Deploying CV Classification Model

scratchpad.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Deploying CV Classification Model",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/jfthuong/a1364ab40189cb1b8d8add3095bb7ebf/scratchpad.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xUQiA23ju0J1"
      },
      "source": [
        "# Deploying a trained model for CV Classification using Streamlit"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n6I5aGgDu5aK"
      },
      "source": [
        "# Upgrade ipykernel"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FZ7NYvDms-Sh"
      },
      "source": [
        "Note: it seems we need recent version of ipykernel to run streamlit."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "anHqkYesoV6L"
      },
      "source": [
        "!pip install -q ipykernel>=5.1.2\n",
        "!pip install -q pydeck"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-P_5g2iSu7h6"
      },
      "source": [
        "... after that, you need to reset the Runtime."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JL6cHcF_vE48"
      },
      "source": [
        "## Install dependencies"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lIYdn1woOS1n"
      },
      "source": [
        "!pip install -q unpackai[deploy]\n",
        "!pip install -qU fastai"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qCSSIDFHueCT"
      },
      "source": [
        "Checking versions installed"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xYBDSRPMuhGt",
        "outputId": "b358dd21-ce87-4fcb-ae15-722d41fe06f9"
      },
      "source": [
        "!pip list | grep \"unpackai\"\n",
        "!pip list | grep \"fastai\""
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "unpackai                      0.1.8.16\n",
            "fastai                        2.5.3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g8ud5FKFjRsU"
      },
      "source": [
        "from unpackai.deploy import deploy_app, StreamlitAppCVClassif"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uWPwd374jdVE",
        "outputId": "8f49bdd0-ab2d-48ac-8d32-ed3e21ebdb8f"
      },
      "source": [
        "app = StreamlitAppCVClassif().render(\n",
        "    title=\"My Mini App\",\n",
        "    author=\"Jeff\",\n",
        "    model=\"/content/model.pkl\",\n",
        "    implem_4_model=\"is_cat = dummy_function\",\n",
        ")\n",
        "app.save(\"app.py\")"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Saved app 'My Mini App' to 'app.py'\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uH7sQ7VvtQZY"
      },
      "source": [
        "**WARNING**\n",
        "\n",
        "Currently, there is a bug, you need to edit the file `app.py` \n",
        "1. Double click on it from file editor\n",
        "2. Replace all `from .deploy import ...` by `from unpackai.deploy import ...`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P4Syo6gjtKRi"
      },
      "source": [
        "NOTE: You can also customize your app.\n",
        "\n",
        "All the Streamlit widgets are listed here:\n",
        "https://docs.streamlit.io/library/api-reference#image"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0FG6UpbsuzGN"
      },
      "source": [
        "## Deploy\n",
        "\n",
        "1. Drag & Drop your model in the list of files\n",
        "2. Run the cell below\n",
        "\n",
        "Notes:\n",
        "* if your model is not named *model.pkl*, adjust the path in the cell that creates the app (or in \"app.py\")\n",
        "* You might have to run the cell below several times if you see \"Connection Error\" "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7-B1h7gOkaAP",
        "outputId": "0c7a5161-911f-4cae-eebe-a573ce15a1c1"
      },
      "source": [
        "deploy_app()\n",
        "!nohup streamlit run app.py"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "  1. Create a new cell with: !nohup streamlit run app.py\n",
            "  2. Run that cell\n",
            "  3. Click on this link: https://c59c-35-197-72-180.ngrok.io\n",
            "\n",
            "Note: this will link to local address http://localhost:8501\n",
            "\n",
            "nohup: ignoring input and appending output to 'nohup.out'\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lz6SIF9kpdvN"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}