpg\\r\\nhttp://farm2.static.flickr.com/1153/1342047740_b8639a8c96.jpg\\r\\nhttp://farm3.static.flickr.com/2035/2525264755_e9d6194b64.jpg\\r\\nhttp://farm1.static.flickr.com/42/120051826_c6cd0bfc42.jpg\\r\\nhttp://farm1.static.flickr.com/158/423544592_ccdc979eea.jpg\\r\\nhttp://farm1.static.flickr.com/187/468322192_7e4cb119c2.jpg\\r\\nhttp://farm1.static.flickr.com/6/5764846_45024d12f6.jpg\\r\\nhttp://static.flickr.com/51/137124335_7407a6eca9.jpg\\r\\nhttp://static.flickr.com/42/99902130_6a3cc2a68d.jpg\\r\\nhttp://farm4.static.flickr.com/3173/2729125983_43bbcd7449.jpg\\r\\nhttp://www.spacewar.com/images/littoral-combat-ship-freedom-lcs1-bg.jpg\\r\\nhttp://ships.greenet.pl/photo/borussia_pbs.jpg\\r\\nhttp://www.jamstec.org/image/ship-yokosuka.jpg\\r\\nhttp://msnbcmedia3.msn.com/j/ap/tok80202170048.hmedium.jpg\\r\\nhttp://www.adcp.com/images/military_vessel.jpg\\r\\nhttp://farm1.static.flickr.com/165/423543550_51be0c8682.jpg\\r\\nhttp://farm3.static.flickr.com/2208/2237664545_d420eaa0b9.jpg\\r\\nhttp://ropacorp.com/images/OfiShip.jpg\\r\\nhttp://farm1.static.flickr.com/139/345479194_6ceee30c16.jpg\\r\\nhttp://lh3.ggpht.com/_D1cMk4w94mg/Rq6nTPwsP4I/AAAAAAAAAMc/zFmvvya1zng/IMG_0562.JPG\\r\\nhttp://farm3.static.flickr.com/2071/2264736195_a11fd4fe70.jpg\\r\\nhttp://farm1.static.flickr.com/212/481565115_9ea60daab2.jpg\\r\\nhttp://www.cruiseweb.com/rssc-imagelibrary/EXP_hor_low555.jpg\\r\\nhttp://farm1.static.flickr.com/74/226369393_13fa872f4b.jpg\\r\\nhttp://www.smartmarine-engg.com/images/jaya.jpg\\r\\nhttp://farm2.static.flickr.com/1186/909148198_376e253b25.jpg\\r\\nhttp://farm1.static.flickr.com/76/190217795_0f408b1bf4.jpg\\r\\nhttp://www.cnn.com/US/9911/01/egyptair.03/new.search.ship.jpg\\r\\nhttp://farm3.static.flickr.com/2026/2172811002_09a3fbfec9.jpg\\r\\nhttp://farm3.static.flickr.com/2287/2407246981_42c09f5799.jpg\\r\\nhttp://farm1.static.flickr.com/208/508330683_ef8bde5bf6.jpg\\r\\nhttp://lh4.ggpht.com/_6wTDRL3WDHw/RyQ-ZHB7WVI/AAAAAAAAAM0/yNgCls3-tP8/Shanghai+123.jpg\\r\\nhttp://gcaptain.com/maritime/blog/wp-content/uploads/2008/04/boats-on-ship.jpg\\r\\nhttp://vesseltrax.com/images/garden_main.jpg\\r\\nhttp://farm3.static.flickr.com/2359/2088879659_8c2ef1a9c6.jpg\\r\\nhttp://farm3.static.flickr.com/2263/2493550298_2dce838904.jpg\\r\\nhttp://farm3.static.flickr.com/2053/1502533095_3da237ea2c.jpg\\r\\nhttp://farm2.static.flickr.com/1263/1465188463_c918a5c72d.jpg\\r\\nhttp://farm4.static.flickr.com/3118/2626090538_1d41a9a244.jpg\\r\\nhttp://farm4.static.flickr.com/3063/2717711593_ab83dff4bb.jpg\\r\\nhttp://www.nap.usace.army.mil/mdc/images/stbd%20profile.JPG\\r\\nhttp://farm1.static.flickr.com/87/233780775_a907ca769e.jpg\\r\\nhttp://www.warshipsifr.com/media/burningShip2.jpg\\r\\nhttp://farm1.static.flickr.com/52/139958936_54c62ced5f.jpg\\r\\nhttp://farm3.static.flickr.com/2393/2170342903_1f7b6f968c.jpg\\r\\nhttp://farm2.static.flickr.com/1004/908912364_c1e90878b8.jpg\\r\\nhttp://img.dailymail.co.uk/i/pix/2007/08_01/arcticship_468x338.jpg\\r\\nhttp://farm1.static.flickr.com/161/355484307_bf5bc86b9a.jpg\\r\\nhttp://www.sirimarine.nl/images/HL2pic4.jpg\\r\\nhttp://www.capeharrison.com/images/spirit.jpg\\r\\nhttp://farm1.static.flickr.com/145/423544764_0c1653316d.jpg\\r\\nhttp://farm3.static.flickr.com/2266/2266059410_434f38c256.jpg\\r\\nhttp://static.flickr.com/96/243067750_5a03a5852c.jpg\\r\\nhttp://farm3.static.flickr.com/2227/2536638073_0eb08a8df5.jpg\\r\\nhttp://farm4.static.flickr.com/3068/2470188056_5bfd120e9b.jpg\\r\\nhttp://farm1.static.flickr.com/222/506449959_ee5547ee05.jpg\\r\\nhttp://farm2.static.flickr.com/1312/964593323_10f8ac322d.jpg\\r\\nhttp://farm4.static.flickr.com/3248/2713077827_3a46bbc454.jpg\\r\\nhttp://farm3.static.flickr.com/2280/2051783458_1090c1997c.jpg\\r\\nhttp://forms.rnicol.com/sites/22232/images/Taras-Shevchenko_ship_profile.jpg\\r\\nhttp://farm2.static.flickr.com/1288/637100731_7a1822cd69.jpg\\r\\nhttp://static.flickr.com/56/179386368_c990345a5a.jpg\\r\\nhttp://static.flickr.com/2412/2051374169_407658f5c2.jpg\\r\\nhttp://static.flickr.com/93/232492989_896599fa03.jpg\\r\\nhttp://farm4.static.flickr.com/3186/2725877376_e437b71e7e.jpg\\r\\nhttp://static.flickr.com/23/29314372_43de2f2e8d.jpg\\r\\nhttp://farm4.static.flickr.com/3243/2686150490_4a50a55d07.jpg\\r\\nhttp://farm1.static.flickr.com/47/145543199_d3298fcc7e.jpg\\r\\nhttp://farm2.static.flickr.com/1194/1164982747_68fba489d0.jpg\\r\\nhttp://nstp.ucis.dal.ca/images/zouppa/userimages/ship.jpg\\r\\nhttp://farm1.static.flickr.com/44/120057929_5810028950.jpg\\r\\nhttp://farm4.static.flickr.com/3223/2596418132_ddde78200b.jpg\\r\\nhttp://www.expertcruiser.com/wp-content/uploads/cwnantucket.jpg\\r\\nhttp://static.flickr.com/82/232513016_435c102b82.jpg\\r\\nhttp://www.elgrecomanila.com/shipchandlers/images/JuanS.jpg\\r\\nhttp://farm4.static.flickr.com/3046/2656327346_dfcaf9903a.jpg\\r\\nhttp://farm4.static.flickr.com/3181/2700506085_fdfddd1761.jpg\\r\\nhttp://farm3.static.flickr.com/2406/2061110732_8de8afa3b7.jpg\\r\\nhttp://farm3.static.flickr.com/2046/2116299488_b97c205ffb.jpg\\r\\nhttp://farm1.static.flickr.com/19/120052210_011270638a.jpg\\r\\nhttp://farm4.static.flickr.com/3221/2491959569_a763a0fe67.jpg\\r\\nhttp://farm3.static.flickr.com/2059/2537464418_9f4526afdc.jpg\\r\\nhttp://farm3.static.flickr.com/2140/2520230273_9868b70707.jpg\\r\\nhttp://farm3.static.flickr.com/2308/2636994466_e1b78208cd.jpg\\r\\nhttp://farm1.static.flickr.com/56/138503749_e4d3610b17.jpg\\r\\nhttp://www.naval-technology.com/features/feature_images/navaloil/2-uss-independence.jpg\\r\\nhttp://www.naudiassoc.com/img/ship.jpg\\r\\nhttp://farm3.static.flickr.com/2094/2381720735_41bc9a9149.jpg\\r\\nhttp://farm4.static.flickr.com/3084/2482609912_9485093dc9.jpg\\r\\nhttp://cache.daylife.com/imageserve/04EW6VAaOw3vT/610x.jpg\\r\\nhttp://farm2.static.flickr.com/1070/909174994_c649f7b60d.jpg\\r\\nhttp://farm1.static.flickr.com/171/479719856_be1f50fb76.jpg\\r\\nhttp://www.hkmw.com.hk/UpFiles/Images/2004/9/22/175743.jpg\\r\\nhttp://farm1.static.flickr.com/79/283240831_09b18addf6.jpg\\r\\nhttp://image.guim.co.uk/Guardian/world/gallery/2007/nov/23/1/GD5416356@Mandatory-Credit-Phot-8793.jpg\\r\\nhttp://www.wright-international.com/rv2.jpg\\r\\nhttp://farm3.static.flickr.com/2360/1797965586_1aae74e28f.jpg\\r\\nhttp://image.guim.co.uk/sys-images/Environment/Pix/pictures/2008/01/15/whaling460.jpg\\r\\nhttp://farm3.static.flickr.com/2186/2482380763_8010e9cc00.jpg\\r\\nhttp://farm4.static.flickr.com/3143/2608167405_090dfeef07.jpg\\r\\nhttp://farm1.static.flickr.com/4/7120493_952750e014.jpg\\r\\nhttp://farm4.static.flickr.com/3250/2497668224_2e843998c9.jpg\\r\\nhttp://farm1.static.flickr.com/41/87983009_25afa91e7f.jpg\\r\\nhttp://farm3.static.flickr.com/2116/2370921919_025ae215d9.jpg\\r\\nhttp://z.about.com/d/cruises/1/5/a/8/2/liberty178.jpg\\r\\nhttp://static.flickr.com/79/235765341_eed512b428.jpg\\r\\nhttp://farm4.static.flickr.com/3063/2567179687_4f191c7e3a.jpg\\r\\nhttp://ohioboating.net/b2/media/ghostship1a.jpg\\r\\nhttp://farm3.static.flickr.com/2400/2627593601_1a212398bf.jpg\\r\\nhttp://farm1.static.flickr.com/62/190893400_e91cbf2e61.jpg\\r\\nhttp://farm1.static.flickr.com/45/138503475_592911d5a0.jpg\\r\\nhttp://farm1.static.flickr.com/61/194005005_7d4affcc21.jpg\\r\\nhttp://farm2.static.flickr.com/1436/1465809088_2d8bab9870.jpg\\r\\nhttp://www.whoi.edu/cms/images/oceanus/Knorr_550_64256.jpg\\r\\nhttp://farm1.static.flickr.com/39/122631954_399f5fcd1a.jpg\\r\\nhttp://farm4.static.flickr.com/3114/2713077813_2e1fe794d6.jpg\\r\\nhttp://www.infotools.hfpeurope.org/energyinfos__e/hydrogen/pics/ship.jpg\\r\\nhttp://farm1.static.flickr.com/229/524399668_4fa32f091d.jpg\\r\\nhttp://voyage.pierwisconsin.org/Titanic/explore/vessel/ronaldbrown.jpg\\r\\nhttp://gcaptain.com/maritime/blog/wp-content/uploads/2007/06/ulstein-x-bow-container-ship.png\\r\\nhttp://www.americaswaterwaywatch.org/images/ship_moored.jpg\\r\\nhttp://static.flickr.com/89/231667795_bb5f1a6d4b.jpg\\r\\nhttp://farm3.static.flickr.com/2267/2525265827_04b9d1a1e6.jpg\\r\\nhttp://farm2.static.flickr.com/1096/1046391929_039a159831.jpg\\r\\nhttp://www.tripmall.com/sites/22232/images/river-voyager_ship_profile.jpg\\r\\nhttp://farm3.static.flickr.com/2415/2123407487_7c7f367403.jpg\\r\\nhttp://farm4.static.flickr.com/3243/2473589475_3c2cbc0462.jpg\\r\\nhttp://farm4.static.flickr.com/3016/2680043875_8fc94ed945.jpg\\r\\nhttp://www.ship-find.com/images/VLCC-RAM2006.jpg\\r\\nhttp://www.ship-technology.com/projects/occl/images/image2.jpg\\r\\nhttp://www.puertocadiz.com/imagenes/asti2.jpg\\r\\nhttp://farm1.static.flickr.com/175/479718442_fb183d127a.jpg\\r\\nhttp://farm3.static.flickr.com/2406/2266039068_982487d312.jpg\\r\\nhttp://farm2.static.flickr.com/1214/1467649072_a8f36caa8d.jpg\\r\\nhttp://static.flickr.com/81/235765335_5c75400819.jpg\\r\\nhttp://farm4.static.flickr.com/3037/2620748832_36eea76d23.jpg\\r\\nhttp://farm1.static.flickr.com/124/328769798_53baf8df47.jpg\\r\\nhttp://www.globalmaritime.com/static/images/shipping1.jpg\\r\\nhttp://www.foxnews.com/images/325856/7_63_ship3_320.jpg\\r\\nhttp://farm4.static.flickr.com/3158/2602276130_d138f1b196.jpg\\r\\nhttp://www.oilpubs.com/_ni/africa.jpg\\r\\nhttp://static.flickr.com/207/447330098_97f34811ab.jpg\\r\\nhttp://farm4.static.flickr.com/3255/2406848165_4e335444da.jpg\\r\\nhttp://farm4.static.flickr.com/3096/2459580733_19a54aea8c.jpg\\r\\nhttp://farm3.static.flickr.com/2284/2173436108_99dc6ab812.jpg\\r\\nhttp://static.flickr.com/216/447334119_92b738415b.jpg\\r\\nhttp://www.pscoman.com/marketing/DSC00551.JPG\\r\\nhttp://the.honoluluadvertiser.com/dailypix/2004/Oct/10/il36a_b.jpg\\r\\nhttp://farm3.static.flickr.com/2043/1538219367_eb93ebebd0.jpg\\r\\nhttp://farm3.static.flickr.com/2283/2207697157_3dc6ceb3c6.jpg\\r\\nhttp://ahoy.tk-jk.net/Images8/ContainerShipLoading.jpg\\r\\nhttp://media.monster.com/xfincaitx/nave.jpg\\r\\nhttp://farm4.static.flickr.com/3120/2483597518_c6ec0c80da.jpg\\r\\nhttp://farm4.static.flickr.com/3056/2640449331_4764c2d816.jpg\\r\\nhttp://farm4.static.flickr.com/3107/2667547903_9d5ce89440.jpg\\r\\nhttp://static.flickr.com/54/132333136_3fb55fc68d.jpg\\r\\nhttp://www.yachtingbrokers.com/scan0011.jpg\\r\\nhttp://static.flickr.com/95/232556370_71d4ec8693.jpg\\r\\nhttp://farm4.static.flickr.com/3180/2640552448_4964dfbf45.jpg\\r\\nhttp://farm4.static.flickr.com/3157/2698421840_1ecd742888.jpg\\r\\nhttp://static.flickr.com/225/499435965_223fb13a6c.jpg\\r\\nhttp://www.chinaembassy.org.np/science_tech/images/ship.jpg\\r\\nhttp://farm4.static.flickr.com/3160/2640550752_d20df03edc.jpg\\r\\nhttp://farm3.static.flickr.com/2108/2314839807_db10f09283.jpg\\r\\nhttp://farm3.static.flickr.com/2296/2526092106_0872d81814.jpg\\r\\nhttp://www.fugro.no/_Rainbow/images/default/GeoBarents_left_14x10cm_80dpi_rgb.jpg\\r\\nhttp://farm3.static.flickr.com/2249/1543300967_9c2e8a04d7.jpg\\r\\nhttp://farm2.static.flickr.com/1187/1424285374_f3ed2ec574.jpg\\r\\nhttp://farm2.static.flickr.com/1395/908107861_749b7a96ed.jpg\\r\\nhttp://cache.daylife.com/imageserve/04iu9Qz8B5fJf/610x.jpg\\r\\nhttp://farm3.static.flickr.com/2116/2167812190_20ee677dcd.jpg\\r\\nhttp://farm4.static.flickr.com/3142/2713291445_9e0becf411.jpg\\r\\nhttp://farm3.static.flickr.com/2344/2626321083_9c0b448583.jpg\\r\\nhttp://www.germes-online.com/direct/dbimage/50284631/Fishing_Vessel.jpg\\r\\nhttp://farm1.static.flickr.com/162/423543520_0dbd7c9213.jpg\\r\\nhttp://static.flickr.com/2214/2072384832_d3e7e64df8.jpg\\r\\nhttp://static.flickr.com/2100/1796049980_81ffda6ec3.jpg\\r\\nhttp://farm1.static.flickr.com/189/524458601_d095cb868e.jpg\\r\\nhttp://farm3.static.flickr.com/2215/2213619571_92e502c624.jpg\\r\\nhttp://www.manana.com/index_files/image001.jpg\\r\\nhttp://www.shipconstructor.com/images/misc/4250b.jpg\\r\\nhttp://farm2.static.flickr.com/1070/1040051126_65b310c692.jpg\\r\\nhttp://www.alaskanadventuretours.com/images/Alaskan_Leader_side.jpg\\r\\nhttp://www.festombord.com/_bilder-nerladdning/drottningholm-sommar2.jpg\\r\\nhttp://farm1.static.flickr.com/188/423543425_134e567962.jpg\\r\\nhttp://farm4.static.flickr.com/3252/2505741253_9103a53560.jpg\\r\\nhttp://www.nakilat.com.qa/English/images/AboutUs2.jpg\\r\\nhttp://farm3.static.flickr.com/2109/2378457074_f93866674e.jpg\\r\\nhttp://static.flickr.com/1198/552271593_c657d37e95.jpg\\r\\nhttp://farm1.static.flickr.com/48/145543190_65a0af08c2.jpg\\r\\nhttp://deilmancruises.com/VesselImages/PageSize/1048_ship-2.jpg\\r\\nhttp://farm1.static.flickr.com/47/120050731_c9bbd273af.jpg\\r\\nhttp://farm2.static.flickr.com/1244/631438141_9c571b49be.jpg\\r\\nhttp://english.kbs.co.kr/ICSFiles/artimage/2005/07/12/c_nws_zom/050712_l02.jpg\\r\\nhttp://farm1.static.flickr.com/227/459352894_0983f93dc8.jpg\\r\\nhttp://www.elitechoice.org/wp-content/uploads/2008/05/skylge.jpg\\r\\nhttp://farm4.static.flickr.com/3020/2620749908_6704d282de.jpg\\r\\nhttp://static.flickr.com/80/231667793_84442a656b.jpg\\r\\nhttp://farm4.static.flickr.com/3109/2641276874_ba9c949a08.jpg\\r\\nhttp://www.sas.jolt.co.uk/images/MapPics/ContainerShipOBJ1.jpg\\r\\nhttp://farm4.static.flickr.com/3084/2601480523_0c7f5500dc.jpg\\r\\nhttp://farm2.static.flickr.com/1368/1407985683_40f6b8d0d6.jpg\\r\\nhttp://farm1.static.flickr.com/54/137036483_cc03f1027a.jpg\\r\\nhttp://static.flickr.com/1103/538082893_c72009822c.jpg\\r\\nhttp://farm1.static.flickr.com/47/138502729_3f51506cc5.jpg\\r\\nhttp://static.flickr.com/2251/1732887858_3ef60402d3.jpg\\r\\nhttp://farm4.static.flickr.com/3213/2300900806_90c9984455.jpg\\r\\nhttp://www.nescaffold.com/html/images/Ship/SHIP%204.jpg\\r\\nhttp://farm4.static.flickr.com/3186/2631766149_50e9041e3b.jpg\\r\\nhttp://farm1.static.flickr.com/47/385835228_61431a5da2.jpg\\r\\nhttp://farm3.static.flickr.com/2002/2085938522_09ece75c66.jpg\\r\\nhttp://farm4.static.flickr.com/3128/2629807199_9c90a94741.jpg\\r\\nhttp://english.cri.cn/mmsource/images/2007/12/17/4086ship.jpg\\r\\nhttp://farm3.static.flickr.com/2025/2308916972_f7c531bb32.jpg\\r\\nhttp://www.stockphotography.co.uk/Upload/Stock/Watermarked/20274.jpg\\r\\nhttp://static.flickr.com/2207/1842172673_d97f75e51f.jpg\\r\\nhttp://www.shipagent.ru/eng/contentimage/13_202.jpg\\r\\nhttp://www.ontarioferries.com/chi/english/images/about/ships_large/manitoulinL.jpg\\r\\nhttp://static.flickr.com/95/242913660_512de1c1e3.jpg\\r\\nhttp://www.zetnet.co.uk/oigs/derbyshire/derby.gif\\r\\nhttp://farm3.static.flickr.com/2295/2519477565_f2acd19d86.jpg\\r\\nhttp://www.imcbrokers.com/blog/wp-content/uploads/2007/06/dutch-landing-craft-l9528.jpg\\r\\nhttp://farm3.static.flickr.com/2244/2483801176_9fc734c0a4.jpg\\r\\nhttp://farm1.static.flickr.com/53/120055267_71dd255748.jpg\\r\\nhttp://www.dainferry.co.kr/image/vessel/SHIP.jpg\\r\\nhttp://farm3.static.flickr.com/2290/2239345415_4b74c3b32a.jpg\\r\\nhttp://farm3.static.flickr.com/2077/2435231907_443e03232a.jpg\\r\\nhttp://farm2.static.flickr.com/1105/974037616_cbbfaa9b15.jpg\\r\\nhttp://www.oilpubs.com/_ni/angola.jpg\\r\\nhttp://static.flickr.com/1197/546233189_8388aa5abb.jpg\\r\\nhttp://farm4.static.flickr.com/3043/2519296663_2591e47250.jpg\\r\\nhttp://farm1.static.flickr.com/55/138503050_db61c70a7c.jpg?v=0\\r\\nhttp://www.marinelog.com/IMAGESMMIV/300cobrajudy.jpg\\r\\nhttp://media-seals.seashepherd2.org/photos/SSCS_Coast_Guard_almost_Ramming_FM_and_Sealing_Ship.jpg\\r\\nhttp://farm3.static.flickr.com/2122/2004729557_5cf7bbc74d.jpg\\r\\nhttp://farm3.static.flickr.com/2037/2387756759_4dc3a945db.jpg\\r\\nhttp://farm1.static.flickr.com/24/58989336_b2ea6ca077.jpg\\r\\nhttp://www.alohadestinations.com/cruises/carnival-spirit-cruise_md.jpg\\r\\nhttp://www.cnn.com/WORLD/Newsbriefs/9601/01-11/freighter.rtv.jpg\\r\\nhttp://farm4.static.flickr.com/3037/2582711281_c35d9df4c8.jpg\\r\\nhttp://farm1.static.flickr.com/42/76710747_c8b4e6025e.jpg\\r\\nhttp://static.flickr.com/2357/2084338770_d4400c36e2.jpg\\r\\nhttp://farm3.static.flickr.com/2202/1810841698_bb7029a5e8.jpg\\r\\nhttp://farm2.static.flickr.com/1021/537727729_16ea136b2b.jpg\\r\\nhttp://farm2.static.flickr.com/1349/1458373981_36b90f0125.jpg\\r\\nhttp://farm3.static.flickr.com/2083/2069832214_3ee51e6383.jpg\\r\\nhttp://farm2.static.flickr.com/1259/1408865824_86b460f867.jpg\\r\\nhttp://farm4.static.flickr.com/3202/2508377300_d524977fc2.jpg\\r\\nhttp://farm3.static.flickr.com/2075/2478034035_61207ba88a.jpg\\r\\nhttp://farm3.static.flickr.com/2017/2539154120_248e15e747.jpg\\r\\nhttp://static.flickr.com/85/233570565_eab16936ba.jpg\\r\\nhttp://www.globalpulse.net/images/emma-maersk-lg.jpg\\r\\nhttp://www.terradaily.com/images/research-vessel-ship-knorr-black-sea-bg.jpg\\r\\nhttp://farm4.static.flickr.com/3215/2646831534_2bc007fea6.jpg\\r\\nhttp://farm4.static.flickr.com/3004/2297382436_fd83165611.jpg\\r\\nhttp://news.holidayhypermarket.co.uk/Images/18312286/Cunard_Will_Build_Queen_Elizabeth_Vessel_large.jpg\\r\\nhttp://farm3.static.flickr.com/2301/2518416999_fc70e78cc9.jpg\\r\\nhttp://farm1.static.flickr.com/28/145308546_d81a68234e.jpg\\r\\nhttp://research.dnv.com/marmil/images/image106.jpg\\r\\nhttp://www.expertcruiser.com/wp-content/uploads/worlddiscoverer.jpg\\r\\nhttp://static.flickr.com/1430/1070931170_390263a5c3.jpg\\r\\nhttp://www.coscoship-qd.com/images/ship1.jpg\\r\\nhttp://farm4.static.flickr.com/3104/2725241425_dd5af9820a.jpg\\r\\nhttp://www.eatmussels.com/ship.jpg\\r\\nhttp://bridgehands.com/Reviews/Bridge_Cruises/Norwegian_Pride_of_Aloha.jpg\\r\\nhttp://farm1.static.flickr.com/45/149697030_a9561474db.jpg\\r\\nhttp://farm3.static.flickr.com/2299/2392339220_4a0bbaa6ef.jpg\\r\\nhttp://www.arcus.org/alias/images/millerfreeman1.jpg\\r\\nhttp://farm3.static.flickr.com/2002/2492743983_e3f5bcaf1a.jpg\\r\\nhttp://www.revistainterforum.com/english/imagesen/052702Collins2.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/7/7b/SK_Alfred_Krupp_Kiel2007.jpg/800px-SK_Alfred_Krupp_Kiel2007.jpg\\r\\nhttp://www.boatingsf.com/photos/031105/DSC_3833_edited-1.jpg\\r\\nhttp://farm1.static.flickr.com/94/261341449_2933ae1a24.jpg\\r\\nhttp://farm4.static.flickr.com/3294/2454374341_562860d23a.jpg\\r\\nhttp://farm1.static.flickr.com/103/255442281_76f199eb22.jpg\\r\\nhttp://farm1.static.flickr.com/96/232105585_81510cdbae.jpg\\r\\nhttp://kcaw.org/modules/local_news/media/pictures/ship%20from%20in%20ship%20vertical.jpg\\r\\nhttp://www.caribbeaninsurers.com/img/phts/ciml-cargo-ship.jpg\\r\\nhttp://farm4.static.flickr.com/3117/2675453414_7b342f9911.jpg\\r\\nhttp://farm4.static.flickr.com/3260/2732467519_400bc1a313.jpg\\r\\nhttp://static.flickr.com/109/291489105_629b29352e.jpg\\r\\nhttp://farm1.static.flickr.com/45/144880536_263dc674d6.jpg\\r\\nhttp://msnbcmedia1.msn.com/j/msnbc/Components/Photos/060815/060815_tanker_vlrg_3a.widec.jpg\\r\\nhttp://www.jamstec.org/image/ship-natsushima.jpg\\r\\nhttp://cache.daylife.com/imageserve/02cE5SU4lR2Ue/610x.jpg\\r\\nhttp://farm4.static.flickr.com/3247/2495712650_29ed278c0c.jpg\\r\\nhttp://farm3.static.flickr.com/2099/2142237887_be7dde8614.jpg\\r\\nhttp://farm4.static.flickr.com/3062/2718110538_77f399f1b7.jpg\\r\\nhttp://farm1.static.flickr.com/108/289991937_d587e56024.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/a/a8/FS_Mistral_02.jpg\\r\\nhttp://farm3.static.flickr.com/2339/2348324535_f7e15f3b25.jpg\\r\\nhttp://www.tq-international.com/NorthPole2008/Images/Yamal.jpg\\r\\nhttp://farm4.static.flickr.com/3177/2639483240_516f98cbfc.jpg\\r\\nhttp://farm3.static.flickr.com/2200/2050999697_0cb861b13b.jpg\\r\\nhttp://farm1.static.flickr.com/149/360485107_6fcb1b740b.jpg\\r\\nhttp://farm3.static.flickr.com/2091/2051783536_de01c46c1a.jpg\\r\\nhttp://farm1.static.flickr.com/22/34369761_783cfb312e.jpg\\r\\nhttp://farm4.static.flickr.com/3239/2307805889_7eafed8edb.jpg\\r\\nhttp://farm3.static.flickr.com/2239/2291097948_657a49345a.jpg\\r\\nhttp://farm3.static.flickr.com/2133/2380557163_7a635ae02b.jpg\\r\\nhttp://www.nordcapital.com/main/nca/images/Cosco_Napoli_(130606-07).JPG\\r\\nhttp://static.flickr.com/84/232527718_2bb0a05b66.jpg\\r\\nhttp://farm1.static.flickr.com/46/120056432_d25e9aaa21.jpg\\r\\nhttp://static.flickr.com/101/266567586_c7436c75df.jpg\\r\\nhttp://farm1.static.flickr.com/46/183951497_9592d70cfd.jpg\\r\\nhttp://farm4.static.flickr.com/3289/2712827249_5c94c70d9e.jpg\\r\\nhttp://farm3.static.flickr.com/2196/2442677327_3d1913d3af.jpg\\r\\nhttp://cache.eb.com/eb/image?id=101832&rendTypeId=4\\r\\nhttp://farm4.static.flickr.com/3092/2734968931_1f53cb0c16.jpg\\r\\nhttp://www.gomr.mms.gov/images_opt/graphics/whale/GG-and-Whale.jpg\\r\\nhttp://farm2.static.flickr.com/1370/1408866292_ffe89967c8.jpg\\r\\nhttp://farm3.static.flickr.com/2225/2252880199_cf53d1226d.jpg\\r\\nhttp://farm1.static.flickr.com/48/144869183_19b2e3880f.jpg\\r\\nhttp://farm1.static.flickr.com/50/144130441_675d4864a1.jpg\\r\\nhttp://farm2.static.flickr.com/1407/1488286588_2e0f3d91e1.jpg\\r\\nhttp://farm1.static.flickr.com/47/141558177_2f5c5fc1b1.jpg\\r\\nhttp://farm1.static.flickr.com/147/380907187_6013af580d.jpg\\r\\nhttp://farm3.static.flickr.com/2369/2525069034_00de587679.jpg\\r\\nhttp://farm2.static.flickr.com/1437/612572640_a316d4b185.jpg\\r\\nhttp://farm3.static.flickr.com/2252/2431278912_77615eb673.jpg\\r\\nhttp://farm4.static.flickr.com/3055/2336857819_fdce456a8a.jpg\\r\\nhttp://farm4.static.flickr.com/3113/2426139663_6b9d94d1c7.jpg\\r\\nhttp://farm1.static.flickr.com/53/145572328_bbf15d53f5.jpg\\r\\nhttp://farm1.static.flickr.com/49/145283853_cd10568152.jpg\\r\\nhttp://farm4.static.flickr.com/3185/2608185975_3a817c6bb9.jpg\\r\\nhttp://farm4.static.flickr.com/3124/2408542413_63c85d83ed.jpg\\r\\nhttp://farm1.static.flickr.com/125/416408431_cf30813eda.jpg\\r\\nhttp://farm1.static.flickr.com/55/145275700_fea643510e.jpg\\r\\nhttp://farm1.static.flickr.com/81/232105446_d24a4d1345.jpg\\r\\nhttp://farm2.static.flickr.com/1408/624185759_93e0413d9a.jpg\\r\\nhttp://farm3.static.flickr.com/2240/2219289346_5d747222cf.jpg\\r\\nhttp://farm4.static.flickr.com/3179/2407520558_3e1c7f8304.jpg\\r\\nhttp://farm1.static.flickr.com/146/447341253_46d8c4eeda.jpg\\r\\nhttp://farm2.static.flickr.com/1439/1144182741_3253b26c6e.jpg\\r\\nhttp://farm1.static.flickr.com/56/145572324_8f1421b128.jpg\\r\\nhttp://farm3.static.flickr.com/2086/2253349310_351a07a421.jpg\\r\\nhttp://farm4.static.flickr.com/3009/2583541408_e17b06531f.jpg\\r\\nhttp://farm4.static.flickr.com/3264/2391524129_b1e266cc9f.jpg\\r\\nhttp://farm4.static.flickr.com/3251/2492780820_b17ae31893.jpg\\r\\nhttp://www.ascinfo.co.uk/9/pdcnewsitem/00/79/32/Edda.jpg\\r\\nhttp://farm4.static.flickr.com/3133/2537452896_0698856a91.jpg\\r\\nhttp://farm1.static.flickr.com/60/184727106_b7be1e3a6d.jpg\\r\\nhttp://farm1.static.flickr.com/2/3691887_33c5903ba0.jpg\\r\\nhttp://gcaptain.com/maritime/blog/wp-content/uploads/2007/10/norweigian-gem-ugly-ship.jpg\\r\\nhttp://farm1.static.flickr.com/153/425314549_7b159bd2b7.jpg\\r\\nhttp://static.flickr.com/59/184727110_5562a5c65f.jpg\\r\\nhttp://www.aw-kreuzfahrten.de/content/awkVessel/detail/aura_ship1_web.jpg\\r\\nhttp://farm3.static.flickr.com/2416/2454426554_d85225f03d.jpg\\r\\nhttp://farm1.static.flickr.com/219/500110649_17af9c6433.jpg\\r\\nhttp://farm4.static.flickr.com/3275/2640532242_81ba51a532.jpg\\r\\nhttp://farm3.static.flickr.com/2336/2254066096_9b9e2cd538.jpg\\r\\nhttp://farm4.static.flickr.com/3289/2295514667_a634679ecb.jpg\\r\\nhttp://www.marina.difesa.it/storia/almanacco/parte04/images/Foto04g/n0433-00bg.jpg\\r\\nhttp://www.tug-barge.com/p502.h32.jpg\\r\\nhttp://static.flickr.com/81/232539596_d368faaea7.jpg\\r\\nhttp://www.bulgariablog.net/ClientFiles/fc6a5dd7-7d7b-4ebe-9e72-0135c814ee23/the-ship-serenade.jpg\\r\\nhttp://farm3.static.flickr.com/2312/2492727595_54d6ee1d70.jpg\\r\\nhttp://static.flickr.com/90/231622240_52337a5022.jpg\\r\\nhttp://static.flickr.com/67/184178730_e53177d26e.jpg\\r\\nhttp://farm4.static.flickr.com/3084/2602355690_dab4165b1c.jpg\\r\\nhttp://farm2.static.flickr.com/1013/907941771_f034bee066.jpg\\r\\nhttp://farm4.static.flickr.com/3101/2627593165_ded60289ef.jpg\\r\\nhttp://farm2.static.flickr.com/1058/695362550_f7dc00b852.jpg\\r\\nhttp://farm1.static.flickr.com/156/447346520_2f1bb2fdc2.jpg\\r\\nhttp://farm3.static.flickr.com/2280/2383615610_3090b5b748.jpg\\r\\nhttp://www.cruiseman.com/images/image25g.jpg\\r\\nhttp://static.flickr.com/90/262578215_8f2d8a6658.jpg\\r\\nhttp://farm1.static.flickr.com/81/216558076_b08f4b6a02.jpg\\r\\nhttp://farm4.static.flickr.com/3024/2348285813_2acf617273.jpg\\r\\nhttp://farm2.static.flickr.com/1107/1408865734_c28f714cdd.jpg\\r\\nhttp://www.ship-technology.com/contractor_images/abas/2-pearl-vessel.jpg\\r\\nhttp://farm3.static.flickr.com/2124/2069021175_3899d44f68.jpg\\r\\nhttp://farm2.static.flickr.com/1306/917357980_3e32b06b28.jpg\\r\\nhttp://www.worldwaterways.com/VesselImages/PageSize/1039_ship.jpg\\r\\nhttp://farm3.static.flickr.com/2154/2381905813_b66d25772f.jpg\\r\\nhttp://www.expertcruiser.com/wp-content/uploads/explorersinking.jpg\\r\\nhttp://bridgehands.com/Reviews/Bridge_Cruises/Regatta.jpg\\r\\nhttp://farm4.static.flickr.com/3070/2681282489_e296f23f70.jpg\\r\\nhttp://static.flickr.com/1360/942994547_b78d9c862a.jpg\\r\\nhttp://static.flickr.com/28/103339931_50d4debfbe.jpg\\r\\nhttp://static.flickr.com/2153/2069823186_83993fde16.jpg\\r\\nhttp://www.news.wisc.edu/story_images/0000/0389/drill_ship_Chikyu07.jpg\\r\\nhttp://farm3.static.flickr.com/2193/2533189582_f8c76e4c0d.jpg\\r\\nhttp://farm4.static.flickr.com/3151/2640447799_9cd7fa28e7.jpg\\r\\nhttp://www.yachtsales.com/text/dr106a.jpg\\r\\nhttp://farm3.static.flickr.com/2240/2069028593_6caeda87a1.jpg\\r\\nhttp://farm2.static.flickr.com/1068/1408869514_4964f74bd6.jpg\\r\\nhttp://farm3.static.flickr.com/2145/2499377854_b361ba6a31.jpg\\r\\nhttp://proftech.org/industrial/marinus_green_ship_photo.jpg\\r\\nhttp://farm3.static.flickr.com/2102/2467851995_c30a3a2c0e.jpg\\r\\nhttp://farm3.static.flickr.com/2094/2250103232_d41704097d.jpg\\r\\nhttp://static.flickr.com/2401/1759129409_58dabbe2a0.jpg\\r\\nhttp://farm4.static.flickr.com/3012/2353693226_22b8518bab.jpg\\r\\nhttp://farm2.static.flickr.com/1101/1270320135_c0c57a9ac0.jpg\\r\\nhttp://farm4.static.flickr.com/3048/2557053996_aa74419cfe.jpg\\r\\nhttp://lh6.ggpht.com/_72LnLKSXh88/RoKT03yseQI/AAAAAAAAAGk/zW2VJojmPNo/sub+and+park.jpg\\r\\nhttp://www.ausspecialforces.com/images/heroinship.jpg\\r\\nhttp://farm4.static.flickr.com/3228/2657012825_3fd4c5eef6.jpg\\r\\nhttp://farm1.static.flickr.com/209/493553628_778d980fe2.jpg\\r\\nhttp://www.maritimesales.com/images/Others/UM10-02.jpg\\r\\nhttp://farm2.static.flickr.com/1125/538089369_22aa2df916.jpg\\r\\nhttp://farm2.static.flickr.com/1211/1058104085_f1b9276fa0.jpg\\r\\nhttp://static.flickr.com/1360/1325075361_b6e154b323.jpg\\r\\nhttp://www.tripmall.com/sites/22232/images/DnieperStar_ship_profile.jpg\\r\\nhttp://farm1.static.flickr.com/107/286660507_385b8a95ea.jpg\\r\\nhttp://farm3.static.flickr.com/2109/2228206153_b8701278e7.jpg\\r\\nhttp://farm4.static.flickr.com/3015/2642996595_6f28f4b59a.jpg\\r\\nhttp://farm4.static.flickr.com/3010/2540276656_816b58d622.jpg\\r\\nhttp://farm1.static.flickr.com/49/179483975_7646e022dc.jpg\\r\\nhttp://farm2.static.flickr.com/1253/611957433_b271e3c647.jpg\\r\\nhttp://lh3.ggpht.com/_hjN28VPmEag/R8P9tFwXfZI/AAAAAAAAB7o/A7b6ZfwBslk/P1010091.JPG\\r\\nhttp://farm2.static.flickr.com/1014/1407984059_ae101b29bd.jpg\\r\\nhttp://farm1.static.flickr.com/200/461416497_8af3948c95.jpg\\r\\nhttp://farm4.static.flickr.com/3001/2504324614_09af21c7b2.jpg\\r\\nhttp://farm4.static.flickr.com/3202/2648505672_39f838521b.jpg\\r\\nhttp://www.koreatimes.co.kr/upload/news/080330_p9_northmain.jpg\\r\\nhttp://www.spacewar.com/images/north-korean-ship-vessel-yalu-river-2006-afp-bg.jpg\\r\\nhttp://www.pscoman.com/marketing/DSC02082.JPG\\r\\nhttp://farm1.static.flickr.com/98/218902390_72424b5d55.jpg\\r\\nhttp://static.flickr.com/196/490506413_157b6d945e.jpg\\r\\nhttp://www.chinadaily.com.cn/english/doc/2005-08/30/xin_380802310811574252945.jpg\\r\\nhttp://farm1.static.flickr.com/116/277213278_64aa9a8668.jpg\\r\\nhttp://farm3.static.flickr.com/2113/2237668229_47fb6c108f.jpg\\r\\nhttp://farm1.static.flickr.com/166/423543279_56bc160edb.jpg\\r\\nhttp://farm1.static.flickr.com/21/25475879_5a7a103976.jpg\\r\\nhttp://farm4.static.flickr.com/3216/2601528129_97422e759b.jpg\\r\\nhttp://farm1.static.flickr.com/146/356156131_148bb5f8eb.jpg\\r\\nhttp://farm4.static.flickr.com/3072/2640448757_45a57858ff.jpg\\r\\nhttp://farm3.static.flickr.com/2171/2435231351_6e23ee166b.jpg\\r\\nhttp://static.flickr.com/1163/1299676771_e95ff05cff.jpg\\r\\nhttp://farm3.static.flickr.com/2056/1933864683_24ee92046e.jpg\\r\\nhttp://farm2.static.flickr.com/1142/708059659_a261bae3bf.jpg\\r\\nhttp://farm3.static.flickr.com/2029/2483600790_bd8430f627.jpg\\r\\nhttp://farm1.static.flickr.com/211/502687253_b7bec59323.jpg\\r\\nhttp://www.scanlandltd.co.uk/Ship2.jpg\\r\\nhttp://farm2.static.flickr.com/1180/908838452_cea2449ed6.jpg\\r\\nhttp://www.hoegh.com/autoliners/multimedia/archive/00002/DSC_0039_2800e.jpg\\r\\nhttp://farm1.static.flickr.com/92/268474470_72917cac50.jpg\\r\\nhttp://farm2.static.flickr.com/1179/908281823_c4f1a39650.jpg\\r\\nhttp://static.flickr.com/1297/1164959379_4d33464b05.jpg\\r\\nhttp://farm3.static.flickr.com/2038/2050999867_6e42ec9787.jpg\\r\\nhttp://www.eturbonews.com/files/imagecache/teaser/files/cruiseship.bmp\\r\\nhttp://farm3.static.flickr.com/2385/2069024941_77140826bf.jpg\\r\\nhttp://farm4.static.flickr.com/3068/2555558779_cc6131aa6f.jpg\\r\\nhttp://farm1.static.flickr.com/190/524458575_d46e01af74.jpg\\r\\nhttp://farm2.static.flickr.com/1257/1408869782_01195178e8.jpg\\r\\nhttp://farm2.static.flickr.com/1346/1219656987_71e3335c69.jpg\\r\\nhttp://farm2.static.flickr.com/1323/917279812_e42f5a2274.jpg\\r\\nhttp://proteus.pca.state.mn.us/artwork/water/ballast-ship.jpg\\r\\nhttp://aycu17.webshots.com/image/31456/2000682631551131311_rs.jpg\\r\\nhttp://farm4.static.flickr.com/3204/2336857699_5461510a13.jpg\\r\\nhttp://static.flickr.com/2095/2084398631_4448655224.jpg\\r\\nhttp://farm1.static.flickr.com/77/179386849_ff962e74cb.jpg\\r\\nhttp://farm4.static.flickr.com/3080/2475656774_462cdb0c15.jpg\\r\\nhttp://farm1.static.flickr.com/69/188303507_d2bebd99ae.jpg\\r\\nhttp://farm4.static.flickr.com/3186/2701572918_40f29fd04b.jpg\\r\\nhttp://farm2.static.flickr.com/1414/908769474_5396ca081a.jpg\\r\\nhttp://www.bymnews.com/july/images/big/MSC-Napoli-MCA.jpg\\r\\nhttp://farm1.static.flickr.com/54/145308539_b8d3f67af7.jpg\\r\\nhttp://farm2.static.flickr.com/1112/529734115_c29d2de053.jpg\\r\\nhttp://farm3.static.flickr.com/2258/2097608924_1890988861.jpg\\r\\nhttp://www.liveexport-indefensible.com/images/250-live_export_ship.jpg\\r\\nhttp://farm2.static.flickr.com/1145/908975482_c445b5946c.jpg\\r\\nhttp://www.alexsurvey.com/images/vessel_3.gif\\r\\nhttp://farm3.static.flickr.com/2324/2423603583_84902d8055.jpg\\r\\nhttp://www.firstship.co.jp/images/ship_jupt.jpg\\r\\nhttp://farm4.static.flickr.com/3262/2609018598_c88012cf14.jpg\\r\\nhttp://www.procargo.at/images/vessels/multipurposeship.jpg\\r\\nhttp://sanctuaries.noaa.gov/sse/education/summits/Images/SBmimi.jpeg\\r\\nhttp://www.bp.com/liveassets/bp_internet/bp_shipping/bp_shipping_english/STAGING/local_assets/images/deck375x250.jpg\\r\\nhttp://farm2.static.flickr.com/1107/1425397193_4627146a56.jpg\\r\\nhttp://farm3.static.flickr.com/2214/2278312426_8776f0a720.jpg\\r\\nhttp://farm2.static.flickr.com/1270/917238580_ff32f46034.jpg\\r\\nhttp://static.flickr.com/98/238522463_aedef8ef47.jpg\\r\\nhttp://www.spanopoulos-group.com/images/vosporos.jpg\\r\\nhttp://farm3.static.flickr.com/2234/2522875802_a4d40e51df.jpg\\r\\nhttp://farm3.static.flickr.com/2001/2051782446_050b491150.jpg\\r\\nhttp://www.solarnavigator.net/images/stealth_vessel_lockheed_martin_usa_navy_base_leaving_port.jpg\\r\\nhttp://www.procargo.at/images/vessels/ungearedship.jpg\\r\\nhttp://farm1.static.flickr.com/45/156433319_1f3c44e603.jpg\\r\\nhttp://farm3.static.flickr.com/2063/2483595736_8433dd9b66.jpg\\r\\nhttp://farm2.static.flickr.com/1056/537076121_4fe0ae6c0e.jpg\\r\\nhttp://farm4.static.flickr.com/3055/2410183436_dfa205bab4.jpg\\r\\nhttp://farm1.static.flickr.com/44/120055982_678ef4e268.jpg\\r\\nhttp://farm3.static.flickr.com/2377/2051783908_725e3e99e3.jpg\\r\\nhttp://farm2.static.flickr.com/1121/1400337655_e4225e4193.jpg\\r\\nhttp://farm4.static.flickr.com/3250/2531192974_9c7a340006.jpg\\r\\nhttp://cache.daylife.com/imageserve/018pfM0aa1fFI/610x.jpg\\r\\nhttp://home.att.net/~lgrosner/Solomons/Ship.JPG\\r\\nhttp://www.kadmar.com/images/tramp_s.jpg\\r\\nhttp://www.irvingshipbuilding.com/images/shipbuildpic1.jpg\\r\\nhttp://farm3.static.flickr.com/2159/2264736209_f739d173fb.jpg\\r\\nhttp://farm1.static.flickr.com/156/423543311_00be222566.jpg\\r\\nhttp://www.shipphotos.co.uk/images/solitaire.jpg\\r\\nhttp://www.dolphin-hellas.gr/Cruises/Louis-cruises/Ships/sea-diamond-cruise-ship.jpg\\r\\nhttp://www.lm-illustration.dk/test/TT/FenderDeploymentShip12.jpg\\r\\nhttp://coastguardnews.com/images/77219.jpg\\r\\nhttp://farm3.static.flickr.com/2061/2266079836_f8afa07f78.jpg\\r\\nhttp://farm1.static.flickr.com/103/277215410_3d09a4aebe.jpg\\r\\nhttp://debragaines.com/imgphotos/Ghost_Ship_1.jpg\\r\\nhttp://farm3.static.flickr.com/2068/1816398149_0ca5faff16.jpg\\r\\nhttp://farm1.static.flickr.com/132/348304598_593bac6f20.jpg\\r\\nhttp://static.flickr.com/2246/2118115345_7ab6f36cd1.jpg\\r\\nhttp://farm3.static.flickr.com/2317/2370969101_97557e7bc3.jpg\\r\\nhttp://farm3.static.flickr.com/2076/2492753681_fe1539218e.jpg\\r\\nhttp://www.coastlinemaritime.com/vessels/big_marianos.jpg\\r\\nhttp://farm1.static.flickr.com/164/435684070_883b4e51f0.jpg\\r\\nhttp://farm3.static.flickr.com/2262/1701278736_9bad4b036f.jpg\\r\\nhttp://www.nordcapital.com/main/nca/images/ER_Shanghai_druck.jpg\\r\\nhttp://farm2.static.flickr.com/1212/759954686_c1f2c88330.jpg\\r\\nhttp://www.bensfriends.com/gadabout/archives/10ship.jpg\\r\\nhttp://tugster.files.wordpress.com/2007/08/bl.jpg\\r\\nhttp://farm4.static.flickr.com/3059/2498548707_de1332760a.jpg\\r\\nhttp://static.flickr.com/87/235766476_471b0635ce.jpg\\r\\nhttp://farm4.static.flickr.com/3091/2462154988_d9c76315ff.jpg\\r\\nhttp://farm4.static.flickr.com/3037/2499317690_084be80d72.jpg\\r\\nhttp://farm1.static.flickr.com/46/145283844_1d86019c9c.jpg\\r\\nhttp://farm3.static.flickr.com/2187/2525754478_e88a5db7ff.jpg\\r\\nhttp://farm4.static.flickr.com/3168/2579450119_3991e73c02.jpg\\r\\nhttp://farm4.static.flickr.com/3161/2337692192_089e860b6c.jpg\\r\\nhttp://farm2.static.flickr.com/1112/908946016_de33051506.jpg\\r\\nhttp://farm1.static.flickr.com/186/423544562_985c4dcd2d.jpg\\r\\nhttp://farm3.static.flickr.com/2209/2498181288_a40896fcb3.jpg\\r\\nhttp://static.flickr.com/1040/986859599_50e6b6f502.jpg\\r\\nhttp://farm3.static.flickr.com/2320/2537566391_cc31642364.jpg\\r\\nhttp://www.emarald.net/images/container_ship_1.jpg\\r\\nhttp://farm2.static.flickr.com/1439/1491235090_affe95ba90.jpg\\r\\nhttp://farm3.static.flickr.com/2133/2455887067_4d97ccf778.jpg\\r\\nhttp://static.flickr.com/83/232105652_8fbce30be6.jpg\\r\\nhttp://farm3.static.flickr.com/2103/2072165767_e2118ae4b7.jpg\\r\\nhttp://farm3.static.flickr.com/2352/1950216096_65323ce161.jpg\\r\\nhttp://www.marcon.com/library/sales_reports/2001_2003/2003salesfiles/oceandiverIV.jpg\\r\\nhttp://static.flickr.com/51/145283839_30f3974550.jpg\\r\\nhttp://www.smh.com.au/ffximage/2007/02/15/nisshin_maru_wideweb__470x281,0.jpg\\r\\nhttp://static.flickr.com/220/522462888_660ba8bef3.jpg\\r\\nhttp://www.zeezand.nl/fotos%5Cschip1.jpg\\r\\nhttp://farm3.static.flickr.com/2314/2529739762_aa17b9b157.jpg\\r\\nhttp://www.coastlinemaritime.com/vessels/big_msvexplorer.jpg\\r\\nhttp://www.eagle.org/news/press/images/EvergreenEverDev.jpg\\r\\nhttp://farm1.static.flickr.com/126/337366522_a7c617e3a2.jpg\\r\\nhttp://farm1.static.flickr.com/240/447337145_40c4107e86.jpg\\r\\nhttp://www.spacedaily.com/images/ship-stealth-bg.jpg\\r\\nhttp://farm4.static.flickr.com/3233/2308540240_5e120b1dcd.jpg\\r\\nhttp://farm2.static.flickr.com/1137/906575172_e168a03d70.jpg\\r\\nhttp://farm3.static.flickr.com/2167/1801220647_c0470d1fe7.jpg\\r\\nhttp://static.flickr.com/34/122631936_c1cf2338dd.jpg\\r\\nhttp://farm1.static.flickr.com/130/388461949_cc6c0b1de6.jpg\\r\\nhttp://farm3.static.flickr.com/2113/2301031424_5c2e98529d.jpg\\r\\nhttp://farm4.static.flickr.com/3287/2503505123_4a5d2f5bf0.jpg\\r\\nhttp://farm3.static.flickr.com/2215/2509537110_3cd6af01a6.jpg\\r\\nhttp://farm3.static.flickr.com/2221/2361692767_3214fa6b2f.jpg\\r\\nhttp://farm1.static.flickr.com/215/484282102_65666cb269.jpg\\r\\nhttp://farm2.static.flickr.com/1134/1061344395_ff2e47a52a.jpg\\r\\nhttp://www.bio.gc.ca/research/overview/images/chsbigthumb.jpg\\r\\nhttp://www.russianempirecruises.com/img/4exov_ship_profile.jpg\\r\\nhttp://farm3.static.flickr.com/2180/1949277451_9962c31494.jpg\\r\\nhttp://farm3.static.flickr.com/2406/2069035025_f72d43a00d.jpg\\r\\nhttp://farm2.static.flickr.com/1181/1364223131_4f3a346c6a.jpg\\r\\nhttp://static.flickr.com/90/232551472_c7087cd104.jpg\\r\\nhttp://static.flickr.com/217/535851660_b1ac04dade.jpg\\r\\nhttp://farm1.static.flickr.com/106/308610539_77d644be78.jpg\\r\\nhttp://farm2.static.flickr.com/1168/1355600071_0d4141a7c3.jpg\\r\\nhttp://farm3.static.flickr.com/2347/2164121134_4dfee48a3e.jpg\\r\\nhttp://farm2.static.flickr.com/1162/943821610_1de7e570dd.jpg\\r\\nhttp://www.navis.gr/manager/images/bollard2.jpg\\r\\nhttp://www.cadea.hr/WWW_Cadea_Files/Podravina1.jpg\\r\\nhttp://farm1.static.flickr.com/45/149483568_0ed1f881e7.jpg\\r\\nhttp://farm4.static.flickr.com/3105/2725867756_92b12c679e.jpg\\r\\nhttp://lh4.ggpht.com/_sQuBFg48_FE/RwgTPzzg_EI/AAAAAAAAAh8/qCXM6L-HoW8/IMG_1424.JPG\\r\\nhttp://farm1.static.flickr.com/52/138503857_e5700a2aea.jpg\\r\\nhttp://farm3.static.flickr.com/2245/2164094896_515489a6ef.jpg\\r\\nhttp://farm3.static.flickr.com/2012/2084300994_ca1387280e.jpg\\r\\nhttp://www.sallys-place.com/images/travel/asia/vietnam/p1-1-ship.jpg\\r\\nhttp://farm4.static.flickr.com/3117/2662213656_d9201ec4cb.jpg\\r\\nhttp://farm1.static.flickr.com/46/145278524_2dc35be996.jpg\\r\\nhttp://farm2.static.flickr.com/1117/882800755_abe0b050c7.jpg\\r\\nhttp://farm3.static.flickr.com/2146/2117752707_1ede8e035d.jpg\\r\\nhttp://www.alfinal.com/Antartida/alej3.jpg\\r\\nhttp://www.mcdoa.org.uk/images/Laleston_in_Norway.jpg\\r\\nhttp://farm3.static.flickr.com/2005/2089668420_76fda9fd52.jpg\\r\\nhttp://farm3.static.flickr.com/2070/2482611078_055832f74b.jpg\\r\\nhttp://retireflorida.org/community/images/stories/activities/cruise-ship.jpg\\r\\nhttp://farm1.static.flickr.com/53/141148236_8adeb2fa8a.jpg\\r\\nhttp://static.flickr.com/93/238527975_7d9b9be4fb.jpg\\r\\nhttp://farm3.static.flickr.com/2143/1949399789_734c641dcb.jpg\\r\\nhttp://www.painetworks.com/photos/hy/hy2187.JPG\\r\\nhttp://www.expertcruiser.com/wp-content/uploads/nclpearl.jpg\\r\\nhttp://static.flickr.com/241/447333703_f3db903969.jpg\\r\\nhttp://farm1.static.flickr.com/6/7968724_eae58d9180.jpg\\r\\nhttp://farm3.static.flickr.com/2280/2239984046_75fda70fe2.jpg\\r\\nhttp://www.kouwan.metro.tokyo.jp/english/port/images/ship.jpg\\r\\nhttp://farm3.static.flickr.com/2273/2410181930_f786249364.jpg\\r\\nhttp://farm3.static.flickr.com/2298/2060588997_7533f5e731.jpg\\r\\nhttp://farm3.static.flickr.com/2408/1979322244_9c0afab5dd.jpg\\r\\nhttp://www.highprofile.nl/images/dynamic/highlight/voc390.jpg\\r\\nhttp://farm4.static.flickr.com/3277/2640482050_55c904fe1f.jpg\\r\\nhttp://www.ihi.co.jp/iem/image/d-conversion.jpg\\r\\nhttp://www.repubblica.it/2007/08/sezioni/cronaca/nave-cnr/nave-cnr/este_03161528_15440.jpg\\r\\nhttp://archive.greenpeace.org/marine/sirius.jpg\\r\\nhttp://media-cdn.tripadvisor.com/media/photo-s/00/17/ea/be/booze-cruise-pirate-ship.jpg\\r\\nhttp://www.solstad.no/web/internet.nsf/(docs)/851D11F90E45E4BBC12573B7002A6F97/$FILE/ship.jpg\\r\\nhttp://farm4.static.flickr.com/3255/2706986094_43567e44fb.jpg\\r\\nhttp://farm4.static.flickr.com/3094/2297156101_9b58648208.jpg\\r\\nhttp://farm3.static.flickr.com/2109/2113356463_4de5dcbda8.jpg\\r\\nhttp://farm3.static.flickr.com/2324/2172715859_a9d3ee1ec4.jpg\\r\\nhttp://www.navytrieste.com/Nave_Espero_F_576/nave_espero_012.jpg\\r\\nhttp://static.flickr.com/155/411592676_871f6c31ef.jpg\\r\\nhttp://farm2.static.flickr.com/1346/907907833_ab4941b0e9.jpg\\r\\nhttp://farm1.static.flickr.com/137/321405543_ab7df9b078.jpg\\r\\nhttp://farm3.static.flickr.com/2263/1698530500_c0f6ae8778.jpg\\r\\nhttp://farm1.static.flickr.com/254/447323928_3149ea1ff7.jpg\\r\\nhttp://static.flickr.com/66/196935059_411208f5d4.jpg\\r\\nhttp://www.marina.difesa.it/unita/pagine/navi/sangiusto/images/01_small.jpg\\r\\nhttp://farm1.static.flickr.com/195/494913937_5eb8569a7c.jpg\\r\\n\\n\"\n",
            "b\"http://farm2.static.flickr.com/1089/537335018_9813e94757.jpg\\r\\nhttp://members.lycos.nl/zwawe/hpbimg/fiets3.jpg\\r\\nhttp://communities.canada.com/shareit/photos/bicycles/images/64838/425x286.aspx\\r\\nhttp://farm2.static.flickr.com/1090/1441125154_341230126c.jpg\\r\\nhttp://farm4.static.flickr.com/3219/2681908945_dbe4c38deb.jpg\\r\\nhttp://farm3.static.flickr.com/2186/2358375851_fa98bf5314.jpg\\r\\nhttp://farm3.static.flickr.com/2375/2463989860_28a9d6743c.jpg\\r\\nhttp://bluh.nl/images/portfolio/grafisch/fiets.jpg\\r\\nhttp://www.bakkeveen.nl/Ondernemers/Bolhuis/2003_3_22_A.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/a/a1/A_Bicycle_in_Oxford.JPG\\r\\nhttp://farm1.static.flickr.com/220/503379999_8301b3ed00.jpg\\r\\nhttp://farm3.static.flickr.com/2266/2539943000_005b996b1e.jpg\\r\\nhttp://home.kooee.com.au/nobananas/images/clayton-bike-1.jpg\\r\\nhttp://farm1.static.flickr.com/21/26188741_0a4a18d992.jpg\\r\\nhttp://farm3.static.flickr.com/2085/1642604624_4ca84a32c4.jpg\\r\\nhttp://www.bikeracks.us/images/bicycle-rack-480.jpg\\r\\nhttp://blog.makezine.com/Custom_Beach_Bike.jpg\\r\\nhttp://static.flickr.com/1118/605735667_9021756ea3.jpg\\r\\nhttp://www.bloggingcopenhagen.net/images/maria/may06/bicycle.jpg\\r\\nhttp://farm2.static.flickr.com/1387/681917428_1f767c8f48.jpg\\r\\nhttp://www.freephoto1.com/photo/photo-bicycle-3.jpg\\r\\nhttp://www.hypebeast.com/image/2008/03/walking-bike-max-knight-2.jpg\\r\\nhttp://static.flickr.com/1060/1079523963_a9d207db1f.jpg\\r\\nhttp://www.mijnmobielefietsenmaker.nl/images/opoe%20fiets.jpg\\r\\nhttp://www.raftri.org.uk/images/kit/louisbike.jpg\\r\\nhttp://static.flickr.com/26/52374684_faa75036e0.jpg\\r\\nhttp://bhznet.ods.org/bhznet/nieuws/2008/fiets_dr_de_jong.jpg\\r\\nhttp://slektsforskning.com/images/bicycle.jpg\\r\\nhttp://www.scienceshareware.com/articles/diy-byo-bike-generator-coachella/images/coachella-cell-phone-charger-bike-generators.jpg\\r\\nhttp://farm2.static.flickr.com/1122/1489498198_266203c606.jpg\\r\\nhttp://static.flickr.com/1398/539935043_061e6dc02c.jpg\\r\\nhttp://farm2.static.flickr.com/1388/770996708_e39a1be4c2.jpg\\r\\nhttp://farm2.static.flickr.com/1042/537336078_c9fa0bbc31.jpg\\r\\nhttp://www.pnk.com.cn/other/art_perform/image/bicycle%20feats1.JPG\\r\\nhttp://farm1.static.flickr.com/4/8284677_a79ebc283c.jpg\\r\\nhttp://farm3.static.flickr.com/2151/2545091901_a55857e00b.jpg\\r\\nhttp://farm1.static.flickr.com/96/218815287_4bd065ea90.jpg\\r\\nhttp://farm1.static.flickr.com/187/424303691_642d267746.jpg\\r\\nhttp://farm2.static.flickr.com/1212/1035528071_89a333ce9c.jpg\\r\\nhttp://boothillchoppers.com/boothill_bike-large-chrome.gif\\r\\nhttp://farm1.static.flickr.com/93/253840770_e929fcc324.jpg\\r\\nhttp://farm3.static.flickr.com/2176/2439902648_f405443d87.jpg\\r\\nhttp://www.apartmenttherapy.com/images/uploads/6-13-cycloc-bicycle-rollo.jpg\\r\\nhttp://farm3.static.flickr.com/2236/2109406787_e058f459f4.jpg\\r\\nhttp://static.flickr.com/1173/537741196_bb4fa12ffe.jpg\\r\\nhttp://farm2.static.flickr.com/1253/1366342164_84c22335ca.jpg\\r\\nhttp://www.abikestore.com/Merchant2/graphics/00000004/fuji-patrol-bike.jpg\\r\\nhttp://farm3.static.flickr.com/2017/2255172606_3972ebd8d6.jpg\\r\\nhttp://farm1.static.flickr.com/47/166471686_652d09ffbd.jpg\\r\\nhttp://www.japannewbie.com/images/journal/phat_bicycle/pimp_bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2416765321_5da5fae259.jpg\\r\\nhttp://farm1.static.flickr.com/26/65718856_4be779e500.jpg\\r\\nhttp://fotos.marktplaats.nl/kopen/f/8b/6Q4MaZriXniv1oBjpJwyEQ==.jpg\\r\\nhttp://farm1.static.flickr.com/234/524353354_1c582c147f.jpg\\r\\nhttp://farm4.static.flickr.com/3175/2495236674_02e57c97de.jpg\\r\\nhttp://www.freephoto1.com/photo/photo-bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/39/123467197_8888d4194f.jpg\\r\\nhttp://gizmodo.com/gadgets/images/bmc_cnt_bike.jpg\\r\\nhttp://static.flickr.com/79/276360647_65ff1ced3d.jpg\\r\\nhttp://farm1.static.flickr.com/21/24613090_737ddffe11.jpg\\r\\nhttp://farm1.static.flickr.com/45/178890059_00a6550475.jpg\\r\\nhttp://graphics8.nytimes.com/images/2007/07/10/nyregion/bike600.gif\\r\\nhttp://static.flickr.com/166/333144978_44f8db54ee.jpg\\r\\nhttp://farm1.static.flickr.com/53/139808954_f4807aaca7.jpg\\r\\nhttp://farm4.static.flickr.com/3087/2485940034_8871d7876a.jpg\\r\\nhttp://farm3.static.flickr.com/2293/2367388822_d3a5b2f455.jpg\\r\\nhttp://farm3.static.flickr.com/2078/2342328797_a67e8bee25.jpg\\r\\nhttp://farm1.static.flickr.com/100/308436598_adea2ee1a9.jpg\\r\\nhttp://farm3.static.flickr.com/2385/2197666898_691606b8d3.jpg\\r\\nhttp://farm2.static.flickr.com/1400/871614535_43d76cc3fd.jpg\\r\\nhttp://farm3.static.flickr.com/2283/2099419714_f1535c9c87.jpg\\r\\nhttp://farm3.static.flickr.com/2164/2011693248_a9f121c0b3.jpg\\r\\nhttp://farm2.static.flickr.com/1308/1398487316_9ddb8077a6.jpg\\r\\nhttp://farm1.static.flickr.com/17/23126696_ca81f42e6d.jpg\\r\\nhttp://farm2.static.flickr.com/1205/771288091_341fdb5b05.jpg\\r\\nhttp://farm1.static.flickr.com/51/161396613_5679861120.jpg\\r\\nhttp://farm1.static.flickr.com/29/56310475_60b308e92a.jpg\\r\\nhttp://static.flickr.com/23/94214717_a6379270c8.jpg\\r\\nhttp://farm2.static.flickr.com/1316/806464012_46b8f49e3b.jpg\\r\\nhttp://farm1.static.flickr.com/59/208409989_c9b2800013.jpg\\r\\nhttp://farm3.static.flickr.com/2013/2433346479_14ab32fa40.jpg\\r\\nhttp://farm1.static.flickr.com/152/411792418_5aafa0b4f8.jpg\\r\\nhttp://farm4.static.flickr.com/3149/2401138741_fde4bfd876.jpg\\r\\nhttp://farm3.static.flickr.com/2063/2414331598_414ab24004.jpg\\r\\nhttp://farm1.static.flickr.com/54/114893320_3b5260dc64.jpg\\r\\nhttp://img.alibaba.com/photo/100925639/Pinarello_Fp2_Aluminum_Road_Bike_With_Shimano_Ultegra_Components_Bicycle.jpg\\r\\nhttp://static.flickr.com/11/13714272_127b07ae63.jpg\\r\\nhttp://farm1.static.flickr.com/81/237944042_7aaacc608d.jpg\\r\\nhttp://farm2.static.flickr.com/1423/1157813343_3db570b0e2.jpg\\r\\nhttp://farm1.static.flickr.com/140/328016201_c3f7986227.jpg\\r\\nhttp://www.aerofast.com/aerofast/graphics/2500_lg.jpg\\r\\nhttp://www.jollygood.com/images/mtb2007/2nd%20diary/BAZZA%20TAIL%20WHIP%20KONA%20BIKE%20PARK.JPG\\r\\nhttp://static.flickr.com/40/125738021_d4e7a2d5ff.jpg\\r\\nhttp://farm1.static.flickr.com/243/462723992_fc622933f3.jpg\\r\\nhttp://blog.thisnext.com/storage/thisnext.com%20surfin%20bike.bmp\\r\\nhttp://www.ski-epic.com/amsterdam_bicycles/pu5b_man_three_children_and_dog_all_on_bike.jpg\\r\\nhttp://farm4.static.flickr.com/3269/2598632423_c53386171e.jpg\\r\\nhttp://farm4.static.flickr.com/3101/2579252782_ab91d99881.jpg\\r\\nhttp://farm3.static.flickr.com/2028/2229824418_0da85463f2.jpg\\r\\nhttp://farm3.static.flickr.com/2041/2267007857_e47b276a14.jpg\\r\\nhttp://static.flickr.com/41/75374694_1b93400e35.jpg\\r\\nhttp://static.flickr.com/169/388259694_286b240c76.jpg\\r\\nhttp://farm3.static.flickr.com/2210/2073935487_ea304ec64d.jpg\\r\\nhttp://farm1.static.flickr.com/208/468473391_79c176c898.jpg\\r\\nhttp://www.atomicrack.com/Bicycle-Rack-Images/Bicycle-Rack-In-House.jpg\\r\\nhttp://static.flickr.com/1125/1094868168_303b099264.jpg\\r\\nhttp://farm4.static.flickr.com/3152/2462545360_ebe66f0fe0.jpg\\r\\nhttp://farm1.static.flickr.com/46/142402524_30ca16ef75.jpg\\r\\nhttp://farm1.static.flickr.com/51/148631945_4b0e795923.jpg\\r\\nhttp://z.about.com/d/bicycling/1/5/v/4/-/-/pt_cruiser_beach_bike.jpg\\r\\nhttp://farm3.static.flickr.com/2262/2284034781_d2c7a1c630.jpg\\r\\nhttp://images.buzzillions.com/images_products/01/52/marin_novato_bike_07_closeout_reviews_1179057_300.jpg\\r\\nhttp://farm2.static.flickr.com/1376/530709833_b48e71159c.jpg\\r\\nhttp://farm4.static.flickr.com/3020/2628139772_d81a185253.jpg\\r\\nhttp://farm3.static.flickr.com/2169/2256348962_90312a4f78.jpg\\r\\nhttp://farm1.static.flickr.com/78/161396411_489683f0ba.jpg\\r\\nhttp://farm1.static.flickr.com/26/53007868_696531f997.jpg\\r\\nhttp://farm1.static.flickr.com/49/152528259_31980b87e9.jpg\\r\\nhttp://farm1.static.flickr.com/1/301279_55dd3c1d7f.jpg\\r\\nhttp://farm1.static.flickr.com/89/220493498_6607e73a2b.jpg\\r\\nhttp://farm4.static.flickr.com/3154/2530613101_97aa40692d.jpg\\r\\nhttp://farm4.static.flickr.com/3066/2592855120_fcd55c2e23.jpg\\r\\nhttp://static.flickr.com/1009/548683369_f0f50dd487.jpg\\r\\nhttp://farm1.static.flickr.com/233/517274301_3030ee8ccd.jpg\\r\\nhttp://farm4.static.flickr.com/3045/2282418707_f5d3920c35.jpg\\r\\nhttp://i1.trekearth.com/photos/12282/bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/44/113893980_5518314749.jpg\\r\\nhttp://farm1.static.flickr.com/1/1042449_692e7502be.jpg\\r\\nhttp://farm3.static.flickr.com/2146/1777855778_a50f4703a9.jpg\\r\\nhttp://farm2.static.flickr.com/1296/1234288300_f45cbe4939.jpg\\r\\nhttp://farm1.static.flickr.com/209/520416998_e354588c32.jpg\\r\\nhttp://img.2dehands.be/f/view/38021302-piste-fiets-ridley-oval-2007.jpg\\r\\nhttp://farm2.static.flickr.com/1255/585128165_d8fdf6dbb4.jpg\\r\\nhttp://farm3.static.flickr.com/2138/2212952263_b06229a268.jpg\\r\\nhttp://farm4.static.flickr.com/3165/2332757847_7b931da5f9.jpg\\r\\nhttp://static.flickr.com/61/154936093_0bf214600e.jpg\\r\\nhttp://farm2.static.flickr.com/1100/536334391_f9f9da92c6.jpg\\r\\nhttp://farm1.static.flickr.com/37/112212389_16a3fb8d04.jpg\\r\\nhttp://farm2.static.flickr.com/1069/1304417567_ec1d461f0c.jpg\\r\\nhttp://www.tecaeromex.com/imagenes/rocket-bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/45/164131219_5a68057be9.jpg\\r\\nhttp://farm1.static.flickr.com/224/489005496_f92a5b2dc3.jpg\\r\\nhttp://farm1.static.flickr.com/77/167688023_5b941caf48.jpg\\r\\nhttp://farm1.static.flickr.com/54/141232626_84571f6ff0.jpg\\r\\nhttp://www.cascadebicyclestudio.com/img/billboard.jpg\\r\\nhttp://farm3.static.flickr.com/2144/2156223654_e1f12cb82c.jpg\\r\\nhttp://static.flickr.com/51/148631945_4b0e795923.jpg\\r\\nhttp://www.lovehkfilm.com/panasia/aj6293/beijing_bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/23/29297545_593a2c930d.jpg\\r\\nhttp://farm1.static.flickr.com/107/311738880_4102d67bdb.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/7/72/Womens-Bicycle-Race-Thuringia-2006.jpg\\r\\nhttp://www.thespiderawards.com/2007presentation/photoshow/nominations/24_silhouette/images/smart-sheila_bicycle-silhouette.jpg\\r\\nhttp://static.flickr.com/95/280347673_81caa34b0c.jpg\\r\\nhttp://static.flickr.com/2144/1923487358_bd176ac513.jpg\\r\\nhttp://farm3.static.flickr.com/2368/1515481757_1901c20f02.jpg\\r\\nhttp://farm4.static.flickr.com/3257/2453287040_fc3c07f776.jpg\\r\\nhttp://static.flickr.com/215/551227594_0a9d45d65c.jpg\\r\\nhttp://farm3.static.flickr.com/2189/2011770058_1375c081ed.jpg\\r\\nhttp://farm2.static.flickr.com/1199/537451983_93d92bcebf.jpg\\r\\nhttp://farm1.static.flickr.com/174/415957295_4f1e1ea65c.jpg\\r\\nhttp://farm3.static.flickr.com/2234/2222790283_b3327098a9.jpg\\r\\nhttp://static.flickr.com/117/279930013_b6eb1e364f.jpg\\r\\nhttp://farm1.static.flickr.com/163/331912830_0b8bef7a85.jpg\\r\\nhttp://static.flickr.com/52/127443097_231e00fd83.jpg\\r\\nhttp://static.flickr.com/165/373356640_630b3a1781.jpg\\r\\nhttp://static.flickr.com/1095/640953777_839210a689.jpg\\r\\nhttp://farm3.static.flickr.com/2170/2450482306_da9ca27265.jpg\\r\\nhttp://farm2.static.flickr.com/1286/853677031_530bc097b0.jpg\\r\\nhttp://farm1.static.flickr.com/84/264340201_7b70127d5b.jpg\\r\\nhttp://farm1.static.flickr.com/170/422259410_37e2e3d8d8.jpg\\r\\nhttp://www.metaefficient.com/wordpress/wp-content/uploads/2007/12/strida_5_folding_bike.jpg\\r\\nhttp://img.alibaba.com/photo/11060089/Chopper_Bicycle.jpg\\r\\nhttp://www.joskaal.nl/nl/images/stories/tda_fiets_gecom.jpg\\r\\nhttp://www.middelkerke.be/toerisme/nieuw/images/fietsers.jpg\\r\\nhttp://farm3.static.flickr.com/2165/2485906564_421fc330f7.jpg\\r\\nhttp://farm2.static.flickr.com/1232/1010688789_8824194e3c.jpg\\r\\nhttp://farm3.static.flickr.com/2133/2092823034_25fef82716.jpg\\r\\nhttp://farm4.static.flickr.com/3255/2414739612_abdbed7a56.jpg\\r\\nhttp://static.flickr.com/57/175374245_bb285893d2.jpg\\r\\nhttp://farm1.static.flickr.com/27/51966134_a4883ef500.jpg\\r\\nhttp://farm3.static.flickr.com/2070/2399200405_e781e207bb.jpg\\r\\nhttp://farm3.static.flickr.com/2121/2430605371_2111abc823.jpg\\r\\nhttp://farm3.static.flickr.com/2092/2338532088_6cf3a5be72.jpg\\r\\nhttp://farm1.static.flickr.com/8/8643700_dd099b1cac.jpg\\r\\nhttp://static.flickr.com/2/1456113_38c3235301.jpg\\r\\nhttp://farm3.static.flickr.com/2265/2274660194_79d8af4ece.jpg\\r\\nhttp://farm2.static.flickr.com/1127/1056307083_cde066b65d.jpg\\r\\nhttp://farm2.static.flickr.com/1289/604919059_756681cb7e.jpg\\r\\nhttp://static.flickr.com/117/367199683_2493946632.jpg\\r\\nhttp://farm4.static.flickr.com/3115/2563117055_1155274f4b.jpg\\r\\nhttp://farm2.static.flickr.com/1322/751576420_72fbe9d01c.jpg\\r\\nhttp://farm3.static.flickr.com/2382/2310736848_1ecf2044ae.jpg\\r\\nhttp://farm1.static.flickr.com/191/486304152_50a2977164.jpg\\r\\nhttp://farm1.static.flickr.com/176/483380111_0336277e20.jpg\\r\\nhttp://img.2dehands.nl/f/view/32544954-sparta-ion-elektrische-fiets-nieuw.jpg\\r\\nhttp://farm3.static.flickr.com/2108/2211782503_70e275cc98.jpg\\r\\nhttp://farm1.static.flickr.com/93/238903523_1899dfcd13.jpg\\r\\nhttp://static.flickr.com/98/237370680_7722a225fe.jpg\\r\\nhttp://fietsvakanties.web-log.nl/fietsvakanties/images/2008/04/24/drentsefietsvierdaagseschapen.jpg\\r\\nhttp://farm4.static.flickr.com/3044/2669912045_a877c0ecd4.jpg\\r\\nhttp://static.flickr.com/83/241248855_46f1227587.jpg\\r\\nhttp://farm1.static.flickr.com/194/482307332_476cc19394.jpg\\r\\nhttp://farm3.static.flickr.com/2097/2054006848_475ba6e8f9.jpg\\r\\nhttp://bikeforpeace.org/images/bike_move-nov_2005/bike_move_nov_2005_boulder%20(22).jpg\\r\\nhttp://farm4.static.flickr.com/3210/2323257635_1d6991e81a.jpg\\r\\nhttp://static.flickr.com/2318/1762518012_c25aa1f92b.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/7/73/Herzl_bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3177/2394247910_99f2a050c3.jpg\\r\\nhttp://farm3.static.flickr.com/2197/2182735915_78bdcff93d.jpg\\r\\nhttp://farm1.static.flickr.com/65/221421852_e0af7d858f.jpg\\r\\nhttp://farm1.static.flickr.com/169/411793134_b8390d255e.jpg\\r\\nhttp://www.sportenzo.com/contents/media/fiets.jpg\\r\\nhttp://farm1.static.flickr.com/8/8173265_6eb27ebae6.jpg\\r\\nhttp://www.szlgnews.com/jdxl/images/001558d90b3307cb63df02.JPG\\r\\nhttp://farm1.static.flickr.com/223/517248558_0c581913fe.jpg\\r\\nhttp://farm1.static.flickr.com/213/496023108_bd916b3817.jpg\\r\\nhttp://i.treehugger.com/images/2007/10/24/baum-bike-detail.jpg\\r\\nhttp://farm1.static.flickr.com/95/214575396_8d0ca34f9d.jpg\\r\\nhttp://farm1.static.flickr.com/53/171680719_57c97cec77.jpg\\r\\nhttp://farm1.static.flickr.com/57/161396591_c81a7b5783.jpg\\r\\nhttp://www.mtbr.com/reviews/Bike/Raleigh_M800_Bike.jpg\\r\\nhttp://farm3.static.flickr.com/2296/2490591019_6223fe8b29.jpg\\r\\nhttp://farm1.static.flickr.com/111/306102827_6ab087135d.jpg\\r\\nhttp://static.flickr.com/172/400438492_2cdb6b6e90.jpg\\r\\nhttp://farm3.static.flickr.com/2361/2536927595_63fb5da29d.jpg\\r\\nhttp://www.treehugger.com/areaware-noel-strida-bike-headshot.jpg\\r\\nhttp://farm3.static.flickr.com/2057/2092678594_1f542a2978.jpg\\r\\nhttp://farm1.static.flickr.com/24/52507084_7b6cae49a1.jpg\\r\\nhttp://farm1.static.flickr.com/177/458930384_96b323c418.jpg\\r\\nhttp://farm2.static.flickr.com/1337/537451263_9d659cd346.jpg\\r\\nhttp://www.thelensflare.com/large/bicycle_2392.jpg\\r\\nhttp://farm4.static.flickr.com/3043/2642779150_806e77d720.jpg\\r\\nhttp://farm1.static.flickr.com/168/411794146_d8a4fadcaa.jpg\\r\\nhttp://farm1.static.flickr.com/34/73776972_cdd72a49d9.jpg\\r\\nhttp://farm1.static.flickr.com/31/171677267_f059e846b1.jpg\\r\\nhttp://farm1.static.flickr.com/242/550443342_6ff8a2ed62.jpg\\r\\nhttp://www.hansrey.com/images/Hans_trials_bike.jpg\\r\\nhttp://farm1.static.flickr.com/6/8593342_f426aa90f7.jpg\\r\\nhttp://farm1.static.flickr.com/79/249205924_bc3cbde5f2.jpg\\r\\nhttp://static.flickr.com/215/505942057_4448afe31d.jpg\\r\\nhttp://farm1.static.flickr.com/82/262887715_9a8d34738b.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/a/a0/Itera_plastic_bicycle.jpg\\r\\nhttp://farm3.static.flickr.com/2017/2220013685_95878c521f.jpg\\r\\nhttp://www.kailashmanasarover.com/images/bikesandhikes.jpg\\r\\nhttp://farm1.static.flickr.com/43/108789541_c8f9d53e41.jpg\\r\\nhttp://www.graphpaper.com/wp/wp-content/uploads/2007/06/arrow-bicycle.jpg\\r\\nhttp://www.gadgetgarden.nl/beeld/Gucci_fiets_olympische.jpg\\r\\nhttp://farm1.static.flickr.com/33/58498441_1a1fdb3e83.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/9/9c/Frieght_bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/16/20817767_b31d03e32c.jpg\\r\\nhttp://www.scientologyhandbook.org/img/p13_2.jpg\\r\\nhttp://farm1.static.flickr.com/31/52368761_32eab05fe9.jpg\\r\\nhttp://www.britishcycling.org.uk/web/MultimediaFiles/20060328_JOSH_BRYCELAND_AND_BIKE_400.JPG\\r\\nhttp://static.flickr.com/109/286665506_29b87aacdc.jpg\\r\\nhttp://farm1.static.flickr.com/134/333818977_bf87219dd6.jpg\\r\\nhttp://farm1.static.flickr.com/110/298296548_1b255aca0e.jpg\\r\\nhttp://farm4.static.flickr.com/3152/2587912040_2d0e900865.jpg\\r\\nhttp://etniesbmx.com/site-images/news/article/202736aaron_ross_bike_01.jpg\\r\\nhttp://farm3.static.flickr.com/2295/2509946461_05f8034df8.jpg\\r\\nhttp://farm3.static.flickr.com/2028/2138777528_b3cdecf23b.jpg\\r\\nhttp://farm3.static.flickr.com/2243/2490690324_945b8c7214.jpg\\r\\nhttp://farm3.static.flickr.com/2188/2379700108_4419a09433.jpg\\r\\nhttp://farm4.static.flickr.com/3196/2403216626_eba4ef41d0.jpg\\r\\nhttp://farm3.static.flickr.com/2161/2091898013_23a8197085.jpg\\r\\nhttp://farm1.static.flickr.com/136/351424919_51dcfe23ed.jpg\\r\\nhttp://www.weblogcartoons.com/cb/a-bike-evangelist.jpg\\r\\nhttp://farm1.static.flickr.com/5/10712153_60902f5531.jpg\\r\\nhttp://farm1.static.flickr.com/22/30385313_8cd4fc3c30.jpg\\r\\nhttp://farm3.static.flickr.com/2122/2239116607_b8cb43a999.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/0/01/Crushed_Bike.JPG/800px-Crushed_Bike.JPG\\r\\nhttp://farm1.static.flickr.com/175/427088506_0f166a1a5e.jpg\\r\\nhttp://farm1.static.flickr.com/17/21133087_d4c9541ed5.jpg\\r\\nhttp://farm2.static.flickr.com/1003/536966880_5065a99d1d.jpg\\r\\nhttp://farm1.static.flickr.com/65/168615275_b29cd6c189.jpg\\r\\nhttp://farm3.static.flickr.com/2113/2508552144_515e09ef68.jpg\\r\\nhttp://farm1.static.flickr.com/43/129183752_43bed0646a.jpg\\r\\nhttp://farm1.static.flickr.com/70/203418977_94309432bd.jpg\\r\\nhttp://farm4.static.flickr.com/3043/2700703001_628f525212.jpg\\r\\nhttp://www.liveforjesus.org/images/bicycle%20&%20gloves.jpg\\r\\nhttp://img.photobucket.com/albums/v422/Jos/Picture013_28Jan06.jpg\\r\\nhttp://farm2.static.flickr.com/1140/965631678_2d9b3d0364.jpg\\r\\nhttp://farm2.static.flickr.com/1362/996366850_a329e95ace.jpg\\r\\nhttp://farm4.static.flickr.com/3186/2526825056_95bb9903f6.jpg\\r\\nhttp://farm3.static.flickr.com/2138/2313936378_b43b3ef09e.jpg\\r\\nhttp://farm1.static.flickr.com/1/122066370_1b0bc6cf6d.jpg\\r\\nhttp://farm1.static.flickr.com/253/521264930_ae3cc6f2a4.jpg\\r\\nhttp://farm1.static.flickr.com/48/139778256_93fe6e59f3.jpg\\r\\nhttp://members.lycos.nl/zwawe/hpbimg/bicycle.gif\\r\\nhttp://images.enjoywhistler.com/whistlerReservations/whistlerMountainBike1.jpg\\r\\nhttp://farm1.static.flickr.com/174/366097965_4088fae251.jpg\\r\\nhttp://farm1.static.flickr.com/37/119300296_1f656275ab.jpg\\r\\nhttp://farm3.static.flickr.com/2190/2536928319_def3d5ee9d.jpg\\r\\nhttp://farm3.static.flickr.com/2175/1777008697_7ab852cfa5.jpg\\r\\nhttp://farm3.static.flickr.com/2307/1602963686_2c828a3139.jpg\\r\\nhttp://www.wingene.be/_STUDIOEMMA_WWW/uploads/Image/Nieuws/2008_05/KoersOpRollen/Aankondiging/Fiets.jpg\\r\\nhttp://farm3.static.flickr.com/2379/2539655388_6d3c6a10f0.jpg\\r\\nhttp://farm1.static.flickr.com/38/74462190_5f04dc867e.jpg\\r\\nhttp://farm1.static.flickr.com/141/388298079_2afa9e8b24.jpg\\r\\nhttp://bikehugger.com/images/blog/rattan_bike.jpg\\r\\nhttp://static.flickr.com/169/416409900_2ee47b95b9.jpg\\r\\nhttp://www.abikestore.com/Merchant2/graphics/00000004/sw-tactical-bike.jpg\\r\\nhttp://farm3.static.flickr.com/2076/2197897958_7f9d32cd1d.jpg\\r\\nhttp://farm1.static.flickr.com/79/216012106_a8bcba2a27.jpg\\r\\nhttp://farm3.static.flickr.com/2185/2272222737_5710fbcaaa.jpg\\r\\nhttp://farm1.static.flickr.com/82/238906463_7ea39958d1.jpg\\r\\nhttp://static.flickr.com/161/356563315_203d0500c0.jpg\\r\\nhttp://farm1.static.flickr.com/81/237280022_2a9da3c166.jpg\\r\\nhttp://farm1.static.flickr.com/6/6006676_2145d67b70.jpg\\r\\nhttp://farm1.static.flickr.com/232/514283910_181e7a5824.jpg\\r\\nhttp://farm4.static.flickr.com/3130/2588795677_471e569402.jpg\\r\\nhttp://static.flickr.com/3/3610353_b8e7216b83.jpg\\r\\nhttp://www.thelensflare.com/large/old-time-bicycle_45990.jpg\\r\\nhttp://static.flickr.com/38/102733145_8ba6d678d7.jpg\\r\\nhttp://farm1.static.flickr.com/55/138362081_24325e57b8.jpg\\r\\nhttp://www.fmbrussel.be/assets/img/varia/08fiets/080521fiets.jpg\\r\\nhttp://www.gadgetgarden.nl/Puma_glow-in-the-dark_bike_fiets_a-thumb.jpg\\r\\nhttp://farm3.static.flickr.com/2399/2018753663_e3d702fa10.jpg\\r\\nhttp://farm1.static.flickr.com/75/179790162_9ae331b9cc.jpg\\r\\nhttp://farm4.static.flickr.com/3155/2684610599_18f5fc9e87.jpg\\r\\nhttp://farm4.static.flickr.com/3100/2670985775_d011bf56a6.jpg\\r\\nhttp://static.flickr.com/2/2156016_a0da35e7e0.jpg\\r\\nhttp://farm1.static.flickr.com/72/189900965_aa74dce707.jpg\\r\\nhttp://www.alternativewireless.com/krusell/krusell-images/krusell-bicycle-holder.jpg\\r\\nhttp://farm4.static.flickr.com/3067/2672813268_ba541a20c7.jpg\\r\\nhttp://www.opinios.com/images/other/DanishBikeLock.jpg\\r\\nhttp://farm1.static.flickr.com/80/235913416_4e64099e56.jpg\\r\\nhttp://www.reinharts.net/images/ItalianBicycle.jpg\\r\\nhttp://bmxmuseum.com/image/8.20.03.bike_lg.jpg\\r\\nhttp://farm3.static.flickr.com/2089/2211504309_1615559f19.jpg\\r\\nhttp://farm3.static.flickr.com/2374/2399973093_fac4419910.jpg\\r\\nhttp://static.flickr.com/52/191293774_efbcf77043.jpg\\r\\nhttp://static.flickr.com/78/180392066_866dad2283.jpg\\r\\nhttp://www.cbc.ca/gfx/topstory/montgomery0915.jpg\\r\\nhttp://farm3.static.flickr.com/2255/2075414128_cf5f4c9f9e.jpg\\r\\nhttp://static.flickr.com/48/117123178_4377f813fb.jpg\\r\\nhttp://i141.photobucket.com/albums/r46/Motorbiker_photos/NewsPics2/Harley-Davidson-Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/183/411795436_7ac20d891d.jpg\\r\\nhttp://farm3.static.flickr.com/2360/1915565936_c9f67c49b5.jpg\\r\\nhttp://farm3.static.flickr.com/2351/1676264435_e51b3682a2.jpg\\r\\nhttp://farm3.static.flickr.com/2219/2539829404_6e1e72871d.jpg\\r\\nhttp://farm1.static.flickr.com/153/355840159_5b6116603c.jpg\\r\\nhttp://www.stirn-vanham.com/Later_pages/Photo_Archiv/Children/Bicycle.jpg\\r\\nhttp://farm3.static.flickr.com/2221/2173621527_c67a92ec53.jpg\\r\\nhttp://byfiles.storage.live.com/y1pCAGuF2zovk3KmLM3M43SYv7uyD-ZeJ6UcWn9H4SHYlh8luQHUF6jNznOMoCbJADYhFrIWn0vnIg\\r\\nhttp://static.flickr.com/45/141137996_f366d15184.jpg\\r\\nhttp://farm1.static.flickr.com/178/394084247_af88a1d9fd.jpg\\r\\nhttp://farm4.static.flickr.com/3132/2481010079_e93eaebf5c.jpg\\r\\nhttp://farm4.static.flickr.com/3105/2688544514_cfd64ef6a9.jpg\\r\\nhttp://farm3.static.flickr.com/2375/2386877368_6181001827.jpg\\r\\nhttp://farm1.static.flickr.com/61/165374189_ec1fddeb31.jpg\\r\\nhttp://static.flickr.com/173/436375711_4ac997f9fc.jpg\\r\\nhttp://farm2.static.flickr.com/1318/1006967417_22c88e6f05.jpg\\r\\nhttp://farm1.static.flickr.com/41/101785441_67ab962d33.jpg\\r\\nhttp://farm4.static.flickr.com/3174/2559517077_687c0b58d6.jpg\\r\\nhttp://farm2.static.flickr.com/1153/591290644_7d8660d469.jpg\\r\\nhttp://farm1.static.flickr.com/7/8190940_6303a42644.jpg\\r\\nhttp://img.photobucket.com/albums/v159/zaan944/fun/airco-op-fiets.jpg\\r\\nhttp://farm1.static.flickr.com/202/479404349_b11a4fd74a.jpg\\r\\nhttp://www.pedalinghistory.com/Images/ShwBkFtm.gif\\r\\nhttp://farm1.static.flickr.com/48/116080340_ea3bc42b9d.jpg\\r\\nhttp://farm4.static.flickr.com/3271/2394071695_3bdcd3d01b.jpg\\r\\nhttp://farm2.static.flickr.com/1199/804474351_28bfa32ff0.jpg\\r\\nhttp://farm1.static.flickr.com/131/389462733_13beef4199.jpg\\r\\nhttp://static.flickr.com/2070/2126683137_2ba0e09106.jpg\\r\\nhttp://farm1.static.flickr.com/8/8518681_cc5dbef755.jpg\\r\\nhttp://static.zooomr.com/images/4682559_cad326077d.jpg\\r\\nhttp://farm1.static.flickr.com/105/300626850_bfb8d15784.jpg\\r\\nhttp://static.flickr.com/163/391319158_e580825d91.jpg\\r\\nhttp://www.toystore.info/images/Polaris-Scrambler-Bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3072/2284032629_6922a80e3d.jpg\\r\\nhttp://farm4.static.flickr.com/3284/2434731827_30b3c7f415.jpg\\r\\nhttp://static.flickr.com/161/413380853_e443953d80.jpg\\r\\nhttp://farm2.static.flickr.com/1266/1259666254_245fab7b6d.jpg\\r\\nhttp://farm2.static.flickr.com/1135/1385598687_cfd4c97945.jpg\\r\\nhttp://farm4.static.flickr.com/3128/2540709816_8293e2c6f1.jpg\\r\\nhttp://farm4.static.flickr.com/3138/2331398841_936e0e3063.jpg\\r\\nhttp://static.flickr.com/1312/705660319_dc195aba00.jpg\\r\\nhttp://farm3.static.flickr.com/2248/2288094961_057eac1e16.jpg\\r\\nhttp://static.flickr.com/141/407577417_07d404b8c9.jpg\\r\\nhttp://www.maartenklijnstra.nl/i/materiaal/fiets.jpg\\r\\nhttp://www.geocities.com/SiliconValley/3839/bike/bike_amf_rdmaster_6a_550h.jpg\\r\\nhttp://static.flickr.com/53/143339472_4d9c93836f.jpg\\r\\nhttp://moworldphotos.com/imagesssdp/bicycle_place_silver_spring.jpg\\r\\nhttp://farm1.static.flickr.com/5/10681211_fa1ed324b7.jpg\\r\\nhttp://farm1.static.flickr.com/169/490979743_76ea264bdc.jpg\\r\\nhttp://farm1.static.flickr.com/52/158398466_ee96830692.jpg\\r\\nhttp://static.flickr.com/1437/1455097012_c6331bc229.jpg\\r\\nhttp://www.localhistory.scit.wlv.ac.uk/Museum/Transport/bicycles/onward2.jpg\\r\\nhttp://farm3.static.flickr.com/2217/2284821318_e9fd78095a.jpg\\r\\nhttp://www.horizonsunlimited.com/newsletter/images/2002-05-01_Frazier_AMoturCycle.jpg\\r\\nhttp://images.asia.ru/img/alibaba/photo/51675505/Electric_Bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3290/2435978673_55524d9ee3.jpg\\r\\nhttp://farm1.static.flickr.com/11/14624146_fbca109687.jpg\\r\\nhttp://static.flickr.com/1255/716752413_fa4a834740.jpg\\r\\nhttp://i2.photobucket.com/albums/y10/corinecolijn/fiets.jpg\\r\\nhttp://farm1.static.flickr.com/12/18527096_216b36619a.jpg\\r\\nhttp://static.flickr.com/1391/1016816239_82fcb3000d.jpg\\r\\nhttp://farm2.static.flickr.com/1029/945554794_7312e3b88c.jpg\\r\\nhttp://farm1.static.flickr.com/58/171679517_8ac3e8f3ef.jpg\\r\\nhttp://farm2.static.flickr.com/1102/864444338_a0c21a72fc.jpg\\r\\nhttp://farm3.static.flickr.com/2001/2431430947_d83be53af6.jpg\\r\\nhttp://farm1.static.flickr.com/145/360353888_200a3bc478.jpg\\r\\nhttp://farm1.static.flickr.com/83/243906555_7984fe5489.jpg\\r\\nhttp://farm2.static.flickr.com/1294/1399623170_4eb31b2811.jpg\\r\\nhttp://farm4.static.flickr.com/3189/2291964434_c3c459e9a9.jpg\\r\\nhttp://farm3.static.flickr.com/2113/2230406882_43a7d6941e.jpg\\r\\nhttp://farm3.static.flickr.com/2070/2069915996_990712e2ea.jpg\\r\\nhttp://www.dlmark.net/vancouvercollage/Bicycle%20Rack.jpg\\r\\nhttp://farm1.static.flickr.com/131/411802713_753c3ffb03.jpg\\r\\nhttp://farm3.static.flickr.com/2154/2080611502_88c9746c03.jpg\\r\\nhttp://farm2.static.flickr.com/1177/1482550155_8a5af6ee1a.jpg\\r\\nhttp://static.flickr.com/53/156786585_8f71ff63ff.jpg\\r\\nhttp://farm3.static.flickr.com/2287/2064023465_562893ae46.jpg\\r\\nhttp://farm1.static.flickr.com/231/498573244_67d1a55add.jpg\\r\\nhttp://farm1.static.flickr.com/216/453552406_f70e8156cc.jpg\\r\\nhttp://farm2.static.flickr.com/1202/1281906948_60ea4578ae.jpg\\r\\nhttp://www.wikihow.com/images/3/3f/Bcycldrft.jpg\\r\\nhttp://farm3.static.flickr.com/2347/2537744590_4b5745750b.jpg\\r\\nhttp://www.dita.org/bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/214/468184022_31c4753fc4.jpg\\r\\nhttp://media.justjared.com/headlines/2007/12/brad-pitt-bicycle-built-for-four.jpg\\r\\nhttp://farm3.static.flickr.com/2253/1790288053_37ab82a01b.jpg\\r\\nhttp://static.flickr.com/167/488549655_b281713ec1.jpg\\r\\nhttp://farm1.static.flickr.com/26/51455338_4bf867907b.jpg\\r\\nhttp://farm4.static.flickr.com/3012/2553181604_4f1de7eeb8.jpg\\r\\nhttp://newsimg.bbc.co.uk/media/images/41615000/jpg/_41615736_cameron_bike203.jpg\\r\\nhttp://farm3.static.flickr.com/2174/2731597287_c30f3621e9.jpg\\r\\nhttp://farm4.static.flickr.com/3149/2578681427_cd80c24f61.jpg\\r\\nhttp://static.flickr.com/41/107759625_d86d8b8a0b.jpg\\r\\nhttp://static.flickr.com/72/208929211_1d5f922df6.jpg\\r\\nhttp://farm1.static.flickr.com/197/495397055_0dd5ea3b0e.jpg\\r\\nhttp://farm3.static.flickr.com/2261/1755992749_7e512596b3.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/2/20/BMX_bicycle.JPG/800px-BMX_bicycle.JPG\\r\\nhttp://farm2.static.flickr.com/1285/530709701_6e5e165fcd.jpg\\r\\nhttp://farm1.static.flickr.com/198/464900866_1e3644d039.jpg\\r\\nhttp://en.wikivisual.com/images/7/70/Bicycle_two_1886.jpg\\r\\nhttp://farm1.static.flickr.com/8/9845986_79b46e3b05.jpg\\r\\nhttp://www.carlosspaventa.com/images/bicycle-w-flower-motif-v.jpeg\\r\\nhttp://farm1.static.flickr.com/149/379396093_e8d4bcde50.jpg\\r\\nhttp://farm3.static.flickr.com/2080/2366820017_1965fcb6c7.jpg\\r\\nhttp://farm3.static.flickr.com/2174/1517350095_956fa1ce12.jpg\\r\\nhttp://farm1.static.flickr.com/3/4004144_40d23e595e.jpg\\r\\nhttp://farm3.static.flickr.com/2374/2200049049_b4074fb22a.jpg\\r\\nhttp://farm3.static.flickr.com/2044/2400804648_2ebe1d3a15.jpg\\r\\nhttp://farm1.static.flickr.com/22/24673088_8ff1d02042.jpg\\r\\nhttp://farm2.static.flickr.com/1397/542301600_7f2d47ba5a.jpg\\r\\nhttp://farm4.static.flickr.com/3171/2399200409_b779e9d919.jpg\\r\\nhttp://farm1.static.flickr.com/47/139234214_5bd0a3e488.jpg\\r\\nhttp://farm3.static.flickr.com/2069/2128353623_729e8e6bba.jpg\\r\\nhttp://farm1.static.flickr.com/26/96916162_556228f65b.jpg\\r\\nhttp://farm3.static.flickr.com/2013/2274183149_888ed0bcd0.jpg\\r\\nhttp://farm3.static.flickr.com/2181/2275160682_7e6f08200c.jpg\\r\\nhttp://farm2.static.flickr.com/1255/1413051171_e95f73eb7f.jpg\\r\\nhttp://farm1.static.flickr.com/28/38924436_205496f1d3.jpg\\r\\nhttp://farm4.static.flickr.com/3166/2692412306_9c4f4bf464.jpg\\r\\nhttp://farm1.static.flickr.com/129/360545640_4b99336de6.jpg\\r\\nhttp://farm2.static.flickr.com/1003/530748434_4bcb7e9ef6.jpg\\r\\nhttp://farm3.static.flickr.com/2304/2091898257_bb444fc65f.jpg\\r\\nhttp://farm1.static.flickr.com/53/134914360_4234659799.jpg\\r\\nhttp://farm1.static.flickr.com/203/495917504_57231c0f83.jpg\\r\\nhttp://farm3.static.flickr.com/2187/1736910214_fafd285f5a.jpg\\r\\nhttp://farm1.static.flickr.com/183/363695635_2887071d4b.jpg\\r\\nhttp://farm1.static.flickr.com/72/156819512_1ecf9e0f6a.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/e/ed/U_Wash_bicycle_shop_03A.jpg/800px-U_Wash_bicycle_shop_03A.jpg\\r\\nhttp://static.flickr.com/1347/1332028809_bc6b7e89e0.jpg\\r\\nhttp://www.weirdwildrealm.com/filmimages/bicycletrick1899.jpg\\r\\nhttp://farm1.static.flickr.com/49/110080206_cd33f0bbbe.jpg\\r\\nhttp://farm3.static.flickr.com/2044/2078458826_c64da49ff1.jpg\\r\\nhttp://www.rqriley.com/images/JHR-bike-big.jpg\\r\\nhttp://farm3.static.flickr.com/2381/2156993373_8c09442d35.jpg\\r\\nhttp://farm4.static.flickr.com/3027/2385412890_dc5c6bba8e.jpg\\r\\nhttp://farm1.static.flickr.com/13/94357703_a11207eb5b.jpg\\r\\nhttp://farm2.static.flickr.com/1108/1019701447_8283477eeb.jpg\\r\\nhttp://home.tiscali.nl/fotoboekroemenie/Gorj/Targu%20Jiu%20oma%20met%20fiets.jpg\\r\\nhttp://farm3.static.flickr.com/2146/2479617836_c3296f2db9.jpg\\r\\nhttp://farm4.static.flickr.com/3188/2327691064_a830f3bb7b.jpg\\r\\nhttp://farm2.static.flickr.com/1374/1052870912_34dfcebace.jpg\\r\\nhttp://farm2.static.flickr.com/1279/1030748599_7ca082f18b.jpg\\r\\nhttp://www.6smitjes.nl/userfiles/koen%20jaco%20fiets.jpg\\r\\nhttp://farm1.static.flickr.com/32/62355991_774fbf1464.jpg\\r\\nhttp://farm3.static.flickr.com/2218/2225844129_9524681ff8.jpg\\r\\nhttp://abewing.nycap.rr.com/judy%20images%20for%20web%20page/images%20assembly/sept%2006%20bmx%20balance%20bike.gif\\r\\nhttp://farm2.static.flickr.com/1367/1351365656_fb9ec6ae17.jpg\\r\\nhttp://pedaleiro.files.wordpress.com/2007/08/cart_bike.jpg\\r\\nhttp://farm1.static.flickr.com/27/57165602_ef7a4b33ec.jpg\\r\\nhttp://static.flickr.com/76/231017800_4c0b745245.jpg\\r\\nhttp://static.flickr.com/1260/1171835222_0eec3ad516.jpg\\r\\nhttp://farm3.static.flickr.com/2115/2083805821_30e947678c.jpg\\r\\nhttp://farm1.static.flickr.com/110/290518648_c8858af690.jpg\\r\\nhttp://farm3.static.flickr.com/2300/2111110064_9e2cc90a32.jpg\\r\\nhttp://farm3.static.flickr.com/2160/2448973700_f4c289f491.jpg\\r\\nhttp://farm1.static.flickr.com/19/117021856_b69d0c622d.jpg\\r\\nhttp://static.flickr.com/59/184794043_e67bf330db.jpg\\r\\nhttp://farm2.static.flickr.com/1104/985526394_d7f4975e6b.jpg\\r\\nhttp://farm1.static.flickr.com/229/475988721_791ae61953.jpg\\r\\nhttp://farm1.static.flickr.com/45/137524224_c5259ec221.jpg\\r\\nhttp://farm2.static.flickr.com/1116/1301220595_a3baad8102.jpg\\r\\nhttp://farm2.static.flickr.com/1180/714159242_123fb3dc68.jpg\\r\\nhttp://static.flickr.com/69/154550843_0e1936e7a6.jpg\\r\\nhttp://farm2.static.flickr.com/1350/695639829_4fb17ed1b9.jpg\\r\\nhttp://www.freefoto.com/images/1351/14/1351_14_54---Bicycle--The-Latin-Quarter--Paris--France_web.jpg\\r\\nhttp://farm4.static.flickr.com/3119/2613781487_76de9563f6.jpg\\r\\nhttp://farm1.static.flickr.com/215/482325367_fd76e5a6af.jpg\\r\\nhttp://www.mobielvlaanderen.be/bereikbaarheidsgids/images/oostd_fiets.jpg\\r\\nhttp://farm3.static.flickr.com/2105/2406485102_b4988ac977.jpg\\r\\nhttp://www.betterphoto.com/uploads/processed/0025/0512040206281bicycle_dsc_5477-b-fn_thumbnai.jpg\\r\\nhttp://static.flickr.com/111/267679383_5252ff74f7.jpg\\r\\nhttp://farm3.static.flickr.com/2109/2213904187_f2edff7c8e.jpg\\r\\nhttp://farm4.static.flickr.com/3120/2578354387_9a9e32074a.jpg\\r\\nhttp://www.dsb.tudelft.nl/cvb/temp/fiets_groot.jpg\\r\\nhttp://farm1.static.flickr.com/167/385413064_6613200cd7.jpg\\r\\nhttp://biketrailerblog.com/wp-content/uploads/2008/05/bike-tent-trailer-german-1.jpg\\r\\nhttp://farm4.static.flickr.com/3172/2452456787_28643a042d.jpg\\r\\nhttp://farm4.static.flickr.com/3008/2650190165_12b33be4ca.jpg\\r\\nhttp://farm1.static.flickr.com/232/502922189_2482c84023.jpg\\r\\nhttp://farm3.static.flickr.com/2299/2284035329_2a23cd63fc.jpg\\r\\nhttp://www.nycbikemaps.com/wp-content/uploads/2008/03/bike-rack-1.jpg\\r\\nhttp://farm1.static.flickr.com/31/59118899_3f8f696988.jpg\\r\\nhttp://farm3.static.flickr.com/2216/2281678673_9033ec8e06.jpg\\r\\nhttp://static.flickr.com/18/69792339_13867503c9.jpg\\r\\nhttp://static.flickr.com/1347/561940943_71ea5b4766.jpg\\r\\nhttp://yvet1973.punt.nl/upload/nieuwe_fiets.jpg\\r\\nhttp://media.charleston.net/img/photos/2008/04/23/Bike_Fitting_001_t600.jpg\\r\\nhttp://static.flickr.com/34/70619968_3ff5580a57.jpg\\r\\nhttp://images.asia.ru/img/alibaba/photo/51677918/Electric_Bicycle.jpg\\r\\nhttp://static.flickr.com/103/283546600_08e228fa9f.jpg\\r\\nhttp://farm3.static.flickr.com/2039/2485942590_58e550fa15.jpg\\r\\nhttp://farm3.static.flickr.com/2050/2485120553_6a5b11291a.jpg\\r\\nhttp://farm1.static.flickr.com/10/11986298_d9b3c08edf.jpg\\r\\nhttp://farm1.static.flickr.com/153/412963224_099421f761.jpg\\r\\nhttp://www.smokstak.com/gallery/files/3/2/4/6/Ladyonantiquebicyclecarbideheadlight.jpg\\r\\nhttp://www.joicfp.or.jp/eng/i_campaign/item/mccoba/images/tanza_bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3291/2514020110_30c18fb1b8.jpg\\r\\nhttp://farm1.static.flickr.com/208/520638543_e1f5b86af1.jpg\\r\\nhttp://farm1.static.flickr.com/19/104116062_3b4c28b0c0.jpg\\r\\nhttp://farm1.static.flickr.com/159/360994781_60ca7c9a7b.jpg\\r\\nhttp://farm4.static.flickr.com/3024/2383349930_aab4da9b83.jpg\\r\\nhttp://farm1.static.flickr.com/151/408640571_81064472be.jpg\\r\\nhttp://www.bikefriday.com/sites/default/files/images/bike.jpg\\r\\nhttp://www.bosschers.nl/images/fiets_voor_vertrek.jpg\\r\\nhttp://farm1.static.flickr.com/37/76552423_de8515ca90.jpg\\r\\nhttp://i13.photobucket.com/albums/a276/Spoorzoekers/fiets1.jpg\\r\\nhttp://life-electronic.net/blog/wp-content/uploads/2007/10/bicycle.jpg\\r\\nhttp://www.albertburger.com/bicycle%20in%20the%20snow.jpg\\r\\nhttp://farm3.static.flickr.com/2092/1746760003_bec352b65d.jpg\\r\\nhttp://farm4.static.flickr.com/3014/2551155558_c7e5a02c02.jpg\\r\\nhttp://farm1.static.flickr.com/6/74853335_409a1604f8.jpg\\r\\nhttp://www.nulzesentwintig.nl/uploads/Jeff_fiets_nov_2006.jpg\\r\\nhttp://farm3.static.flickr.com/2116/2055242128_00123f655e.jpg\\r\\nhttp://farm2.static.flickr.com/1066/1321801523_2c90710e50.jpg\\r\\nhttp://farm4.static.flickr.com/3103/2401051082_178503d77b.jpg\\r\\nhttp://static.flickr.com/35/111327947_9931ed7db8.jpg\\r\\nhttp://static.flickr.com/59/165189498_25ec3fa279.jpg\\r\\nhttp://farm3.static.flickr.com/2035/2487680095_d405c82bd8.jpg\\r\\nhttp://gizmodo.com/assets/resources/2006/08/3gspyder.jpg\\r\\nhttp://farm2.static.flickr.com/1337/1165245408_f6d2a19d98.jpg\\r\\nhttp://biketheflathead.files.wordpress.com/2008/03/bike-shogun.jpg\\r\\nhttp://farm2.static.flickr.com/1138/984386662_32c189c440.jpg\\r\\nhttp://farm3.static.flickr.com/2246/2196617564_ef627e3c3d.jpg\\r\\nhttp://www.optimalride.com/wp-content/uploads/2008/05/xtracycle-bike-extender-longtail.jpg\\r\\nhttp://farm4.static.flickr.com/3203/2463772447_d3f45a9ed1.jpg\\r\\nhttp://static.flickr.com/110/297455515_8a3ce932ae.jpg\\r\\nhttp://www.ragoossens.nl/images/oud.JPG\\r\\nhttp://farm3.static.flickr.com/2105/2199620030_a50285e89e.jpg\\r\\nhttp://www.krates.nl/media/1/20050517-pssst...fiets%20kopen.jpg\\r\\nhttp://farm2.static.flickr.com/1142/676403779_36507f6431.jpg\\r\\nhttp://farm1.static.flickr.com/249/451356109_e5caa71f7f.jpg\\r\\nhttp://farm4.static.flickr.com/3221/2671994045_d0e4b3c401.jpg\\r\\nhttp://farm1.static.flickr.com/58/190128821_71c06a2fc4.jpg\\r\\nhttp://farm4.static.flickr.com/3280/2495196809_2eb622791d.jpg\\r\\nhttp://farm3.static.flickr.com/2061/2485117675_1c66187164.jpg\\r\\nhttp://farm3.static.flickr.com/2193/2483407272_57e5f3d649.jpg\\r\\nhttp://www.photographyblog.com/gallery/data/506/6307Their_bicycle_is_better.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/a/a4/Bicycle-taxi.jpg\\r\\nhttp://farm2.static.flickr.com/1331/1063663572_d2071882a7.jpg\\r\\nhttp://farm3.static.flickr.com/2290/2150754245_c2ff2671d4.jpg\\r\\nhttp://farm1.static.flickr.com/1/124988_c2969cce0a.jpg\\r\\nhttp://farm3.static.flickr.com/2091/2532730800_5b7b8a2ce5.jpg\\r\\nhttp://farm3.static.flickr.com/2165/2202919708_43bd1a175d.jpg\\r\\nhttp://farm3.static.flickr.com/2403/2262661509_495ecc9134.jpg\\r\\nhttp://farm3.static.flickr.com/2182/2174873205_3534151676.jpg\\r\\nhttp://farm3.static.flickr.com/2337/2249161750_e218cfa2be.jpg\\r\\nhttp://farm1.static.flickr.com/56/138913844_a955a381c0.jpg\\r\\nhttp://farm1.static.flickr.com/93/249938889_59e2e3f4d8.jpg\\r\\nhttp://www.sampabikers.com.br/images/clubebikechoper/bike04.jpg\\r\\nhttp://www.daveches.co.uk/analysis/bikes-images/whitebike.gif\\r\\nhttp://farm3.static.flickr.com/2364/2442708952_34c31a64b3.jpg\\r\\nhttp://farm1.static.flickr.com/118/294959414_6c7dc62bc3.jpg\\r\\nhttp://images.asia.ru/img/alibaba/photo/51690971/Electric_Bicycle.jpg\\r\\nhttp://static.flickr.com/58/175868675_5899d57e3d.jpg\\r\\nhttp://www.storeappeal.com/images/Micargi%20SM100%20Mountain%20Bike%20Full%20Suspension%2020.jpg\\r\\nhttp://farm1.static.flickr.com/79/279815247_36fb281b0e.jpg\\r\\nhttp://farm3.static.flickr.com/2083/1561499481_88275e0414.jpg\\r\\nhttp://img5.travelblog.org/Photos/53923/228788/t/1787571-Op-de-fiets-0.jpg\\r\\nhttp://farm3.static.flickr.com/2316/2188111887_6bd7b95122.jpg\\r\\nhttp://farm3.static.flickr.com/2401/2195141790_5f1ea8ce02.jpg\\r\\nhttp://farm3.static.flickr.com/2083/2092307328_23c86e141f.jpg\\r\\nhttp://farm3.static.flickr.com/2064/2537746136_9fe54fc1cc.jpg\\r\\nhttp://farm2.static.flickr.com/1340/1120054759_03ba160de3.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/0/04/-_Bee_swarm_on_a_bicycle_(2-5)_-.jpg/800px--_Bee_swarm_on_a_bicycle_(2-5)_-.jpg\\r\\nhttp://www.scalaeditions.com/images/AL_F_Girl%20on%20Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/199/452635202_4da4c0f5ca.jpg\\r\\nhttp://farm1.static.flickr.com/36/112594641_7f1b9a4c5a.jpg\\r\\nhttp://farm4.static.flickr.com/3154/2631936667_f6bb90013c.jpg\\r\\nhttp://farm3.static.flickr.com/2137/2202716818_86bb36d6be.jpg\\r\\nhttp://www.antonmol.nl/pictures/almere2007150-fiets.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2560641233_23047a7b3a.jpg\\r\\nhttp://farm4.static.flickr.com/3021/2479824977_5fbaebeeb9.jpg\\r\\nhttp://farm3.static.flickr.com/2251/2519605735_0c591ec553.jpg\\r\\nhttp://static.flickr.com/29/53679500_5b2f172845.jpg\\r\\nhttp://static.flickr.com/160/430030382_dad5bbd2c9.jpg\\r\\nhttp://farm1.static.flickr.com/221/502165854_97121737f5.jpg\\r\\nhttp://farm1.static.flickr.com/15/18762664_814586739a.jpg\\r\\nhttp://www.trashzen.com/photos/toxsinthumb.jpg\\r\\nhttp://farm2.static.flickr.com/1417/634307228_1646f79e75.jpg\\r\\nhttp://farm3.static.flickr.com/2087/2214516476_518c227a96.jpg\\r\\nhttp://farm4.static.flickr.com/3001/2614644845_5d5e97a78b.jpg\\r\\nhttp://farm3.static.flickr.com/2010/1994872812_f6109b83fc.jpg\\r\\nhttp://www.seascape.ns.ca/~ebc/merida-500_lg.jpg\\r\\nhttp://static.flickr.com/33/61461091_effda8d78c.jpg\\r\\nhttp://farm3.static.flickr.com/2188/2246170723_3310230b5f.jpg\\r\\nhttp://blog.estadao.com.br/blog/media/bike%20este.jpg\\r\\nhttp://farm1.static.flickr.com/8/9253288_148b874732.jpg\\r\\nhttp://farm1.static.flickr.com/175/476662134_d6d5d44f77.jpg\\r\\nhttp://farm4.static.flickr.com/3250/2695243921_a3fd020d42.jpg\\r\\nhttp://janhoogvliet.punt.nl/upload/fiets_1.jpg\\r\\nhttp://farm2.static.flickr.com/1122/725760259_80dac8ef60.jpg\\r\\nhttp://farm4.static.flickr.com/3208/2485117967_f81d4608a8.jpg\\r\\nhttp://farm1.static.flickr.com/21/25717909_5a046bfebb.jpg\\r\\nhttp://farm3.static.flickr.com/2147/1542493976_9bef3659d2.jpg\\r\\nhttp://www.travelcreek.com/bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/116/260668108_081d3baf24.jpg\\r\\nhttp://www.camelbackcoaching.com/images/bike-after.jpg\\r\\nhttp://farm3.static.flickr.com/2148/2031225270_0b6b0b59db.jpg\\r\\nhttp://static.flickr.com/25/54620646_0827549b98.jpg\\r\\nhttp://static.flickr.com/145/364464476_b65f4e7144.jpg\\r\\nhttp://farm4.static.flickr.com/3156/2381107428_b5dbe5bb17.jpg\\r\\nhttp://farm4.static.flickr.com/3039/2396293753_6680a15075.jpg\\r\\nhttp://farm4.static.flickr.com/3195/2518492335_3ccc655cbc.jpg\\r\\nhttp://farm1.static.flickr.com/175/411804031_d7738730f8.jpg\\r\\nhttp://img.alibaba.com/photo/100805181/2008_Gtr_Expert_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/47/142340864_9572d4a2a7.jpg\\r\\nhttp://farm4.static.flickr.com/3046/2397606007_84d47d4bab.jpg\\r\\nhttp://farm3.static.flickr.com/2417/2214413138_17c0117cc7.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/d/d4/Mountain_bike_ParcoSibillini.jpg/800px-Mountain_bike_ParcoSibillini.jpg\\r\\nhttp://static.flickr.com/1004/533410224_502877e756.jpg\\r\\nhttp://farm2.static.flickr.com/1023/1433160275_8a2febefce.jpg\\r\\nhttp://farm1.static.flickr.com/149/426504523_6fdf45c366.jpg\\r\\nhttp://farm3.static.flickr.com/2284/2503995363_87b443daa3.jpg\\r\\nhttp://static.flickr.com/217/493983170_9e3c4aa62d.jpg\\r\\nhttp://www.besportier.com/archives/dragon-bike-1.JPG\\r\\nhttp://farm4.static.flickr.com/3136/2572455190_533e32f6d5.jpg\\r\\nhttp://farm1.static.flickr.com/81/238906246_5850a3f761.jpg\\r\\nhttp://www.deondernemer.nl/data/hermes/BS/2008/02/.cache-467x307/DO_BS_PAAL2_3314308-b02el-fiets.JPG\\r\\nhttp://static.flickr.com/3/2508794_3888b55f81.jpg\\r\\nhttp://farm1.static.flickr.com/28/51455132_a5f72f3dc5.jpg\\r\\nhttp://fietsen.vindtplek.nl/Mountainbikes%2026inch/fiets_procicle_computer.jpg\\r\\nhttp://farm1.static.flickr.com/126/409434418_4f9ac11e49.jpg\\r\\nhttp://farm3.static.flickr.com/2157/2468288685_5011828931.jpg\\r\\nhttp://farm1.static.flickr.com/153/434393341_e90fe906bc.jpg\\r\\nhttp://www.deproatstal.nl/foto's%20website/Fiets.JPG\\r\\nhttp://farm1.static.flickr.com/120/305573114_65f62d3ba4.jpg\\r\\nhttp://farm1.static.flickr.com/10/16182860_98d4f11be1.jpg\\r\\nhttp://farm3.static.flickr.com/2375/2248422877_0e66dacbd6.jpg\\r\\nhttp://farm1.static.flickr.com/41/102055252_a4d79e357d.jpg\\r\\nhttp://static.flickr.com/35/70451015_5a3b689239.jpg\\r\\nhttp://farm1.static.flickr.com/68/193687006_031e9d9831.jpg\\r\\nhttp://farm3.static.flickr.com/2002/2688544898_18ff0d8f7a.jpg\\r\\nhttp://www.cyclesuperstore.ie/shop/pc/images/look/look595bike.jpg\\r\\nhttp://farm4.static.flickr.com/3280/2486642420_eafdd2c473.jpg\\r\\nhttp://farm1.static.flickr.com/159/397624872_086f434de4.jpg\\r\\nhttp://www.sonic.net/~ckelly/Seekay/ck_with_bike.jpg\\r\\nhttp://static.flickr.com/184/369246745_53983490ae.jpg\\r\\nhttp://farm3.static.flickr.com/2323/1584486240_fdf63f80d9.jpg\\r\\nhttp://farm1.static.flickr.com/107/291247893_72d2c658f5.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/2/2e/-_Bee_swarm_on_a_bicycle_(1-5)_-.jpg/800px--_Bee_swarm_on_a_bicycle_(1-5)_-.jpg\\r\\nhttp://farm2.static.flickr.com/1331/696449936_d7080c942e.jpg\\r\\nhttp://static.flickr.com/84/273854145_3744a8f780.jpg\\r\\nhttp://farm3.static.flickr.com/2014/2347615373_c7de77c9e5.jpg\\r\\nhttp://www.stichtingloredana.nl/viorel%20heeft%20fiets.JPG\\r\\nhttp://farm3.static.flickr.com/2353/2505175772_fbcc5bdcdd.jpg\\r\\nhttp://www.gedemo.nl/fiets.jpg\\r\\nhttp://www.besportier.com/archives/mission-bicycle.jpg\\r\\nhttp://static.flickr.com/2010/2066159448_289615eafa.jpg\\r\\nhttp://www.northyorkshire.police.uk/aladdin/stolen_bike_2.jpg\\r\\nhttp://www.sierravistacarclub.org/spring/spring04/ortega.jpg\\r\\nhttp://farm3.static.flickr.com/2022/2256244974_b6937863a9.jpg\\r\\nhttp://farm3.static.flickr.com/2225/2391799705_6a023b59e2.jpg\\r\\nhttp://www.untitledname.com/archives/upload/2005/5/bicycle-film-festival-girl-posse.jpg\\r\\nhttp://farm3.static.flickr.com/2128/2115846015_c01f68bbaa.jpg\\r\\nhttp://farm3.static.flickr.com/2152/2285500962_b05aecc49f.jpg\\r\\nhttp://farm1.static.flickr.com/132/376757157_4099a5821e.jpg\\r\\nhttp://farm1.static.flickr.com/185/480233506_82ceaa973c.jpg\\r\\nhttp://farm1.static.flickr.com/146/434442565_fc995057e7.jpg\\r\\nhttp://farm1.static.flickr.com/215/508936165_e2492df740.jpg\\r\\nhttp://static.flickr.com/221/464639780_04810e507b.jpg\\r\\nhttp://farm1.static.flickr.com/61/199994239_0dec63560f.jpg\\r\\nhttp://farm1.static.flickr.com/33/91575201_fa6b20bd00.jpg\\r\\nhttp://static.flickr.com/1413/538444267_3c0a65fa49.jpg\\r\\nhttp://farm4.static.flickr.com/3001/2283964310_99e838a86e.jpg\\r\\nhttp://www.riceroselaw.com/bicycling_200.jpg\\r\\nhttp://farm2.static.flickr.com/1177/815046969_25d378bda0.jpg\\r\\nhttp://farm2.static.flickr.com/1212/528055617_62f0d4418f.jpg\\r\\nhttp://static.tmg.nl/wuz/c/7/7/1210862157177.jpg\\r\\nhttp://farm2.static.flickr.com/1025/616930211_15190b0470.jpg\\r\\nhttp://farm1.static.flickr.com/125/338280952_72a4d434c7.jpg\\r\\nhttp://farm3.static.flickr.com/2353/2114098775_3418976ee6.jpg\\r\\nhttp://farm2.static.flickr.com/1131/528080677_75fbe1de0a.jpg\\r\\nhttp://static.flickr.com/47/166471686_652d09ffbd.jpg\\r\\nhttp://farm1.static.flickr.com/211/454496164_2d8cb109b5.jpg\\r\\nhttp://farm2.static.flickr.com/1094/864148476_64b356734e.jpg\\r\\nhttp://farm1.static.flickr.com/32/57470443_881eab0e05.jpg\\r\\nhttp://farm4.static.flickr.com/3099/2626520850_0d0064eb1c.jpg\\r\\nhttp://farm1.static.flickr.com/65/183589524_d18512bcdc.jpg\\r\\nhttp://farm2.static.flickr.com/1411/1448609011_bdbe6113d8.jpg\\r\\nhttp://www.dropmachine.com/media/data/500/bicycle_004.jpg\\r\\nhttp://farm1.static.flickr.com/71/217197184_55cdeb6535.jpg\\r\\nhttp://farm3.static.flickr.com/2290/1798540041_e49ef7daba.jpg\\r\\nhttp://static.flickr.com/64/210528313_810edfa467.jpg\\r\\nhttp://static.flickr.com/2174/2119229715_dd594d0246.jpg\\r\\nhttp://static.flickr.com/2327/2067379744_8442a0dd44.jpg\\r\\nhttp://farm2.static.flickr.com/1401/695707533_0dd12fe418.jpg\\r\\nhttp://farm4.static.flickr.com/3030/2716492139_dd9134df04.jpg\\r\\nhttp://farm4.static.flickr.com/3256/2548179213_5602b847ca.jpg\\r\\nhttp://farm4.static.flickr.com/3131/2590767203_5ca46200fa.jpg\\r\\nhttp://farm2.static.flickr.com/1234/1345822027_01d4c78c48.jpg\\r\\nhttp://www.harmteunissen.nl/Gennep/webspace/Gennepse%20mensen/1947%20Gennep%20bij%20H%20Hartbeeld-ingslagen%20kogels-%20Hans%20van%20Kempen%20bij%20zijn%20vader%20Harrie%20van%20Kempen%20op%20de%20fiets.jpg\\r\\nhttp://farm1.static.flickr.com/227/466107244_1d976d9ffd.jpg\\r\\nhttp://static.flickr.com/76/160157539_0328cb1ef5.jpg\\r\\nhttp://farm3.static.flickr.com/2278/1562682422_3cd475714b.jpg\\r\\nhttp://farm1.static.flickr.com/191/512360375_c82e2d683c.jpg\\r\\nhttp://www.lokeshdhakar.com/gallery/2006/raleigh-bike-profile.jpg\\r\\nhttp://static.flickr.com/23/34268530_03b40fcaf0.jpg\\r\\nhttp://farm4.static.flickr.com/3027/2487594319_433c27492e.jpg\\r\\nhttp://farm4.static.flickr.com/3239/2284038063_d51993a2d7.jpg\\r\\nhttp://farm4.static.flickr.com/3113/2714105994_db91c9dc67.jpg\\r\\nhttp://farm1.static.flickr.com/74/164249093_eae4f0d47c.jpg\\r\\nhttp://www.hellobikes.com/graficos/img_articulos/Ciber%20Bike%20(457%20x%20419).jpg\\r\\nhttp://static.flickr.com/1028/1233406410_ba5779f2d9.jpg\\r\\nhttp://farm2.static.flickr.com/1332/1128329190_8e0b076faa.jpg\\r\\nhttp://farm4.static.flickr.com/3228/2616229147_f599d7a631.jpg\\r\\nhttp://farm1.static.flickr.com/56/193307828_3d879dd99b.jpg\\r\\nhttp://farm1.static.flickr.com/10/15620839_78db3b428f.jpg\\r\\nhttp://farm2.static.flickr.com/1438/1037210243_318922a9f6.jpg\\r\\nhttp://farm1.static.flickr.com/162/411802569_234038fcf7.jpg\\r\\nhttp://farm4.static.flickr.com/3088/2672001381_ca9728f630.jpg\\r\\nhttp://farm1.static.flickr.com/67/230218581_fc805aa4cf.jpg\\r\\nhttp://farm1.static.flickr.com/83/272461292_8a24728b1e.jpg\\r\\nhttp://farm1.static.flickr.com/231/501791705_db4c19d58d.jpg\\r\\nhttp://farm1.static.flickr.com/53/113837123_2af66d651e.jpg\\r\\nhttp://www.juliadean.com/cubagallery/images/bicycle.jpg\\r\\nhttp://www.bicyclefrenzy.com/wp-content/uploads/2007/11/bicyclecrash.jpg\\r\\nhttp://farm1.static.flickr.com/183/443046093_dbc33748ee.jpg\\r\\nhttp://farm1.static.flickr.com/139/377095800_c916a45705.jpg\\r\\nhttp://cycling.finial.com/blog/wp-content/bike3.jpg\\r\\nhttp://farm2.static.flickr.com/1151/1368811629_b1b21850ef.jpg\\r\\nhttp://farm3.static.flickr.com/2036/2010720319_3fe957cdcd.jpg\\r\\nhttp://farm1.static.flickr.com/122/270823956_db984112da.jpg\\r\\nhttp://farm1.static.flickr.com/21/30452792_bacbe9fae9.jpg\\r\\nhttp://farm1.static.flickr.com/111/306828694_f1688e2f0e.jpg\\r\\nhttp://farm4.static.flickr.com/3075/2568782223_77528cf574.jpg\\r\\nhttp://farm4.static.flickr.com/3186/2606948891_c31427b52e.jpg\\r\\nhttp://farm2.static.flickr.com/1186/538776293_9fec10371f.jpg\\r\\nhttp://farm2.static.flickr.com/1386/1124587435_fde37d1ed7.jpg\\r\\nhttp://farm4.static.flickr.com/3102/2341462808_4b3c21b699.jpg\\r\\nhttp://farm4.static.flickr.com/3222/2729545180_c2db947fb2.jpg\\r\\nhttp://farm4.static.flickr.com/3235/2654089360_ceb915128b.jpg\\r\\nhttp://farm1.static.flickr.com/163/370357484_ebf8ac65ea.jpg\\r\\nhttp://www.sfbike.org/images/resources/triathlon/bikeracergirl.jpg\\r\\nhttp://farm1.static.flickr.com/187/447828233_3b345fcdc2.jpg\\r\\nhttp://images.asia.ru/img/alibaba/photo/51485222/Electric_Bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3011/2571390031_b31a5dc550.jpg\\r\\nhttp://bp2.blogger.com/_Mkd9sz1mhP0/RgYS1RVRQtI/AAAAAAAAAm0/jDll8_yXnIo/s400/steam-bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3183/2485945146_59be950a96.jpg\\r\\nhttp://img2.travelblog.org/Photos/8230/105701/f/706185-Bike-sculpture-San-Gabriel-Ecuador-1.jpg\\r\\nhttp://merksem.net/sport/images/atletiek/atleten/aernouts_bart/fiets_rimini.jpg\\r\\nhttp://farm3.static.flickr.com/2010/2221700387_ed5606684c.jpg\\r\\nhttp://farm1.static.flickr.com/164/406090304_33c151ff65.jpg\\r\\nhttp://farm2.static.flickr.com/1305/700133564_15e361f0c1.jpg\\r\\nhttp://static.flickr.com/232/505600784_9fd367116f.jpg\\r\\nhttp://imagecache2.allposters.com/images/pic/PA2/PPR45022~Albert-Einstein-Riding-Bicycle-Posters.jpg\\r\\nhttp://farm3.static.flickr.com/2274/2135709585_e30396e171.jpg\\r\\nhttp://english.vietnamnet.vn/dataimages/200709/original/images1409239_bicycle.jpg\\r\\nhttp://farm3.static.flickr.com/2004/2536929753_ffe509f075.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/3/33/Half_Wheeler_-_bike.JPG/800px-Half_Wheeler_-_bike.JPG\\r\\nhttp://lowrider.free.fr/eastriddaz/Ricco%27s_Bike01.jpg\\r\\nhttp://farm1.static.flickr.com/38/74133023_6a81315c60.jpg\\r\\nhttp://farm3.static.flickr.com/2163/2268556562_663b76fb09.jpg\\r\\nhttp://www.optimalride.com/wp-content/uploads/2008/05/xtracycle-long-bicycle.jpg\\r\\nhttp://www.rkphoto.com/images/HalfBikes/Our%20bike%20room%203.jpg\\r\\nhttp://farm1.static.flickr.com/25/42973167_5da51f1f40.jpg\\r\\nhttp://farm4.static.flickr.com/3079/2553773859_5ca03a4d4e.jpg\\r\\nhttp://images.asia.ru/img/alibaba/photo/50583654/Electric_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/6/10696241_0e692dc0cf.jpg\\r\\nhttp://farm1.static.flickr.com/37/125261393_668bad7a50.jpg\\r\\nhttp://farm1.static.flickr.com/132/331934393_c0764cc6b3.jpg\\r\\nhttp://static.flickr.com/23/26965413_4d5664b46e.jpg\\r\\nhttp://static.flickr.com/4/5252666_f970bb42c8.jpg\\r\\nhttp://farm1.static.flickr.com/115/306054306_20ed07ddaf.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/0/07/Small_bicycle.JPG/800px-Small_bicycle.JPG\\r\\nhttp://farm1.static.flickr.com/55/131572172_65e0a16727.jpg\\r\\nhttp://farm4.static.flickr.com/3161/2579637351_ca37c554b6.jpg\\r\\nhttp://www.tenerife-training.net/Images/convertable-hybrid-MTB-bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/67/201918981_ecd8c24d47.jpg\\r\\nhttp://fotos.marktplaats.nl/kopen/3/1e/uo+2dP66HDfWff+6uw9MWQ==.jpg\\r\\nhttp://static.flickr.com/217/455588533_b1ef638110.jpg\\r\\nhttp://farm1.static.flickr.com/35/111170812_ee72e35a61.jpg\\r\\nhttp://www.veilingkranten.nl/img/gd/producten/364/363_xl.jpg\\r\\nhttp://farm4.static.flickr.com/3004/2546241902_540fa223b9.jpg\\r\\nhttp://www.crowdspirit.com/blog/images/bike_1.jpg\\r\\nhttp://farm1.static.flickr.com/164/387630195_db6ec4a0d6.jpg\\r\\nhttp://farm4.static.flickr.com/3015/2688979466_d984b669a6.jpg\\r\\nhttp://fietsen.web-log.nl/photos/uncategorized/drentsefietsvierdaagsespandoek_2.jpg\\r\\nhttp://www.ravellobikes.com/ultimate%20travel%20bike%20photo1.JPG\\r\\nhttp://farm1.static.flickr.com/94/229153319_adeb77d44f.jpg\\r\\nhttp://img.alibaba.com/photo/100919202/Scott_Speedster_S10_Road_Racing_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/54/134460305_dc5897c4f8.jpg\\r\\nhttp://farm4.static.flickr.com/3101/2583216129_c41ed17f7a.jpg\\r\\nhttp://farm1.static.flickr.com/167/411784100_9f011df322.jpg\\r\\nhttp://farm2.static.flickr.com/1185/695637437_3733525bcf.jpg\\r\\nhttp://www.cyclingcountry.com/Images/TrekBicycle.gif\\r\\nhttp://farm1.static.flickr.com/116/266322527_dfae0056ca.jpg\\r\\nhttp://farm3.static.flickr.com/2410/2425671762_b6e7fbafd5.jpg\\r\\nhttp://farm2.static.flickr.com/1317/752197807_59cc188166.jpg\\r\\nhttp://farm4.static.flickr.com/3207/2467615971_9fc33f34a4.jpg\\r\\nhttp://farm1.static.flickr.com/128/389563505_157bf59507.jpg\\r\\nhttp://farm2.static.flickr.com/1171/1266713201_60f8525bd7.jpg\\r\\nhttp://www.gadgetizer.com/wp-content/uploads/2007/05/electric-bicycle-piano-classic1.jpg\\r\\nhttp://farm1.static.flickr.com/191/497868903_5d9c20967e.jpg\\r\\nhttp://bp0.blogger.com/_ERGrznxTIwo/RhUlmhvrHlI/AAAAAAAABT0/ar1I5gyxffQ/s400/olympics+bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3151/2580588847_835e97eec8.jpg\\r\\nhttp://farm1.static.flickr.com/12/96600681_98e6b48c81.jpg\\r\\nhttp://farm1.static.flickr.com/92/235062420_42a23f74a9.jpg\\r\\nhttp://farm1.static.flickr.com/214/512491470_c6085722ff.jpg\\r\\nhttp://img.alibaba.com/photo/101149678/2008_Santa_Cruz_V10_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/178/407912267_869180c96d.jpg\\r\\nhttp://static.flickr.com/113/309172582_05800b4025.jpg\\r\\nhttp://farm4.static.flickr.com/3088/2620747691_c497071d2f.jpg\\r\\nhttp://static.flickr.com/1159/527992714_2026a97527.jpg\\r\\nhttp://farm1.static.flickr.com/54/128954275_9a2f4c78dd.jpg\\r\\nhttp://farm1.static.flickr.com/53/140736535_c55f7f0bb8.jpg\\r\\nhttp://www.alphenaandenrijn.nl/INTERNET/Communicatie/Afbeeldingen/IMG_2950.jpg\\r\\nhttp://farm1.static.flickr.com/118/272976939_d335524531.jpg\\r\\nhttp://www.johnclearygallery.com/pics/smith/womanonbicycle.jpg\\r\\nhttp://farm1.static.flickr.com/33/35438322_babc31f188.jpg\\r\\nhttp://farm1.static.flickr.com/155/356779934_f93497c958.jpg\\r\\nhttp://www.njskylands.com/Imageod/Diablo.jpg\\r\\nhttp://www.bmxultra.com/news/2006/intense/intense_2006_team_bike.jpg\\r\\nhttp://farm1.static.flickr.com/129/373037070_89f4b1ed25.jpg\\r\\nhttp://farm4.static.flickr.com/3196/2407377820_9c427e916e.jpg\\r\\nhttp://farm3.static.flickr.com/2406/2298849557_2bfc436e64.jpg\\r\\nhttp://farm1.static.flickr.com/83/228554066_6bb26b299c.jpg\\r\\nhttp://farm3.static.flickr.com/2068/1711095552_be81309744.jpg\\r\\nhttp://farm4.static.flickr.com/3248/2410435169_b3cf1000d5.jpg\\r\\nhttp://farm1.static.flickr.com/62/170789927_5f53d080ea.jpg\\r\\nhttp://farm1.static.flickr.com/26/56890723_4aa44f7f7e.jpg\\r\\nhttp://img.2dehands.nl/f/normal/23108468_2-rih-delta-fiets.jpg\\r\\nhttp://farm1.static.flickr.com/183/481529432_61ede07ca0.jpg\\r\\nhttp://static.flickr.com/194/522676783_53cdcf79de.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/e/e2/Bicycle_1865.jpg\\r\\nhttp://farm3.static.flickr.com/2344/2359220539_62883226d1.jpg\\r\\nhttp://farm4.static.flickr.com/3178/2296432588_e6d76299c3.jpg\\r\\nhttp://farm1.static.flickr.com/217/466004075_3e8588de58.jpg\\r\\nhttp://static.flickr.com/2110/2085754555_3b0f2cb08b.jpg\\r\\nhttp://markharja.web-log.nl/photos/uncategorized/fiets_safari.jpg\\r\\nhttp://blog.makezine.com/bambooBike2.jpg\\r\\nhttp://www.tomsimpson.org/images/chopper_bike.jpg\\r\\nhttp://farm4.static.flickr.com/3246/2652851177_ddff76cf10.jpg\\r\\nhttp://static.flickr.com/145/431536407_af6f7ae078.jpg\\r\\nhttp://farm1.static.flickr.com/186/442476925_609903fddb.jpg\\r\\nhttp://farm4.static.flickr.com/3224/2671999327_67aaaac9f0.jpg\\r\\nhttp://farm4.static.flickr.com/3266/2485933254_576080a8db.jpg\\r\\nhttp://farm1.static.flickr.com/13/18929547_83a6206e32.jpg\\r\\nhttp://www.jimlangley.net/spin/stevens.gif\\r\\nhttp://farm1.static.flickr.com/29/49206696_02f00a1e2a.jpg\\r\\nhttp://farm4.static.flickr.com/3051/2729629318_62c49004b1.jpg\\r\\nhttp://farm4.static.flickr.com/3069/2703854782_eef0cbf953.jpg\\r\\nhttp://www.weeu.com/Mens%20Pro%20Race%201%20-06.JPG\\r\\nhttp://farm3.static.flickr.com/2355/2459056091_7598a1ebf1.jpg\\r\\nhttp://static.flickr.com/29/46120703_c4526dacde.jpg\\r\\nhttp://farm4.static.flickr.com/3143/2632530304_c25ed7ee5b.jpg\\r\\nhttp://www.unm.edu/~archie/gage3yr/Bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3144/2578398759_804511d516.jpg\\r\\nhttp://farm1.static.flickr.com/72/171678555_c8c746f984.jpg\\r\\nhttp://farm1.static.flickr.com/54/141404651_ddcef85059.jpg\\r\\nhttp://www.johnhenrybikes.com/images/merrittBikePark1.jpg\\r\\nhttp://asymptotia.com/wp-images/2006/09/blog_on_a_bike_b.jpg\\r\\nhttp://www.mccullagh.org/db9/10d-7/montreal-bicycle.jpg\\r\\nhttp://farm3.static.flickr.com/2088/2167270703_311c593f31.jpg\\r\\nhttp://farm1.static.flickr.com/176/429847696_22fea4d27a.jpg\\r\\nhttp://farm3.static.flickr.com/2368/2092678702_5545c4cea7.jpg\\r\\nhttp://farm1.static.flickr.com/205/454777119_6ad379581b.jpg\\r\\nhttp://static.flickr.com/1215/1435045040_8c4355bf6a.jpg\\r\\nhttp://farm3.static.flickr.com/2169/2124923368_0e243f0cec.jpg\\r\\nhttp://farm1.static.flickr.com/188/364095600_b52d48b170.jpg\\r\\nhttp://farm3.static.flickr.com/2362/1572209012_5188fdb4a0.jpg\\r\\nhttp://static.flickr.com/1377/1472906593_28895f6b18.jpg\\r\\nhttp://farm3.static.flickr.com/2306/2370196922_a7ce42c72a.jpg\\r\\nhttp://static.flickr.com/53/164216337_ea23a9257d.jpg\\r\\nhttp://farm1.static.flickr.com/172/411793404_10c35850ea.jpg\\r\\nhttp://www.novaksblog.com/pictures/interesting_photos/bike_jump.jpg\\r\\nhttp://farm1.static.flickr.com/38/84715915_ac2e0134a0.jpg\\r\\nhttp://farm1.static.flickr.com/135/321544476_ea270f186c.jpg\\r\\nhttp://static.flickr.com/1428/1398584347_ed32744c70.jpg\\r\\nhttp://farm4.static.flickr.com/3018/2459242666_3bc78bcebb.jpg\\r\\nhttp://farm3.static.flickr.com/2307/1588034079_a7c65fe9ee.jpg\\r\\nhttp://static.flickr.com/108/287769684_d8ae73e674.jpg\\r\\nhttp://farm3.static.flickr.com/2104/2462533558_fb21aaf8ec.jpg\\r\\nhttp://farm4.static.flickr.com/3108/2366255410_9f7576ec0d.jpg\\r\\nhttp://farm3.static.flickr.com/2251/2429016582_c1d9e76458.jpg\\r\\nhttp://static.flickr.com/21/29179084_ad8ed3fb7d.jpg\\r\\nhttp://farm1.static.flickr.com/142/381290422_1f270487cf.jpg\\r\\nhttp://www.destination360.com/europe/italy/images/s/italy-bicycle-tours.jpg\\r\\nhttp://farm1.static.flickr.com/125/318323418_83148937cd.jpg\\r\\nhttp://farm3.static.flickr.com/2279/2233300976_2700073dd1.jpg\\r\\nhttp://www.surf-blog.net/images/bike/trial/trial-bike-2.jpg\\r\\nhttp://farm1.static.flickr.com/7/10829377_44215335bc.jpg\\r\\nhttp://images.ebizautos.com/sites/5282/images/bmw_bicycle_giveaway.jpg\\r\\nhttp://farm3.static.flickr.com/2368/2388654129_5e739eb1fd.jpg\\r\\nhttp://farm1.static.flickr.com/39/78805062_89a5ee127e.jpg\\r\\nhttp://farm3.static.flickr.com/2396/1573055902_e350062e99.jpg\\r\\nhttp://farm2.static.flickr.com/1293/954731854_6c6a423695.jpg\\r\\nhttp://farm1.static.flickr.com/6/11543654_755368c930.jpg\\r\\nhttp://farm1.static.flickr.com/22/27182672_fe89244471.jpg\\r\\nhttp://farm4.static.flickr.com/3116/2564812185_42053de7c6.jpg\\r\\nhttp://web.cs.wpi.edu/~claypool/india/bicycle.gif\\r\\nhttp://farm1.static.flickr.com/35/73051880_1948d613ae.jpg\\r\\nhttp://xprize.wirefly.com/images/rocket-bike.jpg\\r\\nhttp://farm1.static.flickr.com/54/425308028_ce17167bb7.jpg\\r\\nhttp://farm1.static.flickr.com/217/492076204_f3978ab18d.jpg\\r\\nhttp://ftlauderdalevilla.com/images/beach-bicycle.jpg\\r\\nhttp://www.jonco48.com/blog/bike.jpg\\r\\nhttp://farm1.static.flickr.com/52/140898286_c4423857bb.jpg\\r\\nhttp://farm4.static.flickr.com/3196/2393245787_2e47373a5a.jpg\\r\\nhttp://farm3.static.flickr.com/2098/2042131399_e2ce7fc7d8.jpg\\r\\nhttp://static.flickr.com/162/345270380_ef4e602991.jpg\\r\\nhttp://www.funsportmountainbike.com/header.jpg\\r\\nhttp://www.nicmic.com/pix/2007/bo_fiets_samuel_schiermonnikoog.jpg\\r\\nhttp://farm1.static.flickr.com/78/206770543_7c809f738f.jpg\\r\\nhttp://farm1.static.flickr.com/229/497539948_e38394d7b6.jpg\\r\\nhttp://www.washington.edu/commuterservices/images/PHOTOS/BIKE_LOCKER/BIKE_LOCK_DSC_0243_n.jpg\\r\\nhttp://farm3.static.flickr.com/2357/1855703632_1d7f4ce6a2.jpg\\r\\nhttp://farm3.static.flickr.com/2362/2092679034_3ec40d0d42.jpg\\r\\nhttp://farm2.static.flickr.com/1359/892521457_aec6e2a391.jpg\\r\\nhttp://farm1.static.flickr.com/211/508749646_c6f29dc8c7.jpg\\r\\nhttp://farm1.static.flickr.com/196/461379052_34ab1bd0d8.jpg\\r\\nhttp://farm1.static.flickr.com/222/520474948_2c0f891eb7.jpg\\r\\nhttp://community.adn.com/sites/community.adn.com/files/images/OddBike.preview.jpg\\r\\nhttp://farm4.static.flickr.com/3204/2673793639_7061dbe427.jpg\\r\\nhttp://prometheus.med.utah.edu/~bwjones/C1516441744/E20060724141536/Media/Honda%20mountain%20bike.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/f/fc/Bicycle_trial.jpg\\r\\nhttp://static.flickr.com/162/439728714_131c2ed76c.jpg\\r\\nhttp://static.flickr.com/215/488695869_8b2c4e9774.jpg\\r\\nhttp://static.flickr.com/2089/1802032571_3f3b46ff26.jpg\\r\\nhttp://www.sonoma.edu/pubs/mt/images/bigroller.jpg\\r\\nhttp://imagesme.net/tuvie/oryx-bike3.jpg\\r\\nhttp://static.flickr.com/12/14647611_20a823d028.jpg\\r\\nhttp://farm3.static.flickr.com/2041/2278125983_1e1a81eb91.jpg\\r\\nhttp://farm1.static.flickr.com/105/264364045_1b659ff7a8.jpg\\r\\nhttp://farm4.static.flickr.com/3283/2323903395_cf189ef99a.jpg\\r\\nhttp://farm4.static.flickr.com/3006/2464183033_01670f5c03.jpg\\r\\nhttp://farm1.static.flickr.com/67/213756214_61b6a22e49.jpg\\r\\nhttp://www3.mtb-news.de/fotos/data/2/medium/Mein_Bike_1.JPG\\r\\nhttp://urbanscooters.com/P/images/bikemania/ElectricBike_ChopperKit-2.jpg\\r\\nhttp://farm3.static.flickr.com/2268/2121323711_54e9e33c72.jpg\\r\\nhttp://farm2.static.flickr.com/1412/1022947780_a015defc15.jpg\\r\\nhttp://farm4.static.flickr.com/3076/2653077685_8362d4cd01.jpg\\r\\nhttp://farm1.static.flickr.com/54/184882491_530f121fe7.jpg\\r\\nhttp://web.inter.nl.net/hcc/jmm/album/fiets.jpg\\r\\nhttp://www.greatoutdoors.com/files/imagecache/display/files/images/articles/colorado-multi-day-bike-tours5.jpg\\r\\nhttp://farm1.static.flickr.com/218/470298091_92a6baef16.jpg\\r\\nhttp://farm1.static.flickr.com/164/437540009_85b1f2ad4d.jpg\\r\\nhttp://farm1.static.flickr.com/179/411038293_b1bfcfe11d.jpg\\r\\nhttp://farm3.static.flickr.com/2207/2363748109_85421a8cc9.jpg\\r\\nhttp://farm1.static.flickr.com/26/61025453_a49ba3c0ea.jpg\\r\\nhttp://farm4.static.flickr.com/3190/2432444058_4de6f45914.jpg\\r\\nhttp://farm3.static.flickr.com/2248/1599155273_9a5d3184f8.jpg\\r\\nhttp://www.polfed-fedpol.be/dos_ops/Dos_VanUytsel/images/kj291107_fiets.JPG\\r\\nhttp://farm3.static.flickr.com/2011/2188668891_dd29875e3c.jpg\\r\\nhttp://farm1.static.flickr.com/82/266807093_c333c62f0a.jpg\\r\\nhttp://farm3.static.flickr.com/2091/2089182946_3a4e654595.jpg\\r\\nhttp://farm1.static.flickr.com/6/10649366_ef9dbfd34a.jpg\\r\\nhttp://www.burgerpanelrotterdam.nl/Media/fiets/IMG_0565.jpg\\r\\nhttp://farm1.static.flickr.com/33/50888390_9de2ee5694.jpg\\r\\nhttp://farm1.static.flickr.com/21/29662982_5d9ff1580b.jpg\\r\\nhttp://farm1.static.flickr.com/61/161673900_df96acf9c1.jpg\\r\\nhttp://farm1.static.flickr.com/34/67246098_1d4c4840fd.jpg\\r\\nhttp://farm1.static.flickr.com/30/102195809_61b9c5ee97.jpg\\r\\nhttp://farm2.static.flickr.com/1276/810187523_de976fb4f9.jpg\\r\\nhttp://www.arikah.com/encyclopedia/images/3/3b/BicyclePlymouth.jpg\\r\\nhttp://www.autismvox.com/wp-content/boydadbike_2.jpg\\r\\nhttp://static.flickr.com/43/83300412_0d06688737.jpg\\r\\nhttp://farm1.static.flickr.com/226/461725635_04cfe34349.jpg\\r\\nhttp://farm2.static.flickr.com/1228/966038640_817cc17a57.jpg\\r\\nhttp://farm3.static.flickr.com/2344/1919769438_9606c54851.jpg\\r\\nhttp://farm1.static.flickr.com/193/493508388_b0b2959c6d.jpg\\r\\nhttp://farm1.static.flickr.com/48/171678548_103134565a.jpg\\r\\nhttp://img.2dehands.nl/f/normal/32546253_1-sparta-ion-city-seat-een-fiets-die-werkelijk-uw-aandacht.jpg\\r\\nhttp://farm2.static.flickr.com/1118/938915724_3f5554075d.jpg\\r\\nhttp://www.roverklubben.se/bilder/cycles/safety.jpg\\r\\nhttp://static.flickr.com/14/89548434_abe892251a.jpg\\r\\nhttp://www.bicielettriche.net/images/mountainbike.jpg\\r\\nhttp://farm3.static.flickr.com/2360/1659842357_d7a916f9bc.jpg\\r\\nhttp://static.flickr.com/1361/749416317_c41f16bd1f.jpg\\r\\nhttp://farm3.static.flickr.com/2331/2128904451_a82e5dde9c.jpg\\r\\nhttp://farm1.static.flickr.com/78/162490991_33e5f767d7.jpg\\r\\nhttp://farm2.static.flickr.com/1073/1049852114_28795f1503.jpg\\r\\nhttp://farm1.static.flickr.com/21/31421517_3a252348cc.jpg\\r\\nhttp://www.garagescapes.com/images/bike_lift.jpg\\r\\nhttp://static.flickr.com/65/391620098_dc1a7f8e36.jpg\\r\\nhttp://farm3.static.flickr.com/2021/1622809213_8c2b8afdf7.jpg\\r\\nhttp://static.flickr.com/1439/639860677_d42a1b9cbb.jpg\\r\\nhttp://farm3.static.flickr.com/2278/2462902297_9cbf5b997e.jpg\\r\\nhttp://www.rhizomes.nl/bergman%20bedrijfsfietsen.jpg\\r\\nhttp://farm1.static.flickr.com/70/172995947_5030868224.jpg\\r\\nhttp://farm4.static.flickr.com/3277/2350203366_9ccd14e70f.jpg\\r\\nhttp://bikes4.web-log.nl/photos/uncategorized/2008/04/12/foto_fiets_4bikes_klein_2.jpg\\r\\nhttp://www.pelmo.it/fotoshow/images/Mountain%20Bike.JPG\\r\\nhttp://chopinensander.web-log.nl/chopinensander/images/fiets.jpg\\r\\nhttp://static.flickr.com/101/317201392_420358b1bc.jpg\\r\\nhttp://static.flickr.com/50/161672919_267c7e6348.jpg\\r\\nhttp://www.bikefriday.com/sites/default/files/images/WEB_TWINAIR_Lynette_&_Mike_Schuyler_Bike_NY_2002.jpg\\r\\nhttp://www.deventercentraal.nl/wp-content/uploads/groene_fiets.jpg\\r\\nhttp://www.copperhood.com/pic-gallery/biking/images/bicycle-trail.jpg\\r\\nhttp://static.flickr.com/1186/537650647_5f0a65fa13.jpg\\r\\nhttp://farm1.static.flickr.com/32/42384285_74f3b47d48.jpg\\r\\nhttp://static.flickr.com/111/293051263_91cf8c9cb2.jpg\\r\\nhttp://farm3.static.flickr.com/2280/2284823704_64e25971b5.jpg\\r\\nhttp://farm1.static.flickr.com/77/203112143_18e0f95152.jpg\\r\\nhttp://farm1.static.flickr.com/133/411803709_eab82f63ef.jpg\\r\\nhttp://farm3.static.flickr.com/2278/2240067470_1e833e96f9.jpg\\r\\nhttp://farm1.static.flickr.com/173/394025587_4794c856b7.jpg\\r\\nhttp://farm3.static.flickr.com/2329/2284822608_61ae406a79.jpg\\r\\nhttp://i4.photobucket.com/albums/y122/reneoskam/fiets.jpg\\r\\nhttp://farm1.static.flickr.com/31/66499018_6c3f81c12b.jpg\\r\\nhttp://www.outdoorvm.com/photos/march07/minnewanka_bike1.jpg\\r\\nhttp://farm1.static.flickr.com/56/148908928_fa079f3e12.jpg\\r\\nhttp://static.flickr.com/4/4686507_9d820b507d.jpg\\r\\nhttp://farm1.static.flickr.com/49/162303010_7119e102b6.jpg\\r\\nhttp://static.flickr.com/1357/1149157136_c72f22d78c.jpg\\r\\nhttp://farm1.static.flickr.com/150/416682079_c5a7f16434.jpg\\r\\nhttp://farm2.static.flickr.com/1225/528160663_2207c60eaa.jpg\\r\\nhttp://www.thebeijingguide.com/francais/bicycle_beijing/bike_beijing_2.jpg\\r\\nhttp://static.flickr.com/167/401187188_bf62fd156d.jpg\\r\\nhttp://farm1.static.flickr.com/9/12690690_7dd4e7de99.jpg\\r\\nhttp://farm2.static.flickr.com/1167/1142581704_f09eb46947.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/0/04/Burning_figure_on_bicycle_-_after.jpg/400px-Burning_figure_on_bicycle_-_after.jpg\\r\\nhttp://farm1.static.flickr.com/31/54133573_7e60766f6c.jpg\\r\\nhttp://leidseglibber.punt.nl/upload/Fiets_fout,fiets_weg_(7)_(Custom).JPG\\r\\nhttp://www.ski-epic.com/amsterdam_bicycles/ps0b_amsterdam_bicycle_cell.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/1/19/Reflective_bicycle.jpg/450px-Reflective_bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/150/411803026_2730b89a8d.jpg\\r\\nhttp://farm4.static.flickr.com/3150/2397663417_6eb503dfd7.jpg\\r\\nhttp://farm3.static.flickr.com/2357/2166176613_4722083ec5.jpg\\r\\nhttp://static.flickr.com/1251/1010157934_bc50b72b9e.jpg\\r\\nhttp://farm1.static.flickr.com/32/57006264_72a7e6c314.jpg\\r\\nhttp://www3.picturepush.com/photo/a/209221/1024/foto/fiets.jpg\\r\\nhttp://farm3.static.flickr.com/2035/2421294585_fae2a644cf.jpg\\r\\nhttp://farm1.static.flickr.com/2/1801270_5300efe2c9.jpg\\r\\nhttp://farm1.static.flickr.com/83/239465953_77030dcdb6.jpg\\r\\nhttp://static.flickr.com/91/234495039_751fc52a99.jpg\\r\\nhttp://farm4.static.flickr.com/3020/2633036197_9b03a6bcbe.jpg\\r\\nhttp://farm1.static.flickr.com/91/207314543_4487eb0a9b.jpg\\r\\nhttp://farm2.static.flickr.com/1098/1066098490_093827fec5.jpg\\r\\nhttp://www.cloverleafhorsefarm.com/pictures/Vicki%20mountain%20bike.JPG\\r\\nhttp://farm3.static.flickr.com/2175/2363471822_6aaf8c3afa.jpg\\r\\nhttp://farm4.static.flickr.com/3072/2695087448_93b0358176.jpg\\r\\nhttp://farm3.static.flickr.com/2026/2232187581_ee1c0e6301.jpg\\r\\nhttp://farm2.static.flickr.com/1260/685286361_93a93d30fe.jpg\\r\\nhttp://farm2.static.flickr.com/1159/1438071015_7b20d33d04.jpg\\r\\nhttp://farm1.static.flickr.com/148/373273160_74a47ec808.jpg\\r\\nhttp://www.bikearlington.com/cImg/bike%20rodeo.jpg\\r\\nhttp://farm1.static.flickr.com/30/45483822_234c3da8ae.jpg\\r\\nhttp://0518.paginamarkt.nl/foto/2007/04/09/1%20Atra%20fiets%207%20april07.JPG\\r\\nhttp://farm4.static.flickr.com/3101/2565087700_176e3e1191.jpg\\r\\nhttp://farm3.static.flickr.com/2296/2537746990_c75cc082de.jpg\\r\\nhttp://static.flickr.com/49/112916007_0286ec89d7.jpg\\r\\nhttp://static.flickr.com/62/158779964_679eff5b11.jpg\\r\\nhttp://farm4.static.flickr.com/3133/2548661176_ce8d71ca6e.jpg\\r\\nhttp://www.frederickpedalers.org/images/ventoux_babe.jpg\\r\\nhttp://farm4.static.flickr.com/3104/2449186636_3a57082892.jpg\\r\\nhttp://www.viktorschreckengost.org/ThreeWorlds/bicycle-200x141\\r\\nhttp://farm1.static.flickr.com/201/491118033_f6c22ff4e4.jpg\\r\\nhttp://www.kulukin.com/bikes/index_files/electric_bicycle_electrec.jpg\\r\\nhttp://farm2.static.flickr.com/1293/657457113_0053ac68a7.jpg\\r\\nhttp://static.flickr.com/1057/1163488341_adb5becf5f.jpg\\r\\nhttp://farm2.static.flickr.com/1099/695642465_c29e3e71ba.jpg\\r\\nhttp://static.flickr.com/1105/528032678_f5a2ed40a7.jpg\\r\\nhttp://farm1.static.flickr.com/214/494568844_013a8f51eb.jpg\\r\\nhttp://www.craave.net/images/Ferrari_cx_70_bicycle.jpg\\r\\nhttp://tusb.stanford.edu/bike.JPG\\r\\nhttp://farm1.static.flickr.com/77/180551023_bc109c61c3.jpg\\r\\nhttp://farm3.static.flickr.com/2255/2471799154_bee9bb2727.jpg\\r\\nhttp://farm1.static.flickr.com/187/435525680_7367a3217b.jpg\\r\\nhttp://www.ski-epic.com/amsterdam_bicycles/po1s_amsterdam_bicycle_suit.jpg\\r\\nhttp://davie.tv/DavTV/images/BikeRally05.JPG\\r\\nhttp://farm1.static.flickr.com/10/17051385_7be0113a81.jpg\\r\\nhttp://farm3.static.flickr.com/2103/1730155434_6596c76753.jpg\\r\\nhttp://farm1.static.flickr.com/23/34344147_70aa44faa6.jpg\\r\\nhttp://static.flickr.com/16/22966173_d19d24625b.jpg\\r\\nhttp://farm1.static.flickr.com/86/229510007_4bde88ad6e.jpg\\r\\nhttp://images.mirror.co.uk/upl/article/13298887/2008/07/24/12271794.jpg\\r\\nhttp://farm4.static.flickr.com/3017/2574312601_8e57ef8ea4.jpg\\r\\nhttp://farm2.static.flickr.com/1174/1286959674_ebf68cf72c.jpg\\r\\nhttp://farm2.static.flickr.com/1185/527942820_988d66ea54.jpg\\r\\nhttp://farm2.static.flickr.com/1016/1268496403_3f7294cd8e.jpg\\r\\nhttp://static.flickr.com/103/283871397_83debc48a4.jpg\\r\\nhttp://farm3.static.flickr.com/2003/1903295544_11c72b40ce.jpg\\r\\nhttp://nearlife.org/img/tassajara_blk_red.jpg\\r\\nhttp://farm4.static.flickr.com/3091/2432693763_f8eef5bd1e.jpg\\r\\nhttp://www.ixiqi.com/wp-content/uploads/2008/03/080301bike.jpg\\r\\nhttp://farm1.static.flickr.com/225/450709462_d8a9f9bab8.jpg\\r\\nhttp://farm3.static.flickr.com/2273/2099196751_2ed9a356e4.jpg\\r\\nhttp://admissions.uwstout.edu/hdl/article/media/bike1.jpg\\r\\nhttp://farm1.static.flickr.com/151/380566947_713e11ca47.jpg\\r\\nhttp://static.flickr.com/1089/537782512_19d4dc8c33.jpg\\r\\nhttp://www.ganahl.info/5pict/dontstealpuma0863s.jpg\\r\\nhttp://farm1.static.flickr.com/15/21138166_15f4afc40b.jpg\\r\\nhttp://www.greatdreams.com/1987bike.gif\\r\\nhttp://www.rocbike.com/wp-content/uploads/2007/07/adam_bike.jpg\\r\\nhttp://static.flickr.com/54/124401863_082a97821f.jpg\\r\\nhttp://farm1.static.flickr.com/37/116815657_c89be3f452.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/2/25/Blue_Bicycle.JPG\\r\\nhttp://static.flickr.com/1220/898264449_ac1657225f.jpg\\r\\nhttp://farm3.static.flickr.com/2282/1659291172_3b69876bb0.jpg\\r\\nhttp://farm1.static.flickr.com/51/142342159_090cd2f903.jpg\\r\\nhttp://static.flickr.com/8/8227653_a25782fe58.jpg\\r\\nhttp://farm3.static.flickr.com/2223/1657084072_a9e8a0f5e0.jpg\\r\\nhttp://farm2.static.flickr.com/1249/844443982_3f263cc62e.jpg\\r\\nhttp://www.kriskuypers.be/fotos/fiets.jpg\\r\\nhttp://static.flickr.com/31/102132747_51cf40dc18.jpg\\r\\nhttp://www.theolympictower.com/images/Bicycle%20Storage%20Room.JPG\\r\\nhttp://farm1.static.flickr.com/75/204488471_b43b299aa2.jpg\\r\\nhttp://farm1.static.flickr.com/90/257287765_774afe23d9.jpg\\r\\nhttp://www.waterkant.net/suriname/newspictures/images/20061018222219_politie_fiets.jpg\\r\\nhttp://bike-pgh.org/wp-content/uploads/bike-rack-rei.JPG\\r\\nhttp://www.wgbesseling.nl/foto/fiets761.jpg\\r\\nhttp://farm2.static.flickr.com/1411/746707551_1185120944.jpg\\r\\nhttp://farm1.static.flickr.com/83/245449716_18409342c6.jpg\\r\\nhttp://farm1.static.flickr.com/18/24010204_de19f464d7.jpg\\r\\nhttp://www.atob.org.uk/More%20illustrations/Oxygen.jpg\\r\\nhttp://www.treehugger.com/nyc-bicycle-parking-bike-shelter-street-cycling-union-square.jpg\\r\\nhttp://farm1.static.flickr.com/83/209352012_5049dbe558.jpg\\r\\nhttp://farm2.static.flickr.com/1288/554790188_43afd728fc.jpg\\r\\nhttp://www.michigan-sportsman.com/hunting_articles/hunting_library_pictures/mountain_bike_lights.jpg\\r\\nhttp://farm3.static.flickr.com/2413/2217361212_7cd513dfa6.jpg\\r\\nhttp://arnica.csustan.edu/mtbike/bike/bike8.jpg\\r\\nhttp://www.thinkcooper.com/bicycle/FlourescentRat01.jpg\\r\\nhttp://farm1.static.flickr.com/94/278144705_5bd6452401.jpg\\r\\nhttp://farm3.static.flickr.com/2207/2404832112_2815859692.jpg\\r\\nhttp://farm4.static.flickr.com/3025/2538826233_19bc188fa0.jpg\\r\\nhttp://megan.dnserver.net.nz/wp-content/bikes.jpg\\r\\nhttp://www.thenewhomemaker.com/files/bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/131/339745168_0b69e786ec.jpg\\r\\nhttp://farm1.static.flickr.com/64/158325768_59db238fcd.jpg\\r\\nhttp://brandweer.dronten.net/e107_files/images/materieel/FIETS.JPG\\r\\nhttp://farm1.static.flickr.com/212/506098967_14ad6f4557.jpg\\r\\nhttp://farm3.static.flickr.com/2088/2277118126_10888d9c7b.jpg\\r\\nhttp://farm3.static.flickr.com/2146/2303929959_03083c9196.jpg\\r\\nhttp://www.cambriabike.com/Images/product/tomac_carbidexc1_bike_08.jpg\\r\\nhttp://www.audiworld.com/news/03/bike/bike.jpg\\r\\nhttp://farm2.static.flickr.com/1177/1470535933_33e73bf37d.jpg\\r\\nhttp://img.2dehands.nl/f/normal/44321006_6-mooie-goede-dames-fiets-rivel-sirius-tourline-7-versnelling.jpg\\r\\nhttp://farm1.static.flickr.com/71/203251265_e3b0ffeb4d.jpg\\r\\nhttp://bp2.blogger.com/_4cb3Z_sg09Q/R5aQNpTYTNI/AAAAAAAAAj4/jHsVuxl-eYc/s400/bike1.jpg\\r\\nhttp://farm3.static.flickr.com/2140/2288475281_5cff123a64.jpg\\r\\nhttp://static.flickr.com/132/366602235_63c7a524fb.jpg\\r\\nhttp://members.lycos.nl/reinierbremer/hpbimg/mijn%20fiets.jpg\\r\\nhttp://farm3.static.flickr.com/2112/2086907780_81cb31e067.jpg\\r\\nhttp://static.flickr.com/42/96672910_054ed822fd.jpg\\r\\nhttp://static.flickr.com/117/317202101_97ca2fafcd.jpg\\r\\nhttp://farm1.static.flickr.com/168/457619379_4ef9513aa9.jpg\\r\\nhttp://farm1.static.flickr.com/150/416058099_c88dcd01fc.jpg\\r\\nhttp://www.bikearlington.com/cImg/Diana%20bike.jpg\\r\\nhttp://static.flickr.com/1111/1341628857_924ef35994.jpg\\r\\nhttp://farm1.static.flickr.com/168/411795065_7d7c6e44c2.jpg\\r\\nhttp://farm1.static.flickr.com/104/291437114_54dfcd4ee4.jpg\\r\\nhttp://farm3.static.flickr.com/2085/2496272122_d1d9d80ff0.jpg\\r\\nhttp://eduserv.kahosl.be/sovo/Aalst/images/fiets.jpg\\r\\nhttp://static.flickr.com/175/470971336_915fe88779.jpg\\r\\nhttp://farm3.static.flickr.com/2237/2173983253_3f5f629029.jpg\\r\\nhttp://static.flickr.com/212/449013824_06f180a6fc.jpg\\r\\nhttp://farm4.static.flickr.com/3044/2672820822_1582782bc6.jpg\\r\\nhttp://static.flickr.com/1106/1102919475_88330ba8ff.jpg\\r\\nhttp://farm4.static.flickr.com/3275/2458406671_666c275e23.jpg\\r\\nhttp://www.metaefficient.com/bionx_montague_folding_electric_bike.jpg\\r\\nhttp://static.flickr.com/7/10829494_741fcad6e8.jpg\\r\\nhttp://farm3.static.flickr.com/2038/2123359862_90093ee25f.jpg\\r\\nhttp://farm2.static.flickr.com/1378/1333671180_1dd46aa3fb.jpg\\r\\nhttp://static.flickr.com/44/151010230_e06777dcb9.jpg\\r\\nhttp://www.hazelantiques.com/Trader_s_Corner/Radio_FlyerTM/Classic_Bicycle/37_bicycle_large.JPG\\r\\nhttp://sentwest.nl/images/FIETSEN/fiets02.jpg\\r\\nhttp://farm1.static.flickr.com/17/19624966_ebd70866a4.jpg\\r\\nhttp://static.flickr.com/2/1656056_875c4bc563.jpg\\r\\nhttp://static.flickr.com/32/54143389_e2d42eea7e.jpg\\r\\nhttp://farm1.static.flickr.com/163/411791970_ae716a412a.jpg\\r\\nhttp://farm3.static.flickr.com/2145/2450082450_eb096bd5dd.jpg\\r\\nhttp://static.flickr.com/2325/2089878380_6fff3866e0.jpg\\r\\nhttp://farm1.static.flickr.com/182/474524034_797a37aa65.jpg\\r\\nhttp://farm3.static.flickr.com/2043/2283588949_0e5e41cc33.jpg\\r\\nhttp://farm3.static.flickr.com/2200/2182170399_5496753b53.jpg\\r\\nhttp://bicycle-mania.com/yahoo_site_admin/assets/images/BicycleManiaStorePictureInside.148150854_std.JPG\\r\\nhttp://farm1.static.flickr.com/80/240909749_de4d32db2c.jpg\\r\\nhttp://farm3.static.flickr.com/2037/2242985188_01c280481f.jpg\\r\\nhttp://farm1.static.flickr.com/252/524675684_9e3111f3f6.jpg\\r\\nhttp://farm1.static.flickr.com/51/128028854_047590ce8f.jpg\\r\\nhttp://homepage.ntlworld.com/catfoodrob/choppers/images/models/se(alex).jpg\\r\\nhttp://static.flickr.com/2241/2184087767_8268d4f468.jpg\\r\\nhttp://farm3.static.flickr.com/2370/1736910126_02dd4bb615.jpg\\r\\nhttp://static.flickr.com/28/92217891_f158f9ef8f.jpg\\r\\nhttp://farm4.static.flickr.com/3205/2309267985_48a8186632.jpg\\r\\nhttp://farm4.static.flickr.com/3218/2672000901_6df1cd2939.jpg\\r\\nhttp://farm2.static.flickr.com/1345/1475869509_204ef339e3.jpg\\r\\nhttp://farm4.static.flickr.com/3054/2516865912_0e72b8e240.jpg\\r\\nhttp://static.flickr.com/50/142224592_ab84e93fec.jpg\\r\\nhttp://farm1.static.flickr.com/95/216003815_ef509374d2.jpg\\r\\nhttp://farm1.static.flickr.com/59/165189498_25ec3fa279.jpg\\r\\nhttp://farm2.static.flickr.com/1094/824117577_c0453a7283.jpg\\r\\nhttp://farm2.static.flickr.com/1115/530602592_8fc6c99166.jpg\\r\\nhttp://farm1.static.flickr.com/161/413380853_e443953d80.jpg\\r\\nhttp://farm3.static.flickr.com/2012/2485128201_9e96fa297b.jpg\\r\\nhttp://farm3.static.flickr.com/2210/1911967391_b49d0457a2.jpg\\r\\nhttp://farm1.static.flickr.com/37/122322482_c276c02b4e.jpg\\r\\nhttp://farm1.static.flickr.com/78/162302697_0fe9136e95.jpg\\r\\nhttp://farm3.static.flickr.com/2150/2125568070_99d3f23a0e.jpg\\r\\nhttp://farm2.static.flickr.com/1357/1295357883_9acc1d001d.jpg\\r\\nhttp://static.flickr.com/32/40975666_7bfd3b43f5.jpg\\r\\nhttp://farm4.static.flickr.com/3047/2583602856_85232350fb.jpg\\r\\nhttp://weblogs3.nrc.nl/wielen/wp-content/uploads/fiets.jpg\\r\\nhttp://www2.btt.com.ar/foto/o/11/65/1165017253_zoo-python06bike.jpg\\r\\nhttp://farm1.static.flickr.com/70/158388384_99a16718c0.jpg\\r\\nhttp://farm1.static.flickr.com/28/46244394_b67cf694d6.jpg\\r\\nhttp://farm1.static.flickr.com/253/458258255_6dec9b8010.jpg\\r\\nhttp://farm1.static.flickr.com/41/122522019_abab060354.jpg\\r\\nhttp://farm1.static.flickr.com/3/5970859_0c4bd13e09.jpg\\r\\nhttp://farm4.static.flickr.com/3200/2602826204_4e53fcf355.jpg\\r\\nhttp://img.alibaba.com/photo/11015636/26_21_Speed_Dual_Suspension_Mtb_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/31/49902206_cd5ae5e2bc.jpg\\r\\nhttp://farm1.static.flickr.com/23/26833239_865e9ee1b7.jpg\\r\\nhttp://wvs.topleftpixel.com/photos/panning_bike_king_yellow_bag.jpg\\r\\nhttp://farm1.static.flickr.com/249/517427529_7027e9ca42.jpg\\r\\nhttp://farm3.static.flickr.com/2151/1736910192_26883df5f7.jpg\\r\\nhttp://farm1.static.flickr.com/224/527981218_5a2a95084a.jpg\\r\\nhttp://www.mountain-bikes-now.com/mountain_bikes.jpg\\r\\nhttp://weblog.columni.st/blogimages/fiets.jpg\\r\\nhttp://farm2.static.flickr.com/1125/573602062_9063a3d309.jpg\\r\\nhttp://static.flickr.com/44/150361429_6f7ae5a02a.jpg\\r\\nhttp://static.flickr.com/222/458930398_b9de63e446.jpg\\r\\nhttp://static.flickr.com/174/411802860_3db442b09e.jpg\\r\\nhttp://www.scottneyedli.com/FFT/Images/PageImages/SNImages/My%20Bike.jpg\\r\\nhttp://farm1.static.flickr.com/97/255123183_373868cdb0.jpg\\r\\nhttp://farm3.static.flickr.com/2282/2010563357_0656cad455.jpg\\r\\nhttp://www.atomicrack.com/images/Bicycle-Rack-Easy_to_use%20(1).jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/d/d7/NHS_bicycle.jpg/800px-NHS_bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/22/95686545_3accd52eb8.jpg\\r\\nhttp://farm3.static.flickr.com/2355/1803984008_b10471ba0b.jpg\\r\\nhttp://farm4.static.flickr.com/3073/2618776916_e6a11eb718.jpg\\r\\nhttp://www.berlare.be/nieuws/fietsen/fiets_1.jpg\\r\\nhttp://www.prweb.com/prfiles/2006/08/24/428963/izip_splash_01.jpg\\r\\nhttp://static.flickr.com/21/27146129_7acc493960.jpg\\r\\nhttp://farm1.static.flickr.com/160/359125387_31f941f291.jpg\\r\\nhttp://farm3.static.flickr.com/2254/2450584797_93a017651e.jpg\\r\\nhttp://www.rodekruissneek.nl/Rode%20Kruis%20Sneek/Nieuwsfotoos/Elfstedenrijwieltocht2007/Elfstedenrijwieltocht%202007%20004.jpg\\r\\nhttp://static.flickr.com/24/47992802_196494fe20.jpg\\r\\nhttp://farm1.static.flickr.com/137/361960861_58cbf01339.jpg\\r\\nhttp://www.cyclesg.com/images/home_jeep01.jpg\\r\\nhttp://farm4.static.flickr.com/3177/2612623788_63c76e905c.jpg\\r\\nhttp://farm1.static.flickr.com/50/147389863_7b7ace98ef.jpg\\r\\nhttp://farm1.static.flickr.com/77/216014372_431e209988.jpg\\r\\nhttp://farm2.static.flickr.com/1131/943802337_530fc3d452.jpg\\r\\nhttp://farm1.static.flickr.com/73/181949535_2d426d8210.jpg\\r\\nhttp://img181.imageshack.us/img181/4162/dc070220002ab4.jpg\\r\\nhttp://farm1.static.flickr.com/246/528046625_5c09857c6a.jpg\\r\\nhttp://wagaman.org/wp-content/bicycle.gif\\r\\nhttp://static.flickr.com/30/96389180_c5e58dc19a.jpg\\r\\nhttp://farm1.static.flickr.com/24/42493143_6005a80aa4.jpg\\r\\nhttp://www.treehugger.com/mcdonalds-cycle-center-chicago.jpg\\r\\nhttp://www.co.greene.oh.us/parks/images/bike-path-family.jpg\\r\\nhttp://farm2.static.flickr.com/1315/1374917000_5aa3388136.jpg?v=0\\r\\nhttp://www.btt.com.ar/foto/o/10/81/1081539802_my_bike.jpg\\r\\nhttp://farm2.static.flickr.com/1227/571179238_898fbddeaf.jpg\\r\\nhttp://img2.pcauto.com.cn/pcauto/0701/29/426004_1_thumb.jpg\\r\\nhttp://farm4.static.flickr.com/3046/2587878826_a18a8a27a1.jpg\\r\\nhttp://farm3.static.flickr.com/2154/2010942357_f0100afebd.jpg\\r\\nhttp://farm1.static.flickr.com/152/351524443_92b75438cd.jpg\\r\\nhttp://farm4.static.flickr.com/3042/2562237001_9b3cee0b98.jpg\\r\\nhttp://farm1.static.flickr.com/19/107408886_71dfe1bc57.jpg\\r\\nhttp://farm4.static.flickr.com/3171/2575319823_afb24d0461.jpg\\r\\nhttp://byfiles.storage.live.com/y1pKZbmfpECnkYsF33WQrvGiLp7QdPWiCwWySHXCaJX27nvVLeMEXMp2WKQrErRiDBgxv8Nezeu6iw\\r\\nhttp://www.pssc.cn/sucai/pic/sport/002/bike/sport_9898744167.jpg\\r\\nhttp://farm1.static.flickr.com/18/23557193_0a2f906efd.jpg\\r\\nhttp://farm4.static.flickr.com/3132/2403552159_21fff55ae9.jpg\\r\\nhttp://farm4.static.flickr.com/3075/2607527425_6b38438084.jpg\\r\\nhttp://readytoride.biz/wordpress/wp-content/uploads/2007/05/building_the_long_distance_bicycle(2).jpg\\r\\nhttp://farm3.static.flickr.com/2242/1864824837_1b94aed9e3.jpg\\r\\nhttp://farm1.static.flickr.com/53/147389938_bba3e1fbc7.jpg\\r\\nhttp://www.cis.nctu.edu.tw/~whtsai/Pictures%20in%20Personal%20Homepages/Japan%202004%20---%2002%20Riding%20a%202-seated%20bicycle%20with%20Mamma%20in%20Onuma%20Park.jpg\\r\\nhttp://farm3.static.flickr.com/2373/2342769285_3d7b570e5f.jpg\\r\\nhttp://farm4.static.flickr.com/3260/2562209768_3e31df4084.jpg\\r\\nhttp://farm4.static.flickr.com/3296/2308366067_e90f6aeef2.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/thumb/0/07/Parking_a_velo_-_gouda.JPG/800px-Parking_a_velo_-_gouda.JPG\\r\\nhttp://farm2.static.flickr.com/1110/1337272517_3b9c357d8c.jpg\\r\\nhttp://static.flickr.com/96/246600072_c6ee9fb83f.jpg\\r\\nhttp://farm1.static.flickr.com/56/144439295_d5bdc2e693.jpg\\r\\nhttp://farm1.static.flickr.com/49/127052062_1abbeed68a.jpg\\r\\nhttp://www.rtrmag.com/images/BarcroftcolumbiaRecumbent.jpg\\r\\nhttp://farm3.static.flickr.com/2118/1605319376_01eb10de16.jpg\\r\\nhttp://farm1.static.flickr.com/75/164890274_5b1bfb4470.jpg\\r\\nhttp://farm1.static.flickr.com/65/201965684_35a31b67e7.jpg\\r\\nhttp://static.flickr.com/59/171680723_a7efc4f239.jpg\\r\\nhttp://farm3.static.flickr.com/2151/2340342932_6cef2e8f7d.jpg\\r\\nhttp://farm1.static.flickr.com/100/290385318_983a8ec57b.jpg\\r\\nhttp://www.7stanes.com/images/oldbike_200.jpg\\r\\nhttp://www.zeg.de/img/product/subnavi_80.jpg\\r\\nhttp://www.uoguelph.ca/campus/map/img/bike_center.jpg\\r\\nhttp://farm3.static.flickr.com/2205/1752613537_2819c8c712.jpg\\r\\nhttp://thedailyrandonneur.files.wordpress.com/2007/12/kogswell-bars-side-view.jpg\\r\\nhttp://farm1.static.flickr.com/56/116297873_3a6546efb5.jpg\\r\\nhttp://farm4.static.flickr.com/3043/2331679011_37c2ef5f4c.jpg\\r\\nhttp://farm2.static.flickr.com/1437/689042680_46372dd843.jpg\\r\\nhttp://farm1.static.flickr.com/67/158908663_b86e3d12b6.jpg\\r\\nhttp://farm4.static.flickr.com/3100/2304911061_d61a194780.jpg\\r\\nhttp://cycling-vietnam.com/xeden1.jpg\\r\\nhttp://www.depts.ttu.edu/communications/news/stories/images/angledBike-big.jpg\\r\\nhttp://www.chosensites.com/bicycle.jpg\\r\\nhttp://farm4.static.flickr.com/3184/2342847896_07e499b20c.jpg\\r\\nhttp://farm1.static.flickr.com/174/456698204_664a516edb.jpg\\r\\nhttp://www.carbuyersnotebook.com/archives/Mercedes-Bike.jpg\\r\\nhttp://www.mosquito-bikes.co.uk/htdocs/cms_ul/cata_202/smoothieES_brik2%20small.jpg\\r\\nhttp://blog.tiffany.com.tw/up/1146029290.jpg\\r\\nhttp://farm1.static.flickr.com/165/413801550_40a1418eb7.jpg\\r\\nhttp://upload.wikimedia.org/wikipedia/commons/3/38/Bike-park-aarhus.jpg\\r\\nhttp://img.alibaba.com/photo/51302292/Electric_Bicycle.jpg\\r\\nhttp://farm1.static.flickr.com/81/277306005_0ac0b4eed1.jpg\\r\\nhttp://www.fietsberaad.nl/library/repository/nieuws/ing_fiets.jpg\\r\\nhttp://farm4.static.flickr.com/3074/2303177969_661a254d9a.jpg\\r\\nhttp://static.flickr.com/1238/1483444007_a48564e1f8.jpg\\r\\nhttp://farm3.static.flickr.com/2322/2485121641_5fd349882e.jpg\\r\\nhttp://www.highwycombecc.org/images/thames_velo_team.jpg\\r\\nhttp://farm1.static.flickr.com/23/25802576_72e0d28e07.jpg\\r\\nhttp://static.flickr.com/2055/2109407065_5cabdf1cec.jpg\\r\\nhttp://farm3.static.flickr.com/2244/2078780945_b71fb87ab3.jpg\\r\\nhttp://farm1.static.flickr.com/68/216014363_0c513ef3f2.jpg\\r\\nhttp://farm2.static.flickr.com/1326/1154285130_27dbb0bebd.jpg\\r\\n\\n\"\n",
            "1262\n",
            "1345\n",
            "0\n",
            "20\n",
            "40\n",
            "60\n",
            "80\n",
            "100\n",
            "120\n",
            "140\n",
            "0\n",
            "20\n",
            "40\n",
            "60\n",
            "80\n",
            "100\n",
            "120\n",
            "140\n",
            "0\n",
            "20\n",
            "40\n",
            "0\n",
            "20\n",
            "40\n",
            "\n",
            "TRAIN:\n",
            "\n",
            "\n",
            "list the files inside ships directory:\n",
            "\n",
            "img0.jpg    img122.jpg\timg141.jpg  img24.jpg  img46.jpg  img63.jpg  img8.jpg\n",
            "img102.jpg  img125.jpg\timg142.jpg  img25.jpg  img47.jpg  img64.jpg  img91.jpg\n",
            "img103.jpg  img127.jpg\timg143.jpg  img26.jpg  img48.jpg  img66.jpg  img92.jpg\n",
            "img105.jpg  img12.jpg\timg144.jpg  img28.jpg  img49.jpg  img67.jpg  img93.jpg\n",
            "img107.jpg  img130.jpg\timg146.jpg  img29.jpg  img4.jpg   img69.jpg  img94.jpg\n",
            "img108.jpg  img131.jpg\timg147.jpg  img31.jpg  img51.jpg  img6.jpg   img95.jpg\n",
            "img111.jpg  img132.jpg\timg148.jpg  img32.jpg  img52.jpg  img72.jpg  img96.jpg\n",
            "img113.jpg  img133.jpg\timg149.jpg  img33.jpg  img54.jpg  img73.jpg  img97.jpg\n",
            "img115.jpg  img135.jpg\timg14.jpg   img34.jpg  img55.jpg  img74.jpg  img99.jpg\n",
            "img116.jpg  img136.jpg\timg15.jpg   img35.jpg  img56.jpg  img76.jpg\n",
            "img117.jpg  img137.jpg\timg16.jpg   img39.jpg  img59.jpg  img77.jpg\n",
            "img118.jpg  img138.jpg\timg17.jpg   img3.jpg   img5.jpg   img78.jpg\n",
            "img11.jpg   img139.jpg\timg18.jpg   img40.jpg  img60.jpg  img83.jpg\n",
            "img120.jpg  img140.jpg\timg23.jpg   img43.jpg  img62.jpg  img86.jpg\n",
            "\n",
            "list the files inside bikes directory:\n",
            "\n",
            "img100.jpg  img118.jpg\timg138.jpg  img21.jpg  img48.jpg  img67.jpg  img86.jpg\n",
            "img101.jpg  img119.jpg\timg13.jpg   img22.jpg  img4.jpg   img6.jpg   img87.jpg\n",
            "img102.jpg  img120.jpg\timg141.jpg  img26.jpg  img51.jpg  img70.jpg  img88.jpg\n",
            "img103.jpg  img122.jpg\timg142.jpg  img27.jpg  img53.jpg  img71.jpg  img89.jpg\n",
            "img104.jpg  img124.jpg\timg144.jpg  img31.jpg  img54.jpg  img72.jpg  img90.jpg\n",
            "img105.jpg  img126.jpg\timg145.jpg  img33.jpg  img56.jpg  img73.jpg  img93.jpg\n",
            "img106.jpg  img127.jpg\timg146.jpg  img35.jpg  img58.jpg  img74.jpg  img95.jpg\n",
            "img107.jpg  img129.jpg\timg147.jpg  img36.jpg  img59.jpg  img75.jpg  img98.jpg\n",
            "img108.jpg  img130.jpg\timg148.jpg  img37.jpg  img5.jpg   img76.jpg  img99.jpg\n",
            "img10.jpg   img132.jpg\timg149.jpg  img3.jpg   img60.jpg  img77.jpg  img9.jpg\n",
            "img110.jpg  img133.jpg\timg14.jpg   img41.jpg  img63.jpg  img78.jpg\n",
            "img112.jpg  img134.jpg\timg15.jpg   img43.jpg  img64.jpg  img82.jpg\n",
            "img115.jpg  img135.jpg\timg17.jpg   img45.jpg  img65.jpg  img83.jpg\n",
            "img117.jpg  img136.jpg\timg19.jpg   img47.jpg  img66.jpg  img84.jpg\n",
            "\n",
            "VALIDATION:\n",
            "\n",
            "\n",
            "list the files inside ships directory:\n",
            "\n",
            "img0.jpg   img15.jpg  img21.jpg  img34.jpg  img44.jpg  img49.jpg  img7.jpg\n",
            "img11.jpg  img16.jpg  img25.jpg  img36.jpg  img45.jpg  img4.jpg\n",
            "img12.jpg  img17.jpg  img30.jpg  img38.jpg  img47.jpg  img5.jpg\n",
            "img13.jpg  img18.jpg  img31.jpg  img3.jpg   img48.jpg  img6.jpg\n",
            "\n",
            "list the files inside bikes directory:\n",
            "\n",
            "img0.jpg   img15.jpg  img25.jpg  img30.jpg  img35.jpg  img42.jpg  img49.jpg\n",
            "img10.jpg  img18.jpg  img26.jpg  img31.jpg  img37.jpg  img43.jpg  img5.jpg\n",
            "img11.jpg  img19.jpg  img27.jpg  img32.jpg  img38.jpg  img45.jpg  img6.jpg\n",
            "img13.jpg  img21.jpg  img29.jpg  img33.jpg  img40.jpg  img46.jpg  img7.jpg\n",
            "img14.jpg  img24.jpg  img2.jpg\t img34.jpg  img41.jpg  img47.jpg\n",
            "0\n",
            "20\n",
            "40\n",
            "0\n",
            "20\n",
            "40\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sP__IrNZMKf1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import pandas as pd\n",
        "# Path to the train folder\n",
        "original_test = '/content/test/test'\n",
        " \n",
        "filenames = os.listdir(original_test)\n",
        "categories = []\n",
        "for filename in filenames:\n",
        "    category = filename.split('_')[0]\n",
        "    if category == 'ships':\n",
        "        categories.append('ships')\n",
        "    else:\n",
        "        categories.append('bikes')\n",
        "\n",
        "data_test = pd.DataFrame({'filename':filenames,'label':categories})\n",
        "\n",
        "data_test.to_csv(\"original_test.csv\",index=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qPTrBa_oMqmQ",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning using VGG16"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fkxopbQPMr9l",
        "colab_type": "code",
        "outputId": "60b735cc-7702-4eeb-a236-10b3f3b06ed3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 50\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1QIoIKhfOK1A",
        "colab_type": "code",
        "outputId": "b280f421-1ac0-4218-ce47-ba07f4da6069",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QzsrjqrGOSfW",
        "colab_type": "code",
        "outputId": "9a9ed330-33d4-434f-952c-b65f12fc1f42",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.VGG16(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "58892288/58889256 [==============================] - 2s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:12: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "  if sys.path[0] == '':\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:40: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:40: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=50, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "11/11 [==============================] - 11s 1s/step - loss: 0.7714 - accuracy: 0.4912 - val_loss: 0.7943 - val_accuracy: 0.3730\n",
            "Epoch 2/50\n",
            "11/11 [==============================] - 5s 477ms/step - loss: 0.7578 - accuracy: 0.5146 - val_loss: 0.7794 - val_accuracy: 0.3717\n",
            "Epoch 3/50\n",
            "11/11 [==============================] - 5s 477ms/step - loss: 0.7903 - accuracy: 0.4561 - val_loss: 0.7666 - val_accuracy: 0.3751\n",
            "Epoch 4/50\n",
            "11/11 [==============================] - 5s 473ms/step - loss: 0.7252 - accuracy: 0.5511 - val_loss: 0.7745 - val_accuracy: 0.3717\n",
            "Epoch 5/50\n",
            "11/11 [==============================] - 5s 471ms/step - loss: 0.7236 - accuracy: 0.5904 - val_loss: 0.7614 - val_accuracy: 0.3730\n",
            "Epoch 6/50\n",
            "11/11 [==============================] - 5s 475ms/step - loss: 0.7836 - accuracy: 0.5322 - val_loss: 0.7604 - val_accuracy: 0.3751\n",
            "Epoch 7/50\n",
            "11/11 [==============================] - 5s 467ms/step - loss: 0.7375 - accuracy: 0.5556 - val_loss: 0.7377 - val_accuracy: 0.3717\n",
            "Epoch 8/50\n",
            "11/11 [==============================] - 5s 456ms/step - loss: 0.7432 - accuracy: 0.5439 - val_loss: 0.7664 - val_accuracy: 0.3717\n",
            "Epoch 9/50\n",
            "11/11 [==============================] - 5s 459ms/step - loss: 0.7534 - accuracy: 0.5227 - val_loss: 0.7393 - val_accuracy: 0.3741\n",
            "Epoch 10/50\n",
            "11/11 [==============================] - 5s 461ms/step - loss: 0.7566 - accuracy: 0.4795 - val_loss: 0.7537 - val_accuracy: 0.3740\n",
            "Epoch 11/50\n",
            "11/11 [==============================] - 5s 458ms/step - loss: 0.7897 - accuracy: 0.4971 - val_loss: 0.7433 - val_accuracy: 0.3705\n",
            "Epoch 12/50\n",
            "11/11 [==============================] - 5s 446ms/step - loss: 0.7606 - accuracy: 0.4678 - val_loss: 0.8154 - val_accuracy: 0.3728\n",
            "Epoch 13/50\n",
            "11/11 [==============================] - 4s 406ms/step - loss: 0.7517 - accuracy: 0.5263 - val_loss: 0.7882 - val_accuracy: 0.3730\n",
            "Epoch 14/50\n",
            "11/11 [==============================] - 5s 478ms/step - loss: 0.7463 - accuracy: 0.5029 - val_loss: 0.7933 - val_accuracy: 0.3682\n",
            "Epoch 15/50\n",
            "11/11 [==============================] - 5s 476ms/step - loss: 0.7639 - accuracy: 0.5263 - val_loss: 0.7460 - val_accuracy: 0.3786\n",
            "Epoch 16/50\n",
            "11/11 [==============================] - 5s 464ms/step - loss: 0.7816 - accuracy: 0.4795 - val_loss: 0.6865 - val_accuracy: 0.3717\n",
            "Epoch 17/50\n",
            "11/11 [==============================] - 5s 460ms/step - loss: 0.7650 - accuracy: 0.5205 - val_loss: 0.7726 - val_accuracy: 0.3753\n",
            "Epoch 18/50\n",
            "11/11 [==============================] - 5s 469ms/step - loss: 0.7030 - accuracy: 0.5455 - val_loss: 0.7200 - val_accuracy: 0.3671\n",
            "Epoch 19/50\n",
            "11/11 [==============================] - 5s 469ms/step - loss: 0.8057 - accuracy: 0.4458 - val_loss: 0.7643 - val_accuracy: 0.3763\n",
            "Epoch 20/50\n",
            "11/11 [==============================] - 5s 454ms/step - loss: 0.7695 - accuracy: 0.5057 - val_loss: 0.7697 - val_accuracy: 0.3728\n",
            "Epoch 21/50\n",
            "11/11 [==============================] - 5s 461ms/step - loss: 0.7173 - accuracy: 0.5789 - val_loss: 0.7595 - val_accuracy: 0.3719\n",
            "Epoch 22/50\n",
            "11/11 [==============================] - 5s 470ms/step - loss: 0.8023 - accuracy: 0.4620 - val_loss: 0.7247 - val_accuracy: 0.3763\n",
            "Epoch 23/50\n",
            "11/11 [==============================] - 5s 458ms/step - loss: 0.7707 - accuracy: 0.4971 - val_loss: 0.7452 - val_accuracy: 0.3705\n",
            "Epoch 24/50\n",
            "11/11 [==============================] - 5s 438ms/step - loss: 0.7903 - accuracy: 0.4971 - val_loss: 0.7924 - val_accuracy: 0.3728\n",
            "Epoch 25/50\n",
            "11/11 [==============================] - 5s 416ms/step - loss: 0.6951 - accuracy: 0.5614 - val_loss: 0.7862 - val_accuracy: 0.3753\n",
            "Epoch 26/50\n",
            "11/11 [==============================] - 5s 478ms/step - loss: 0.7664 - accuracy: 0.4854 - val_loss: 0.7454 - val_accuracy: 0.3740\n",
            "Epoch 27/50\n",
            "11/11 [==============================] - 5s 479ms/step - loss: 0.7596 - accuracy: 0.5114 - val_loss: 0.7637 - val_accuracy: 0.3694\n",
            "Epoch 28/50\n",
            "11/11 [==============================] - 5s 472ms/step - loss: 0.7801 - accuracy: 0.4880 - val_loss: 0.7744 - val_accuracy: 0.3728\n",
            "Epoch 29/50\n",
            "11/11 [==============================] - 5s 462ms/step - loss: 0.7718 - accuracy: 0.4678 - val_loss: 0.7383 - val_accuracy: 0.3741\n",
            "Epoch 30/50\n",
            "11/11 [==============================] - 5s 467ms/step - loss: 0.7422 - accuracy: 0.5057 - val_loss: 0.7746 - val_accuracy: 0.3705\n",
            "Epoch 31/50\n",
            "11/11 [==============================] - 5s 468ms/step - loss: 0.7409 - accuracy: 0.5482 - val_loss: 0.7886 - val_accuracy: 0.3717\n",
            "Epoch 32/50\n",
            "11/11 [==============================] - 5s 457ms/step - loss: 0.7635 - accuracy: 0.4943 - val_loss: 0.7306 - val_accuracy: 0.3751\n",
            "Epoch 33/50\n",
            "11/11 [==============================] - 5s 461ms/step - loss: 0.7359 - accuracy: 0.5263 - val_loss: 0.8186 - val_accuracy: 0.3707\n",
            "Epoch 34/50\n",
            "11/11 [==============================] - 5s 461ms/step - loss: 0.7484 - accuracy: 0.5263 - val_loss: 0.7212 - val_accuracy: 0.3763\n",
            "Epoch 35/50\n",
            "11/11 [==============================] - 5s 449ms/step - loss: 0.7671 - accuracy: 0.5263 - val_loss: 0.7505 - val_accuracy: 0.3728\n",
            "Epoch 36/50\n",
            "11/11 [==============================] - 5s 445ms/step - loss: 0.7244 - accuracy: 0.5789 - val_loss: 0.7281 - val_accuracy: 0.3717\n",
            "Epoch 37/50\n",
            "11/11 [==============================] - 4s 408ms/step - loss: 0.7943 - accuracy: 0.4886 - val_loss: 0.7545 - val_accuracy: 0.3730\n",
            "Epoch 38/50\n",
            "11/11 [==============================] - 5s 478ms/step - loss: 0.7101 - accuracy: 0.5322 - val_loss: 0.7276 - val_accuracy: 0.3740\n",
            "Epoch 39/50\n",
            "11/11 [==============================] - 5s 474ms/step - loss: 0.7416 - accuracy: 0.4880 - val_loss: 0.7458 - val_accuracy: 0.3728\n",
            "Epoch 40/50\n",
            "11/11 [==============================] - 5s 473ms/step - loss: 0.7933 - accuracy: 0.4773 - val_loss: 0.7623 - val_accuracy: 0.3717\n",
            "Epoch 41/50\n",
            "11/11 [==============================] - 5s 469ms/step - loss: 0.7637 - accuracy: 0.5422 - val_loss: 0.7753 - val_accuracy: 0.3753\n",
            "Epoch 42/50\n",
            "11/11 [==============================] - 5s 467ms/step - loss: 0.7819 - accuracy: 0.4912 - val_loss: 0.7653 - val_accuracy: 0.3705\n",
            "Epoch 43/50\n",
            "11/11 [==============================] - 5s 462ms/step - loss: 0.7922 - accuracy: 0.4545 - val_loss: 0.7875 - val_accuracy: 0.3740\n",
            "Epoch 44/50\n",
            "11/11 [==============================] - 5s 446ms/step - loss: 0.7939 - accuracy: 0.5029 - val_loss: 0.7664 - val_accuracy: 0.3717\n",
            "Epoch 45/50\n",
            "11/11 [==============================] - 5s 452ms/step - loss: 0.7319 - accuracy: 0.5380 - val_loss: 0.7356 - val_accuracy: 0.3730\n",
            "Epoch 46/50\n",
            "11/11 [==============================] - 5s 458ms/step - loss: 0.7686 - accuracy: 0.4880 - val_loss: 0.7856 - val_accuracy: 0.3671\n",
            "Epoch 47/50\n",
            "11/11 [==============================] - 5s 453ms/step - loss: 0.7156 - accuracy: 0.5455 - val_loss: 0.7327 - val_accuracy: 0.3786\n",
            "Epoch 48/50\n",
            "11/11 [==============================] - 5s 444ms/step - loss: 0.7929 - accuracy: 0.4211 - val_loss: 0.7637 - val_accuracy: 0.3728\n",
            "Epoch 49/50\n",
            "11/11 [==============================] - 5s 416ms/step - loss: 0.7711 - accuracy: 0.4561 - val_loss: 0.7174 - val_accuracy: 0.3719\n",
            "Epoch 50/50\n",
            "11/11 [==============================] - 5s 478ms/step - loss: 0.7922 - accuracy: 0.4503 - val_loss: 0.7731 - val_accuracy: 0.3717\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa32e4b5390>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9o1nkiLFICma",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with ResNet50"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h_1mkL4UIHog",
        "colab_type": "code",
        "outputId": "9ea57b46-bdd9-4e82-fb96-798ecb3ff47f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 411
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet50.ResNet50(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:150]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_applications/resnet50.py:265: UserWarning: The output shape of `ResNet50(include_top=False)` has been changed since Keras 2.2.0.\n",
            "  warnings.warn('The output shape of `ResNet50(include_top=False)` '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.2/resnet50_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "94658560/94653016 [==============================] - 3s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 10s 879ms/step - loss: 1.2497 - accuracy: 0.4971 - val_loss: 0.8033 - val_accuracy: 0.4222\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 5s 493ms/step - loss: 1.0912 - accuracy: 0.5146 - val_loss: 0.6990 - val_accuracy: 0.4281\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 5s 489ms/step - loss: 0.9770 - accuracy: 0.5497 - val_loss: 0.7500 - val_accuracy: 0.4511\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 5s 481ms/step - loss: 1.0015 - accuracy: 0.5439 - val_loss: 0.6867 - val_accuracy: 0.6133\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 5s 489ms/step - loss: 0.9862 - accuracy: 0.4886 - val_loss: 0.6469 - val_accuracy: 0.6773\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa322045160>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iu8XFGZeIc5r",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with Xception"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9v5Vg0OCKGCT",
        "colab_type": "code",
        "outputId": "d54dc78a-5f6d-460f-a895-09564d8864bf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.xception.Xception(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.4/xception_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "83689472/83683744 [==============================] - 3s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 13s 1s/step - loss: 0.7097 - accuracy: 0.4854 - val_loss: 0.6995 - val_accuracy: 0.4062\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 505ms/step - loss: 0.7290 - accuracy: 0.5556 - val_loss: 0.6919 - val_accuracy: 0.3901\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 504ms/step - loss: 0.7122 - accuracy: 0.5556 - val_loss: 0.6990 - val_accuracy: 0.3705\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 5s 491ms/step - loss: 0.6966 - accuracy: 0.5614 - val_loss: 0.7144 - val_accuracy: 0.3728\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 5s 497ms/step - loss: 0.7011 - accuracy: 0.5455 - val_loss: 0.7042 - val_accuracy: 0.4096\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa31a86e978>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "45BMWCiwKaV-",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning VGG19"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z4yRWsmYKPPx",
        "colab_type": "code",
        "outputId": "1298a432-6591-4bad-a6a8-43b47b95abab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.vgg19.VGG19(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "80142336/80134624 [==============================] - 3s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 5s 487ms/step - loss: 0.7557 - accuracy: 0.4971 - val_loss: 0.6576 - val_accuracy: 0.7460\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 5s 474ms/step - loss: 0.7602 - accuracy: 0.4620 - val_loss: 0.6543 - val_accuracy: 0.7434\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 5s 478ms/step - loss: 0.7314 - accuracy: 0.4737 - val_loss: 0.6383 - val_accuracy: 0.7480\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 5s 464ms/step - loss: 0.6827 - accuracy: 0.5789 - val_loss: 0.6550 - val_accuracy: 0.7457\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 5s 462ms/step - loss: 0.7456 - accuracy: 0.5000 - val_loss: 0.6473 - val_accuracy: 0.7471\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa31bfd4f28>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ol7GRcJ2KpyM",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning pretrained ResNet101"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fdUgbZ4fKwLY",
        "colab_type": "code",
        "outputId": "9a868c36-34bb-4762-b120-a711373f7d9d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet.ResNet101(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet101_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "171450368/171446536 [==============================] - 5s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 31s 3s/step - loss: 1.1292 - accuracy: 0.4678 - val_loss: 0.6198 - val_accuracy: 0.5755\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 577ms/step - loss: 1.0876 - accuracy: 0.5205 - val_loss: 0.6814 - val_accuracy: 0.5765\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 570ms/step - loss: 0.9273 - accuracy: 0.5497 - val_loss: 0.6560 - val_accuracy: 0.5777\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 6s 564ms/step - loss: 0.8561 - accuracy: 0.5789 - val_loss: 1.0650 - val_accuracy: 0.5754\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 6s 570ms/step - loss: 0.8562 - accuracy: 0.5965 - val_loss: 0.6224 - val_accuracy: 0.5767\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2d1f5fd30>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aWaSVWuWK4d8",
        "colab_type": "text"
      },
      "source": [
        "#Load pretrained ResNet152"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bTblwTMEK-FZ",
        "colab_type": "code",
        "outputId": "e535d482-363f-4f96-91cf-f871fd4b5cca",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet.ResNet152(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet152_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "234700800/234698864 [==============================] - 7s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 50s 5s/step - loss: 1.0395 - accuracy: 0.5146 - val_loss: 0.8266 - val_accuracy: 0.5767\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 8s 703ms/step - loss: 0.8505 - accuracy: 0.5146 - val_loss: 0.9132 - val_accuracy: 0.5731\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 8s 717ms/step - loss: 0.9182 - accuracy: 0.5497 - val_loss: 0.8299 - val_accuracy: 0.4522\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 8s 703ms/step - loss: 0.9119 - accuracy: 0.5341 - val_loss: 0.6978 - val_accuracy: 0.4062\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 8s 720ms/step - loss: 0.7729 - accuracy: 0.6566 - val_loss: 0.7027 - val_accuracy: 0.4382\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2ca368a90>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IfYCPPGBLLYn",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning ResNet50v2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z7dNlsc7LQ1t",
        "colab_type": "code",
        "outputId": "558bf68e-2caf-4d31-abde-39387633390a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet_v2.ResNet50V2(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet50v2_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "94674944/94668760 [==============================] - 3s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 15s 1s/step - loss: 0.8509 - accuracy: 0.5057 - val_loss: 0.6649 - val_accuracy: 0.6121\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 523ms/step - loss: 0.9063 - accuracy: 0.4940 - val_loss: 0.7107 - val_accuracy: 0.6157\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 514ms/step - loss: 0.8848 - accuracy: 0.4620 - val_loss: 0.5903 - val_accuracy: 0.6479\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 6s 501ms/step - loss: 0.8615 - accuracy: 0.5146 - val_loss: 0.5997 - val_accuracy: 0.6617\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 6s 508ms/step - loss: 0.7673 - accuracy: 0.5739 - val_loss: 0.6823 - val_accuracy: 0.5606\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2ca8e49b0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yxByxfN4Lay8",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with ResNet101v2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5ZnbwpBkLg7_",
        "colab_type": "code",
        "outputId": "ae455690-94d6-4f79-f6f7-3b507af608cb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet_v2.ResNet101V2(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet101v2_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "171319296/171317808 [==============================] - 5s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 28s 3s/step - loss: 0.9610 - accuracy: 0.4737 - val_loss: 0.7446 - val_accuracy: 0.5606\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 572ms/step - loss: 0.9509 - accuracy: 0.5146 - val_loss: 0.7002 - val_accuracy: 0.4741\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 565ms/step - loss: 0.8935 - accuracy: 0.4795 - val_loss: 0.6768 - val_accuracy: 0.3901\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 6s 569ms/step - loss: 0.8570 - accuracy: 0.5789 - val_loss: 0.7128 - val_accuracy: 0.4246\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 6s 559ms/step - loss: 0.7829 - accuracy: 0.5284 - val_loss: 0.7346 - val_accuracy: 0.4073\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2be5fe710>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4DfL7H3iLqhS",
        "colab_type": "text"
      },
      "source": [
        "#Fine tune ResNetv152"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HwSm4JfdLvoh",
        "colab_type": "code",
        "outputId": "7d3de22d-41fc-4f13-cc9a-8cdd802d2f84",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.resnet_v2.ResNet152V2(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/resnet/resnet152v2_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "234553344/234545216 [==============================] - 7s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 47s 4s/step - loss: 0.9244 - accuracy: 0.4327 - val_loss: 0.6297 - val_accuracy: 0.5950\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 8s 697ms/step - loss: 0.7573 - accuracy: 0.5455 - val_loss: 0.6497 - val_accuracy: 0.5086\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 8s 703ms/step - loss: 0.8706 - accuracy: 0.4795 - val_loss: 0.6370 - val_accuracy: 0.5616\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 8s 695ms/step - loss: 0.7299 - accuracy: 0.5614 - val_loss: 0.6378 - val_accuracy: 0.4580\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 8s 691ms/step - loss: 0.6785 - accuracy: 0.5602 - val_loss: 0.6734 - val_accuracy: 0.4416\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2b7796a58>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JeVsqtqLL8mb",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with Inceptionv3"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Oi4OWsI1MCbF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.inception_v3.InceptionV3(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KebaDyLGMRmJ",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with InceptionResNetV2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PIX8z9huMXiE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.inception_resnet_v2.InceptionResNetV2(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3YfCh452MgJ7",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with MobileNet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6ZJTx_LfMlhT",
        "colab_type": "code",
        "outputId": "ae3efbec-11f3-47e4-d597-d89babec42cc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 411
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.mobilenet.MobileNet(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_applications/mobilenet.py:207: UserWarning: `input_shape` is undefined or non-square, or `rows` is not in [128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n",
            "  warnings.warn('`input_shape` is undefined or non-square, '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.6/mobilenet_1_0_224_tf_no_top.h5\n",
            "17227776/17225924 [==============================] - 1s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 9s 838ms/step - loss: 1.3790 - accuracy: 0.5029 - val_loss: 0.6894 - val_accuracy: 0.5435\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 5s 488ms/step - loss: 1.2721 - accuracy: 0.5439 - val_loss: 0.6883 - val_accuracy: 0.5270\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 5s 487ms/step - loss: 1.1735 - accuracy: 0.5497 - val_loss: 0.6964 - val_accuracy: 0.5386\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 5s 481ms/step - loss: 1.0442 - accuracy: 0.5682 - val_loss: 0.7078 - val_accuracy: 0.5949\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 5s 472ms/step - loss: 1.0875 - accuracy: 0.5723 - val_loss: 0.6841 - val_accuracy: 0.4211\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2cd2d3c50>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6BNbZvSDMuj8",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with DenseNet121"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a8AJxh_OMtrR",
        "colab_type": "code",
        "outputId": "62547942-262a-4ac0-82b4-95507b40c555",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.densenet.DenseNet121(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/densenet/densenet121_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "29089792/29084464 [==============================] - 1s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 37s 3s/step - loss: 1.0073 - accuracy: 0.4620 - val_loss: 1.0263 - val_accuracy: 0.3856\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 576ms/step - loss: 0.9172 - accuracy: 0.4795 - val_loss: 0.8498 - val_accuracy: 0.4614\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 587ms/step - loss: 0.7732 - accuracy: 0.5625 - val_loss: 0.6734 - val_accuracy: 0.4580\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 6s 577ms/step - loss: 0.8038 - accuracy: 0.5723 - val_loss: 0.7757 - val_accuracy: 0.4902\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 6s 583ms/step - loss: 0.7646 - accuracy: 0.5966 - val_loss: 0.7713 - val_accuracy: 0.5263\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2abf71320>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7CGiKtXdM6xa",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning with DenseNet169"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3MTPa5EZM56D",
        "colab_type": "code",
        "outputId": "afa1ef0f-d53a-474d-e708-ab4981cae08e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.densenet.DenseNet169(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/densenet/densenet169_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "51879936/51877672 [==============================] - 2s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 55s 5s/step - loss: 0.7560 - accuracy: 0.5614 - val_loss: 0.7878 - val_accuracy: 0.4233\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 8s 697ms/step - loss: 0.7402 - accuracy: 0.5380 - val_loss: 0.7887 - val_accuracy: 0.4396\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 8s 687ms/step - loss: 0.7108 - accuracy: 0.5673 - val_loss: 0.8507 - val_accuracy: 0.4568\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 8s 705ms/step - loss: 0.6061 - accuracy: 0.6608 - val_loss: 0.8785 - val_accuracy: 0.4591\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 8s 707ms/step - loss: 0.6338 - accuracy: 0.6818 - val_loss: 0.8597 - val_accuracy: 0.4565\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa2a390b278>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BVC6C_DwNGnB",
        "colab_type": "text"
      },
      "source": [
        "#Fine tuning DenseNet201"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k5jYA89rNM_O",
        "colab_type": "code",
        "outputId": "9532b87a-6379-4cdc-9905-7e89d9b1c679",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.densenet.DenseNet201(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/keras-team/keras-applications/releases/download/densenet/densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "74842112/74836368 [==============================] - 3s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 68s 6s/step - loss: 0.7978 - accuracy: 0.4795 - val_loss: 0.7204 - val_accuracy: 0.4737\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 8s 740ms/step - loss: 0.7606 - accuracy: 0.5439 - val_loss: 0.7762 - val_accuracy: 0.4879\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 8s 756ms/step - loss: 0.7217 - accuracy: 0.5789 - val_loss: 0.7031 - val_accuracy: 0.5293\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 8s 736ms/step - loss: 0.6543 - accuracy: 0.6140 - val_loss: 0.6452 - val_accuracy: 0.6617\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 9s 779ms/step - loss: 0.6089 - accuracy: 0.6705 - val_loss: 0.6205 - val_accuracy: 0.6957\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa29847a9b0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IzSNA8euNSls",
        "colab_type": "text"
      },
      "source": [
        "#Fine tune NasNetLarge"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Yk-rmWBNRpP",
        "colab_type": "code",
        "outputId": "fee28769-839a-457a-ba71-e7b1cde9ae8f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.nasnet.NASNetLarge(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/titu1994/Keras-NASNet/releases/download/v1.2/NASNet-large-no-top.h5\n",
            "343613440/343610240 [==============================] - 10s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 110s 10s/step - loss: 0.8461 - accuracy: 0.5170 - val_loss: 0.6996 - val_accuracy: 0.5412\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 10s 899ms/step - loss: 0.7375 - accuracy: 0.5723 - val_loss: 0.7057 - val_accuracy: 0.5086\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 10s 876ms/step - loss: 0.8200 - accuracy: 0.4943 - val_loss: 0.7142 - val_accuracy: 0.5926\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 10s 873ms/step - loss: 0.8203 - accuracy: 0.4880 - val_loss: 0.6901 - val_accuracy: 0.5788\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 10s 875ms/step - loss: 0.7839 - accuracy: 0.5114 - val_loss: 0.7012 - val_accuracy: 0.5412\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa28b61fba8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-BBtmClwNe44",
        "colab_type": "text"
      },
      "source": [
        "#Fine tune NasNetMobile"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jbuqwNfDNjze",
        "colab_type": "code",
        "outputId": "49133778-9bf4-468e-c08a-7dc5e82082f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 377
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.nasnet.NASNetMobile(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Downloading data from https://github.com/titu1994/Keras-NASNet/releases/download/v1.2/NASNet-mobile-no-top.h5\n",
            "19996672/19993432 [==============================] - 1s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 71s 6s/step - loss: 1.6952 - accuracy: 0.5322 - val_loss: 2.1728 - val_accuracy: 0.4222\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 9s 774ms/step - loss: 1.4433 - accuracy: 0.5029 - val_loss: 1.6433 - val_accuracy: 0.4062\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 8s 753ms/step - loss: 1.1984 - accuracy: 0.5556 - val_loss: 1.4104 - val_accuracy: 0.4235\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 8s 759ms/step - loss: 1.3422 - accuracy: 0.4386 - val_loss: 1.1836 - val_accuracy: 0.4741\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 8s 752ms/step - loss: 1.0892 - accuracy: 0.5322 - val_loss: 1.2120 - val_accuracy: 0.5229\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa27774b2b0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kVKIPWtjNolu",
        "colab_type": "text"
      },
      "source": [
        "#Fine tune MobileNetV2"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QKj6I68PNtx9",
        "colab_type": "code",
        "outputId": "f4e34bab-4738-45bb-e12e-5c7df160a206",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 411
        }
      },
      "source": [
        "from keras import applications\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dropout, Flatten, Dense\n",
        "\n",
        "# dimensions of our images.\n",
        "img_width, img_height = 32, 32\n",
        "\n",
        "train_data_dir = '/content/train'\n",
        "validation_data_dir = '/content/validation'\n",
        "nb_train_samples = 189\n",
        "nb_validation_samples = 59\n",
        "epochs = 5\n",
        "batch_size = 16\n",
        "\n",
        "\n",
        "\n",
        "# prepare data augmentation configuration\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1. / 255,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1. / 255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "    train_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "    validation_data_dir,\n",
        "    target_size=(img_height, img_width),\n",
        "    batch_size=batch_size,\n",
        "    class_mode='binary')\n",
        "\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "# build the VGG16 network\n",
        "\n",
        "input_tensor = Input(shape=(32,32,3))\n",
        "base_model = applications.mobilenet_v2.MobileNetV2(weights='imagenet',include_top= False,input_tensor=input_tensor)\n",
        "top_model = Sequential()\n",
        "top_model.add(Flatten(input_shape=base_model.output_shape[1:]))\n",
        "top_model.add(Dense(256, activation='relu'))\n",
        "top_model.add(Dropout(0.5))\n",
        "top_model.add(Dense(1, activation='sigmoid'))\n",
        "model = Model(input= base_model.input, output= top_model(base_model.output))\n",
        "\n",
        "# note that it is necessary to start with a fully-trained\n",
        "# classifier, including the top classifier,\n",
        "# in order to successfully do fine-tuning\n",
        "\n",
        "# add the model on top of the convolutional base\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "# set the first 25 layers (up to the last conv block)\n",
        "# to non-trainable (weights will not be updated)\n",
        "for layer in model.layers[:25]:\n",
        "    layer.trainable = False\n",
        "\n",
        "# compile the model with a SGD/momentum optimizer\n",
        "# and a very slow learning rate.\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.SGD(lr=1e-4, momentum=0.9),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "# fine-tune the model\n",
        "model.fit_generator(\n",
        "    train_generator,\n",
        "    samples_per_epoch=nb_train_samples,\n",
        "    epochs=epochs,\n",
        "    validation_data=validation_generator,\n",
        "    nb_val_samples=nb_validation_samples)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n",
            "Found 187 images belonging to 2 classes.\n",
            "Found 59 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras_applications/mobilenet_v2.py:294: UserWarning: `input_shape` is undefined or non-square, or `rows` is not in [96, 128, 160, 192, 224]. Weights for input shape (224, 224) will be loaded as the default.\n",
            "  warnings.warn('`input_shape` is undefined or non-square, '\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://github.com/JonathanCMitchell/mobilenet_v2_keras/releases/download/v1.1/mobilenet_v2_weights_tf_dim_ordering_tf_kernels_1.0_224_no_top.h5\n",
            "9412608/9406464 [==============================] - 1s 0us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: Update your `Model` call to the Keras 2 API: `Model(inputs=Tensor(\"in..., outputs=Tensor(\"se...)`\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: The semantics of the Keras 2 argument `steps_per_epoch` is not the same as the Keras 1 argument `samples_per_epoch`. `steps_per_epoch` is the number of batches to draw from the generator at each epoch. Basically steps_per_epoch = samples_per_epoch/batch_size. Similarly `nb_val_samples`->`validation_steps` and `val_samples`->`steps` arguments have changed. Update your method calls accordingly.\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:91: UserWarning: Update your `fit_generator` call to the Keras 2 API: `fit_generator(<keras.pre..., epochs=5, validation_data=<keras.pre..., steps_per_epoch=11, validation_steps=59)`\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "11/11 [==============================] - 14s 1s/step - loss: 1.0535 - accuracy: 0.5029 - val_loss: 0.6855 - val_accuracy: 0.5584\n",
            "Epoch 2/5\n",
            "11/11 [==============================] - 6s 514ms/step - loss: 0.9819 - accuracy: 0.4854 - val_loss: 0.6880 - val_accuracy: 0.5259\n",
            "Epoch 3/5\n",
            "11/11 [==============================] - 6s 503ms/step - loss: 0.8783 - accuracy: 0.5795 - val_loss: 0.6903 - val_accuracy: 0.4568\n",
            "Epoch 4/5\n",
            "11/11 [==============================] - 6s 502ms/step - loss: 0.8197 - accuracy: 0.6023 - val_loss: 0.6960 - val_accuracy: 0.4741\n",
            "Epoch 5/5\n",
            "11/11 [==============================] - 6s 506ms/step - loss: 0.9408 - accuracy: 0.5361 - val_loss: 0.7049 - val_accuracy: 0.5057\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.callbacks.History at 0x7fa277c7e400>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pP66eSTRNzXv",
        "colab_type": "text"
      },
      "source": [
        "#Thanks for completing this notebook"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2V2BAb6wfEVj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}