Stock_Prediction_CNN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Stock_Prediction_CNN.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "k8bQtOidMp5Z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "d73d2ea3-99be-4193-f023-be0a73fc103c"
      },
      "source": [
        "! pip install quandl \n",
        "!pip install yfinance"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting quandl\n",
            "  Downloading https://files.pythonhosted.org/packages/12/64/6314547cb01fa75697ddba6d6d43a9f1d023c1730ffc6abf3db7a0c89222/Quandl-3.4.8-py2.py3-none-any.whl\n",
            "Collecting ndg-httpsclient\n",
            "  Downloading https://files.pythonhosted.org/packages/fb/67/c2f508c00ed2a6911541494504b7cac16fe0b0473912568df65fd1801132/ndg_httpsclient-0.5.1-py3-none-any.whl\n",
            "Requirement already satisfied: numpy>=1.8 in /usr/local/lib/python3.6/dist-packages (from quandl) (1.17.4)\n",
            "Requirement already satisfied: requests>=2.7.0 in /usr/local/lib/python3.6/dist-packages (from quandl) (2.21.0)\n",
            "Requirement already satisfied: pyasn1 in /usr/local/lib/python3.6/dist-packages (from quandl) (0.4.7)\n",
            "Collecting pyOpenSSL\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/9e/de/f8342b68fa9e981d348039954657bdf681b2ab93de27443be51865ffa310/pyOpenSSL-19.1.0-py2.py3-none-any.whl (53kB)\n",
            "\u001b[K     |████████████████████████████████| 61kB 9.8MB/s \n",
            "\u001b[?25hRequirement already satisfied: pandas>=0.14 in /usr/local/lib/python3.6/dist-packages (from quandl) (0.25.3)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from quandl) (1.12.0)\n",
            "Collecting inflection>=0.3.1\n",
            "  Downloading https://files.pythonhosted.org/packages/d5/35/a6eb45b4e2356fe688b21570864d4aa0d0a880ce387defe9c589112077f8/inflection-0.3.1.tar.gz\n",
            "Collecting more-itertools<=5.0.0\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/a4/a6/42f17d065bda1fac255db13afc94c93dbfb64393eae37c749b4cb0752fc7/more_itertools-5.0.0-py3-none-any.whl (52kB)\n",
            "\u001b[K     |████████████████████████████████| 61kB 9.9MB/s \n",
            "\u001b[?25hRequirement already satisfied: python-dateutil in /usr/local/lib/python3.6/dist-packages (from quandl) (2.6.1)\n",
            "Requirement already satisfied: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests>=2.7.0->quandl) (2.8)\n",
            "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests>=2.7.0->quandl) (3.0.4)\n",
            "Requirement already satisfied: urllib3<1.25,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests>=2.7.0->quandl) (1.24.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests>=2.7.0->quandl) (2019.9.11)\n",
            "Collecting cryptography>=2.8\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/ca/9a/7cece52c46546e214e10811b36b2da52ce1ea7fa203203a629b8dfadad53/cryptography-2.8-cp34-abi3-manylinux2010_x86_64.whl (2.3MB)\n",
            "\u001b[K     |████████████████████████████████| 2.3MB 14.2MB/s \n",
            "\u001b[?25hRequirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.14->quandl) (2018.9)\n",
            "Requirement already satisfied: cffi!=1.11.3,>=1.8 in /usr/local/lib/python3.6/dist-packages (from cryptography>=2.8->pyOpenSSL->quandl) (1.13.2)\n",
            "Requirement already satisfied: pycparser in /usr/local/lib/python3.6/dist-packages (from cffi!=1.11.3,>=1.8->cryptography>=2.8->pyOpenSSL->quandl) (2.19)\n",
            "Building wheels for collected packages: inflection\n",
            "  Building wheel for inflection (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for inflection: filename=inflection-0.3.1-cp36-none-any.whl size=6076 sha256=06b945bde63b629318d0ab17087b9d0d83406d3aada337e487d0fdbbc8fdbed8\n",
            "  Stored in directory: /root/.cache/pip/wheels/9f/5a/d3/6fc3bf6516d2a3eb7e18f9f28b472110b59325f3f258fe9211\n",
            "Successfully built inflection\n",
            "\u001b[31mERROR: datascience 0.10.6 has requirement folium==0.2.1, but you'll have folium 0.8.3 which is incompatible.\u001b[0m\n",
            "Installing collected packages: cryptography, pyOpenSSL, ndg-httpsclient, inflection, more-itertools, quandl\n",
            "  Found existing installation: more-itertools 7.2.0\n",
            "    Uninstalling more-itertools-7.2.0:\n",
            "      Successfully uninstalled more-itertools-7.2.0\n",
            "Successfully installed cryptography-2.8 inflection-0.3.1 more-itertools-5.0.0 ndg-httpsclient-0.5.1 pyOpenSSL-19.1.0 quandl-3.4.8\n",
            "Collecting yfinance\n",
            "  Downloading https://files.pythonhosted.org/packages/85/56/5e26c150517acbdce8e31d3e6169530e102a0575f5eb1a4310753e53a424/yfinance-0.1.50.tar.gz\n",
            "Requirement already satisfied: pandas>=0.24 in /usr/local/lib/python3.6/dist-packages (from yfinance) (0.25.3)\n",
            "Requirement already satisfied: numpy>=1.15 in /usr/local/lib/python3.6/dist-packages (from yfinance) (1.17.4)\n",
            "Requirement already satisfied: requests>=2.20 in /usr/local/lib/python3.6/dist-packages (from yfinance) (2.21.0)\n",
            "Requirement already satisfied: multitasking>=0.0.7 in /usr/local/lib/python3.6/dist-packages (from yfinance) (0.0.9)\n",
            "Requirement already satisfied: python-dateutil>=2.6.1 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.24->yfinance) (2.6.1)\n",
            "Requirement already satisfied: pytz>=2017.2 in /usr/local/lib/python3.6/dist-packages (from pandas>=0.24->yfinance) (2018.9)\n",
            "Requirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->yfinance) (3.0.4)\n",
            "Requirement already satisfied: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->yfinance) (2.8)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->yfinance) (2019.9.11)\n",
            "Requirement already satisfied: urllib3<1.25,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests>=2.20->yfinance) (1.24.3)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.6/dist-packages (from python-dateutil>=2.6.1->pandas>=0.24->yfinance) (1.12.0)\n",
            "Building wheels for collected packages: yfinance\n",
            "  Building wheel for yfinance (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for yfinance: filename=yfinance-0.1.50-py2.py3-none-any.whl size=21556 sha256=d87d7215d2e9eede8bedd690462302a7a5ab0467f123fee30feaaf754fb57ec8\n",
            "  Stored in directory: /root/.cache/pip/wheels/19/2a/e6/b4d4cc977576e18220a36866acf2f2cf7241f16af5f8795b17\n",
            "Successfully built yfinance\n",
            "Installing collected packages: yfinance\n",
            "Successfully installed yfinance-0.1.50\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "81GuSj_gM223",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "e6fb350e-4b65-4084-8c22-108cf5959779"
      },
      "source": [
        "import yfinance as yf  \n",
        "import time\n",
        "import quandl \n",
        "import datetime as dt\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "from tqdm import tqdm\n",
        "import matplotlib.pylab as plt\n",
        "import glob, os\n",
        "from fastai.vision import *\n",
        "from fastai.metrics import accuracy\n",
        "import datetime\n",
        "%matplotlib inline\n",
        "from matplotlib.pylab import rcParams\n",
        "rcParams['figure.figsize'] = 15, 6\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "import torch\n",
        "torch.cuda.get_device_name(0)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'Tesla P100-PCIE-16GB'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gru6JrWMM813",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def plot_candles(pricing = None, title=None,\n",
        "                 volume_bars=False,\n",
        "                 color_function=None,\n",
        "                 overlays=None,\n",
        "                 technicals=None,\n",
        "                 technicals_titles=None,\n",
        "                 save =False,\n",
        "                 name = None,\n",
        "                 multi_stocks = None,\n",
        "                 Bollinger_Band = None,\n",
        "                 start=\"2016-01-01\",\n",
        "                 end=dt.datetime.now()):\n",
        "    \"\"\" \n",
        "    The orginal code for this code was obtained\n",
        "    from https://www.quantopian.com/posts/plot-candlestick-charts-in-research \n",
        "    and was modified. \n",
        "    \n",
        "    \n",
        "    Plots a candlestick chart using quantopian pricing data.\n",
        "    \n",
        "    \n",
        "    \n",
        "    Args:\n",
        "      pricing: A pandas dataframe with columns ['open_price', 'close_price', 'high', 'low', 'volume']\n",
        "      title: An optional title for the chart\n",
        "      volume_bars: If True, plots volume bars\n",
        "      color_function: A function which, given a row index and price series, returns a candle color.\n",
        "      overlays: A list of additional data series to overlay on top of pricing.  Must be the same length as pricing.\n",
        "      technicals: A list of additional data series to display as subplots.\n",
        "      technicals_titles: A list of titles to display for each technical indicator.\n",
        "    \"\"\"\n",
        "    def default_color(index, open_price, close_price, low, high):\n",
        "        return 'r' if open_price[index] > close_price[index] else 'g'\n",
        "    color_function = color_function or default_color\n",
        "    overlays = overlays or []\n",
        "    technicals = technicals or []\n",
        "    technicals_titles = technicals_titles or []\n",
        "\n",
        "    \n",
        "    subplot_count = 1\n",
        "    if volume_bars:\n",
        "        subplot_count = 2\n",
        "    if technicals:\n",
        "        subplot_count += len(technicals)\n",
        "    \n",
        "    if multi_stocks: \n",
        "        subplot_count = len(multi_stocks)*2\n",
        "        \n",
        "    if subplot_count == 1:\n",
        "        fig, ax1 = plt.subplots(1, 1)\n",
        "    \n",
        "\n",
        "    else:\n",
        "        ratios = np.insert(np.full(subplot_count - 1, 1), 0, 3)\n",
        "        fig, subplots = plt.subplots(subplot_count, 1, sharex=True, gridspec_kw={'height_ratios': ratios})\n",
        "        \n",
        "        \n",
        "    \n",
        "    \n",
        "    for i, ticker in enumerate(multi_stocks):\n",
        "      # get the data 20 day before start to calculate rolling mean and std and then clip it \n",
        "        \n",
        "        pricing = yf.download(ticker,start- datetime.timedelta(days=40),end)\n",
        "        ax1 = subplots[i*2]\n",
        "   \n",
        "        pricing['30 Day MA'] = pricing['Adj Close'].rolling(window=20).mean()\n",
        "        pricing['30 Day STD'] = pricing['Adj Close'].rolling(window=20).std()\n",
        "        pricing['Upper Band'] = pricing['30 Day MA'] + (pricing['30 Day STD'] * 2)\n",
        "        pricing['Lower Band'] = pricing['30 Day MA'] - (pricing['30 Day STD'] * 2)\n",
        "         # clip the start -20 to start\n",
        "        pricing.reset_index(inplace =True)\n",
        "        pricing = pricing[pricing['Date']>=start_date]\n",
        "        pricing.set_index('Date',inplace = True)      \n",
        "        open_price = pricing['Open']\n",
        "        close_price = pricing['Close']\n",
        "        low = pricing['Low']\n",
        "        high = pricing['High']\n",
        "        \n",
        "        oc_min = pd.concat([open_price, close_price], axis=1).min(axis=1)\n",
        "        oc_max = pd.concat([open_price, close_price], axis=1).max(axis=1)\n",
        "        x = np.arange(len(pricing))\n",
        "        candle_colors = [color_function(i, open_price, close_price, low, high) for i in x]\n",
        "        candles = ax1.bar(x, oc_max-oc_min, bottom=oc_min, color=candle_colors, linewidth=0)\n",
        "        lines = ax1.vlines(x + 0.4, low, high, color=candle_colors, linewidth=1)\n",
        "        ax1.xaxis.grid(False)\n",
        "        ax1.xaxis.set_tick_params(which='major', length=3.0, direction='in', top='off')\n",
        "        ax1.plot(pricing['Close'].values,color = 'b')\n",
        "        # Assume minute frequency if first two bars are in the same day.\n",
        "        frequency = 'minute' if (pricing.index[1] - pricing.index[0]).days == 0 else 'day'\n",
        "        time_format = '%d-%m-%Y'\n",
        "        if frequency == 'minute':\n",
        "            time_format = '%H:%M'\n",
        "\n",
        "        for overlay in overlays:\n",
        "            ax1.plot(x, overlay)\n",
        "        # Plot volume bars if needed\n",
        "        if volume_bars:\n",
        "            ax2 = subplots[i*2+1]\n",
        "            volume = pricing['Volume']\n",
        "            volume_scale = None\n",
        "            scaled_volume = volume\n",
        "            if volume.max() > 1000000:\n",
        "                volume_scale = 'M'\n",
        "                scaled_volume = volume / 1000000\n",
        "            elif volume.max() > 1000:\n",
        "                volume_scale = 'K'\n",
        "                scaled_volume = volume / 1000\n",
        "            ax2.bar(x, scaled_volume, color=candle_colors)\n",
        "            ax2.xaxis.grid(False)\n",
        "        \n",
        "        x_axis = pricing.reset_index().reset_index()['index'] \n",
        "\n",
        "        ax1.fill_between(x_axis, pricing['Upper Band'], pricing['Lower Band'], color='grey')\n",
        "\n",
        "        # Plot additional technical indicators\n",
        "        for (i, technical) in enumerate(technicals):\n",
        "            ax = subplots[i - len(technicals)] # Technical indicator plots are shown last\n",
        "            ax.plot(x, technical)\n",
        "            if i < len(technicals_titles):\n",
        "                ax.set_title(technicals_titles[i])\n",
        "                \n",
        "                \n",
        "    if save: plt.savefig('/content/drive/My Drive/Stocks/'+name) "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LOgM08HuN_EN",
        "colab_type": "text"
      },
      "source": [
        "Load Google Drive"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-A4S95NwNp9F",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 129
        },
        "outputId": "6213df53-5cbf-4880-e7a5-b3d4d904850c"
      },
      "source": [
        "# Load the Drive helper and mount\n",
        "from google.colab import drive\n",
        "# This will prompt for authorization.\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z49CAdjdOdsF",
        "colab_type": "text"
      },
      "source": [
        "Plot AMZN and MSFT"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rAglpH9DNyzl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 411
        },
        "outputId": "773aedc1-c60c-4446-9848-4e729836b90a"
      },
      "source": [
        "windows = 30 \n",
        "start_date = \"2016-01-01\"\n",
        "start_date = pd.to_datetime(start_date)\n",
        "end_date = dt.datetime.now()\n",
        "plot_candles(  volume_bars=True, save = True,name = 'amazon.png', multi_stocks=['AMZN','MSFT'],Bollinger_Band = True, start = start_date, end = end_date )\n"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[*********************100%***********************]  1 of 1 completed\n",
            "[*********************100%***********************]  1 of 1 completed\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3MAAAFlCAYAAABBWbHVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 1080x432 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jIwNHJ5EOO0L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Uncomment this block if you want to delete and create a new folder\n",
        "# ! rm -r \"/content/drive/My Drive/Stocks/\"\n",
        "# ! /bin/mkdir -p \"/content/drive/My Drive/Stocks\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m44QvCvxO26j",
        "colab_type": "text"
      },
      "source": [
        "# Training the model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cc97afrjRfjB",
        "colab_type": "text"
      },
      "source": [
        "Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OBry1wn1RcIX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "# stickers = ['AMZN','MSFT','WMT','NFLX']\n",
        "stickers = ['AMZN']\n",
        "windows = 30 \n",
        "start_date = \"2012-01-01\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P4ySkudgO4-i",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "b9cadfbb-873c-4f99-e8a1-227e30a8764a"
      },
      "source": [
        "start_date = pd.to_datetime(start_date)\n",
        "Final_date = dt.datetime.now() \n",
        "#  some days stock market is closed so I only get data for the day that \n",
        "#  we have data for\n",
        "data = yf.download('AMZN',start_date,Final_date)\n",
        "# Label \n",
        "data['profit']= (data.Close.shift(-1) - data.Close)/data.Close.shift(-1)\n",
        "data.dropna(how = 'any',inplace= True)\n",
        "data['profit']= data['profit'] >= 0.02\n",
        "data['profit']= data['profit'].astype('int32')\n",
        "\n",
        "label= data.reset_index ()\n",
        "label= label[['profit','Date']]\n",
        "# Clean the data\n",
        "\n",
        "train_dir = 'content/drive/My Drive/Stocks/'\n",
        "\n",
        "label = label.shift(-(windows-1))\n",
        "label.dropna(how = 'any',inplace= True)\n",
        "label['profit']= label['profit'].astype('int32')\n",
        "label.drop(label.index[-1], inplace = True)\n",
        "label['path'] = label['Date'].map(lambda x: os.path.join(train_dir,'{}.png'.format(x)))\n",
        "label['Date_str']=label['Date'].astype('str')\n",
        "\n"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\r[*********************100%***********************]  1 of 1 completed\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PSL5nqgzQgtE",
        "colab_type": "text"
      },
      "source": [
        "## Generate Pictures from charts \n",
        "Get the date from dataframe "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mA9SUxxkPr4D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        },
        "outputId": "f833a316-03fe-4c42-d01d-6ac88bab88a2"
      },
      "source": [
        "start_date = label.iloc[0].Date\n",
        "os.chdir(\"/content/drive/My Drive/Stocks\")\n",
        "list_file = []\n",
        "for file in glob.glob(\"*.png\"):\n",
        "    list_file.append(file)\n",
        "\n",
        "   \n",
        "for ii in tqdm(range(label.shape[0])):\n",
        "    plt.clf()\n",
        "    end_date   = label.iloc[ii].Date \n",
        "    start_date = pd.to_datetime(label.iloc[ii].Date) - datetime.timedelta(days=windows)\n",
        "    if label.iloc[ii].Date_str+'.png' not in list_file:\n",
        "       print(\"file name:   \" ,label.iloc[ii].Date_str)\n",
        "       plot_candles(  volume_bars=True, save = True, Bollinger_Band = True, name =label.iloc[ii].Date_str, multi_stocks= stickers , start = start_date, end = end_date )\n",
        "    ii += 1\n",
        "\n",
        "label['Date'] = label['Date'].astype('str')\n",
        "label['path'] = label['Date'].map(lambda x: os.path.join(train_dir,'{}.png'.format(x)))\n",
        "\n",
        "os.chdir(\"../../../..\")"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 1953/1953 [00:01<00:00, 1614.69it/s]\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x432 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1vDy7GN_PDEy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Data Augmentation and Transformation up to Data Bunch\n",
        "tfms = get_transforms()\n",
        "sz= 600 # 600 ,20 good \n",
        "bs =10\n",
        "\n",
        "src = (ImageList.from_df(df=label,path =\"..\",cols='path') \n",
        "        .split_by_idx(valid_idx=range(int(label.shape[0] - label.shape[0]*0.3),label.shape[0]))\n",
        "        .label_from_df(cols='profit') \n",
        "      )\n",
        "data= (src.databunch(bs=bs,num_workers=4) \n",
        "        .normalize(imagenet_stats)      \n",
        "       )"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bobd_4KkSH3n",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 112
        },
        "outputId": "f6624bad-a783-4394-cd79-588713490a3a"
      },
      "source": [
        "manualSeed = 100\n",
        "np.random.seed(manualSeed)\n",
        "torch.manual_seed(manualSeed)\n",
        "torch.cuda.manual_seed_all(manualSeed)\n",
        "learn = cnn_learner(data, models.resnet50)\n",
        "# Let's fit a couple of cycles\n",
        "learn.fit_one_cycle(2)\n"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\n",
              "      <th>epoch</th>\n",
              "      <th>train_loss</th>\n",
              "      <th>valid_loss</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.755021</td>\n",
              "      <td>0.507046</td>\n",
              "      <td>00:59</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1</td>\n",
              "      <td>0.373679</td>\n",
              "      <td>0.331444</td>\n",
              "      <td>00:59</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3N9vN0pJSWSo",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 446
        },
        "outputId": "e1981fa0-a601-43e3-d8a8-1a5ed67c5d27"
      },
      "source": [
        "learn.unfreeze()\n",
        "learn.lr_find()\n",
        "learn.recorder.plot(suggestion=True)\n"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              ""
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n",
            "Min numerical gradient: 1.10E-06\n",
            "Min loss divided by 10: 1.74E-06\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4sAAAF2CAYAAADHgA2NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 1080x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UUmxGmL6SZ8Z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 127
        },
        "outputId": "d2e64a88-f092-43c9-c0c6-d42562f23f82"
      },
      "source": [
        "# Let's fit some more with a more accurate learning rate\n",
        "learn.fit_one_cycle(8, max_lr=slice(1.10E-06))\n",
        "learn.recorder.plot_losses()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='1' class='' max='8', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      12.50% [1/8 00:59<06:54]\n",
              "    </div>\n",
              "    \n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: left;\">\n",
              "      <th>epoch</th>\n",
              "      <th>train_loss</th>\n",
              "      <th>valid_loss</th>\n",
              "      <th>time</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>0</td>\n",
              "      <td>0.234420</td>\n",
              "      <td>0.338849</td>\n",
              "      <td>00:59</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table><p>\n",
              "\n",
              "    <div>\n",
              "        <style>\n",
              "            /* Turns off some styling */\n",
              "            progress {\n",
              "                /* gets rid of default border in Firefox and Opera. */\n",
              "                border: none;\n",
              "                /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
              "                background-size: auto;\n",
              "            }\n",
              "            .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
              "                background: #F44336;\n",
              "            }\n",
              "        </style>\n",
              "      <progress value='60' class='' max='136', style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      44.12% [60/136 00:21<00:27 0.2496]\n",
              "    </div>\n",
              "    "
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pswtmK5ISmdG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "preds,y, loss = learn.get_preds(with_loss=True)\n",
        "# get accuracy\n",
        "acc = accuracy(preds, y)\n",
        "print('The accuracy is {0} %.'.format(acc))\n",
        "\n",
        "# probs from log preds\n",
        "probs = np.exp(preds[:,1])\n",
        "# Compute ROC curve\n",
        "fpr, tpr, thresholds = roc_curve(y, probs, pos_label=1)\n",
        "\n",
        "# Compute ROC area\n",
        "roc_auc = auc(fpr, tpr)\n",
        "print('ROC area is {0}'.format(roc_auc))\n",
        "\n",
        "plt.figure()\n",
        "plt.plot(fpr, tpr, color='darkorange', label='ROC curve (area = %0.2f)' % roc_auc)\n",
        "plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
        "plt.xlim([-0.01, 1.0])\n",
        "plt.ylim([0.0, 1.01])\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive Rate')\n",
        "plt.title('Receiver operating characteristic')\n",
        "plt.legend(loc=\"lower right\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BYxtIJviWOoH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "f00b1aef-5962-46f3-bcce-1da9505f40a3"
      },
      "source": [
        "learn.summary()"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Sequential\n",
              "======================================================================\n",
              "Layer (type)         Output Shape         Param #    Trainable \n",
              "======================================================================\n",
              "Conv2d               [64, 216, 540]       9,408      True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 216, 540]       128        True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [64, 216, 540]       0          False     \n",
              "______________________________________________________________________\n",
              "MaxPool2d            [64, 108, 270]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       4,096      True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       36,864     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 108, 270]      16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 108, 270]      512        True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [256, 108, 270]      0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 108, 270]      16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 108, 270]      512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       36,864     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 108, 270]      16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 108, 270]      512        True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [256, 108, 270]      0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [64, 108, 270]       36,864     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [64, 108, 270]       128        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 108, 270]      16,384     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 108, 270]      512        True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [256, 108, 270]      0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 108, 270]      32,768     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 108, 270]      256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       147,456    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 54, 135]       1,024      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [512, 54, 135]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 54, 135]       131,072    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 54, 135]       1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       147,456    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 54, 135]       1,024      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [512, 54, 135]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       147,456    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 54, 135]       1,024      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [512, 54, 135]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [128, 54, 135]       147,456    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [128, 54, 135]       256        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 54, 135]       65,536     True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 54, 135]       1,024      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [512, 54, 135]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 54, 135]       131,072    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 54, 135]       512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       524,288    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [256, 27, 68]        589,824    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [256, 27, 68]        512        True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [1024, 27, 68]       262,144    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [1024, 27, 68]       2,048      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [1024, 27, 68]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 27, 68]        524,288    True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 27, 68]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 14, 34]        2,359,296  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 14, 34]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [2048, 14, 34]       1,048,576  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [2048, 14, 34]       4,096      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [2048, 14, 34]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [2048, 14, 34]       2,097,152  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [2048, 14, 34]       4,096      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 14, 34]        1,048,576  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 14, 34]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 14, 34]        2,359,296  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 14, 34]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [2048, 14, 34]       1,048,576  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [2048, 14, 34]       4,096      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [2048, 14, 34]       0          False     \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 14, 34]        1,048,576  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 14, 34]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [512, 14, 34]        2,359,296  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [512, 14, 34]        1,024      True      \n",
              "______________________________________________________________________\n",
              "Conv2d               [2048, 14, 34]       1,048,576  True      \n",
              "______________________________________________________________________\n",
              "BatchNorm2d          [2048, 14, 34]       4,096      True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [2048, 14, 34]       0          False     \n",
              "______________________________________________________________________\n",
              "AdaptiveAvgPool2d    [2048, 1, 1]         0          False     \n",
              "______________________________________________________________________\n",
              "AdaptiveMaxPool2d    [2048, 1, 1]         0          False     \n",
              "______________________________________________________________________\n",
              "Flatten              [4096]               0          False     \n",
              "______________________________________________________________________\n",
              "BatchNorm1d          [4096]               8,192      True      \n",
              "______________________________________________________________________\n",
              "Dropout              [4096]               0          False     \n",
              "______________________________________________________________________\n",
              "Linear               [512]                2,097,664  True      \n",
              "______________________________________________________________________\n",
              "ReLU                 [512]                0          False     \n",
              "______________________________________________________________________\n",
              "BatchNorm1d          [512]                1,024      True      \n",
              "______________________________________________________________________\n",
              "Dropout              [512]                0          False     \n",
              "______________________________________________________________________\n",
              "Linear               [2]                  1,026      True      \n",
              "______________________________________________________________________\n",
              "\n",
              "Total params: 25,615,938\n",
              "Total trainable params: 25,615,938\n",
              "Total non-trainable params: 0\n",
              "Optimized with 'torch.optim.adam.Adam', betas=(0.9, 0.99)\n",
              "Using true weight decay as discussed in https://www.fast.ai/2018/07/02/adam-weight-decay/ \n",
              "Loss function : FlattenedLoss\n",
              "======================================================================\n",
              "Callbacks functions applied "
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1rGn3NsQXnOy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}