classfication.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "2f6ddc86-2410-4f92-b718-9fc526892df3",
   "metadata": {},
   "source": [
    "# Import Section\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3a6655ef-d4d1-4f1b-aa15-a82358c0cd4d",
   "metadata": {
    "jupyter": {
     "source_hidden": true
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "This script sets up the necessary imports for a Jupyter Notebook that involves\n",
    "image classification using TensorFlow 2. \n",
    "\"\"\"\n",
    "import shutil\n",
    "import os\n",
    "\n",
    "import ipywidgets as widgets\n",
    "from ipywidgets import interact_manual\n",
    "from ipywidgets import Layout, Box, Dropdown, Label\n",
    "from IPython.display import display, HTML\n",
    "from IPython.display import clear_output\n",
    "from ipyfilechooser import FileChooser"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "47fdfc08-2ec2-446f-9d5b-9dcf05dbf869",
   "metadata": {},
   "source": [
    "# Widgets Control Section\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "ce8b6cd0-2e45-4997-bc77-4aa67ffcba62",
   "metadata": {
    "jupyter": {
     "source_hidden": true
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "class TrainConfigAndCmdsWidgets:\n",
    "    \"\"\"\n",
    "    A class to create and manage widgets for configuring and running training, testing, and model conversion tasks.\n",
    "    Attributes:\n",
    "    \"\"\"\n",
    "    def __init__(self):\n",
    "\n",
    "        self.tflite_file_loc = \"\"\n",
    "\n",
    "        form_item_layout = Layout(\n",
    "            display=\"flex\",\n",
    "            flex_flow=\"row\",\n",
    "            justify_content=\"space-between\",\n",
    "        )\n",
    "\n",
    "        # data exist\n",
    "        self.a_de = widgets.Checkbox(value=True, disabled=False, indent=False)\n",
    "        self.b_de = widgets.Text(value=\"flower_photos\", placeholder=\"Type something\", disabled=False)\n",
    "\n",
    "        form_data_prepare_items = [\n",
    "            Box([Label(value=\"Data Exist\"), self.a_de], layout=form_item_layout),\n",
    "            Box([Label(value=\"Dataset Name\"), self.b_de], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        self.form_data_prepare_exist = Box(\n",
    "            form_data_prepare_items,\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"100%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "        # data download\n",
    "        # Another flowers dataset\n",
    "        # https://storage.googleapis.com/mledu-datasets/cats_and_dogs_filtered.zip\n",
    "        self.a_dp = widgets.Textarea(value=\"https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz\", placeholder=\"Type something\", disabled=False)\n",
    "        # cats_and_dogs.zip\n",
    "        self.b_dp = widgets.Text(value=\"flower_photos.tgz\", placeholder=\"Type something\", disabled=False)\n",
    "\n",
    "        form_data_prepare_items = [\n",
    "            Box([Label(value=\"URL Link\"), self.a_dp], layout=form_item_layout),\n",
    "            Box([Label(value=\"Zip Name\"), self.b_dp], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        self.form_data_prepare_cmd = Box(\n",
    "            form_data_prepare_items,\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                border=\"solid 1px lightblue\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"100%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "        # train\n",
    "        self.a_ta = widgets.Text(value=\"MyTask\", placeholder=\"Type something\", disabled=False)\n",
    "        self.e_ta = Dropdown(\n",
    "            options=[\n",
    "                \"mobilenet_v1\",\n",
    "                \"mobilenet_v2\",\n",
    "                \"mobilenet_v3_mini\",\n",
    "                \"mobilenet_v3\",\n",
    "                \"fdmobilenet_w1\",\n",
    "                \"fdmobilenet_wd2\",\n",
    "                \"fdmobilenet_wd4\",\n",
    "                \"shufflenet_g1_wd4\",\n",
    "                \"shufflenet_g3_wd4\",\n",
    "                \"shufflenet_g1_wd2\",\n",
    "                \"shufflenet_g3_wd2\",\n",
    "                \"efficientnetB0\",\n",
    "                \"efficientnetv2B0\",\n",
    "            ]\n",
    "        )\n",
    "        self.o_ta = Dropdown(options=[\"1: Load Pretrain Model(Not support all setting)\", \"2: Download Pretrain Model Only\", \"3: Train From Scratch\"])\n",
    "        self.b_ta = widgets.IntSlider(value=32, min=4, max=192, step=4)\n",
    "        self.c_ta = widgets.IntSlider(value=224, min=32, max=352, step=32)\n",
    "        self.m_ta = widgets.FloatSlider(value=0.35, min=0.1, max=1.5, step=0.05)\n",
    "        self.d_ta = widgets.FloatSlider(value=0.2, min=0.1, max=0.5, step=0.1)\n",
    "        self.f_ta = widgets.FloatSlider(value=0.2, min=0.0, max=0.5, step=0.1)\n",
    "        self.g_ta = widgets.Checkbox(value=True, disabled=False, indent=False)\n",
    "        self.h_ta = widgets.Text(value=\"10,50\", placeholder=\"Type something\", disabled=False)\n",
    "        self.i_ta = widgets.Text(value=\"0.001,0.0005\", placeholder=\"Type something\", disabled=False)\n",
    "        self.j_ta = Dropdown(value=\"3: Transfer and Fine-Tuning training\", options=[\"1: Show the train data and model\", \"2: Transfer training\", \"3: Transfer and Fine-Tuning training\"])\n",
    "        self.k_ta = widgets.BoundedIntText(value=40, min=0, max=300, step=1, disabled=False)\n",
    "        self.l_ta = widgets.Button(description=\"Run\", layout=Layout(width=\"30%\", height=\"30px\"), button_style=\"success\")\n",
    "\n",
    "        self.fine_tune_la = Label(value=\"Freezing Layers of Fine-Tuning\")\n",
    "        self.k_ta_box = Box([self.fine_tune_la, self.k_ta], layout=form_item_layout)\n",
    "        self.alpha_width_la = Label(value=\"Alpha Width\")\n",
    "        self.m_ta_box = Box([self.alpha_width_la, self.m_ta], layout=form_item_layout)\n",
    "\n",
    "        form_train_items = [\n",
    "            Box([Label(value=\"Project Name\"), self.a_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Model Name\"), self.e_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Model Setting\"), self.o_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Batch Size\"), self.b_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Image Size\"), self.c_ta], layout=form_item_layout),\n",
    "            self.m_ta_box,\n",
    "            Box([Label(value=\"Validation Percent\"), self.d_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Test Percent from Val Dataset\"), self.f_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Data Augmentation Enable\"), self.g_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Epochs (Transfer, Fine-Tuning)\"), self.h_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Learning Rate (Transfer, Fine-Tuning)\"), self.i_ta], layout=form_item_layout),\n",
    "            Box([Label(value=\"Switch Mode\"), self.j_ta], layout=form_item_layout),\n",
    "            self.k_ta_box,\n",
    "            Box([Label(value=\"Start to Execute\"), self.l_ta], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        self.form_output_train_cmd = Box(\n",
    "            form_train_items,\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                border=\"solid 1px lightblue\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"100%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "        # test\n",
    "        self.a_tt = widgets.Button(description=\"Setting\", layout=Layout(width=\"30%\", height=\"30px\"), button_style=\"success\")\n",
    "        self.b_tt = widgets.BoundedIntText(value=1, min=1, max=100, step=1, disabled=False)\n",
    "        self.c_tt = widgets.Button(description=\"Run\", layout=Layout(width=\"30%\", height=\"30px\"), button_style=\"success\")\n",
    "\n",
    "        form_test_items = [\n",
    "            Box([Label(value=\"Choose the tflite file\"), self.a_tt], layout=form_item_layout),\n",
    "            Box([Label(value=\"Batches for test\"), self.b_tt], layout=form_item_layout),\n",
    "            Box([Label(value=\"Start to Test\"), self.c_tt], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        self.form_output_test_cmd = Box(\n",
    "            form_test_items,\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                border=\"solid 3px lightgreen\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"50%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "        # convert model cpp\n",
    "        self.a_cm = widgets.Text(value=r\"..\\workspace\\\\catsdogs\\\\tflite_model\", placeholder=\"Type something\", disabled=False)\n",
    "        self.b_cm = widgets.Text(value=\"mobilenet_v2_int8quant.tflite\", placeholder=\"Type something\", disabled=False)\n",
    "        self.c_cm = widgets.Text(value=r\"..\\workspace\\\\catsdogs\\\\tflite_model\\\\vela\", placeholder=\"Type something\", disabled=False)\n",
    "        self.e_cm = widgets.Button(description=\"Setting\", layout=Layout(width=\"30%\", height=\"30px\"), button_style=\"success\")\n",
    "        self.d_cm = widgets.Button(description=\"Run\", layout=Layout(width=\"30%\", height=\"30px\"), button_style=\"success\")\n",
    "        self.f_cm = widgets.Text(value=\"cats_and_dogs_filtered\", placeholder=\"Type something\", disabled=False)\n",
    "        self.g_cm = widgets.Checkbox(value=True, disabled=False, indent=False)\n",
    "\n",
    "        form_convert_items_paths = [\n",
    "            Box([Label(value=\"Choose the tflite file\"), self.e_cm], layout=form_item_layout),\n",
    "            Box([Label(value=\"MODEL SRC DIR\"), self.a_cm], layout=form_item_layout),\n",
    "            Box([Label(value=\"MODEL SRC FILE\"), self.b_cm], layout=form_item_layout),\n",
    "            Box([Label(value=\"GEN SRC DIR\"), self.c_cm], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        form_convert_items_label = [\n",
    "            Box([Label(value=\"Dataset Name\"), self.f_cm], layout=form_item_layout),\n",
    "            Box([Label(value=\"Labels Including\"), self.g_cm], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        form_convert_items = [\n",
    "            Box(form_convert_items_paths, layout=Layout(display=\"flex\", flex_flow=\"column\", justify_content=\"center\", border=\"dotted 3px lightblue\", align_items=\"stretch\", width=\"70%\")),\n",
    "            Box(form_convert_items_label, layout=Layout(display=\"flex\", flex_flow=\"column\", justify_content=\"center\", border=\"dotted 3px lightblue\", align_items=\"stretch\", width=\"70%\")),\n",
    "            Box([Label(value=\"Convert to cpp & Vela\"), self.d_cm], layout=form_item_layout),\n",
    "        ]\n",
    "\n",
    "        self.form_output_convert_cmd = Box(\n",
    "            form_convert_items,\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                justify_content=\"center\",\n",
    "                border=\"solid 3px lightgreen\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"70%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "    def move_allfiles(self, src_folder, dst_folder):\n",
    "        \"\"\"\n",
    "        Moves all files and directories from the source folder to the destination folder.\n",
    "        \"\"\"\n",
    "        files = os.listdir(src_folder)\n",
    "        for f in files:\n",
    "            fullpath = os.path.join(src_folder, f)\n",
    "            if os.path.isdir(fullpath):  # copy whole folder\n",
    "                shutil.move(fullpath, dst_folder)\n",
    "                print(f\"Copy finish: {f}\")\n",
    "\n",
    "    def show_headline(self, output):\n",
    "        \"\"\"\n",
    "        Displays a headline with the given output text in a styled HTML format.\n",
    "        \"\"\"\n",
    "        html0 = widgets.HTML(value=f\"<b><font color='lightblue'><font size=4>{output}</b>\")\n",
    "        display(html0)\n",
    "\n",
    "    def show_main(self):\n",
    "        \"\"\"\n",
    "        Displays the main interface for data preparation and training settings.\n",
    "        This method creates and displays various widgets including HTML, Accordion, Box, and Tab widgets\n",
    "        to allow users to configure data preparation and training settings. It also sets up observers and\n",
    "        interactive outputs to handle user interactions and updates.\n",
    "        Observers and Interactive Outputs:\n",
    "        - An observer to update dependent values based on the selected model name.\n",
    "        - An interactive output to handle parameter changes and update the visibility of widgets accordingly.\n",
    "        Button Click Handlers:\n",
    "        - Handlers for various buttons to trigger training, model selection, conversion, and testing actions.\n",
    "        \"\"\"\n",
    "\n",
    "        intro_text = \"Please Choose the setting of data prepare & train\"\n",
    "        html_widget = widgets.HTML(value=f\"<b><font color='lightgreen'><font size=6>{intro_text}</b>\")\n",
    "        display(html_widget)\n",
    "\n",
    "        # Create an accordion and put the 2 boxes\n",
    "        accordion = widgets.Accordion(children=[self.form_data_prepare_cmd, self.form_output_train_cmd]).add_class(\"parentstyle\")\n",
    "        display(HTML(\"<style>.parentstyle > .p-Accordion-child > .p-Collapse-header{background-color:green}</style>\"))\n",
    "        accordion.set_title(0, \"Data Download Setting\")\n",
    "        accordion.set_title(1, \"Configure the Training\")\n",
    "\n",
    "        # Create a box combining with 2 elements\n",
    "        box_data_train = Box(\n",
    "            [self.form_data_prepare_exist, accordion],\n",
    "            layout=Layout(\n",
    "                display=\"flex\",\n",
    "                flex_flow=\"column\",\n",
    "                border=\"solid 3px lightgreen\",\n",
    "                align_items=\"stretch\",\n",
    "                width=\"50%\",\n",
    "            ),\n",
    "        )\n",
    "\n",
    "        # Create a tab and put the 2 boxes\n",
    "        tab = widgets.Tab(children=[box_data_train, self.form_output_test_cmd, self.form_output_convert_cmd]).add_class(\"parentstyle\")\n",
    "        tab_contents = [\"Train\", \"Test\", \"Deployment\"]\n",
    "        tab.titles = tab_contents\n",
    "\n",
    "        output_widgets = widgets.Output(layout=Layout(border=\"1px solid green\"))\n",
    "\n",
    "        # Special observe for MODEL_NAME dependent value updating (fine tune layers)\n",
    "        def update(*args):\n",
    "            if self.e_ta.value.count(\"fdmobile\"):\n",
    "                self.k_ta.value = 6  # bcs the tf2cv structure is combining blocks\n",
    "                self.k_ta.max = 10\n",
    "                self.m_ta.layout.visibility = \"hidden\"\n",
    "                self.alpha_width_la.layout.visibility = \"hidden\"\n",
    "            elif self.e_ta.value.count(\"shufflenet\"):\n",
    "                self.k_ta.value = 10  # bcs the tf2cv structure is combining blocks\n",
    "                self.k_ta.max = 17\n",
    "                self.m_ta.layout.visibility = \"hidden\"\n",
    "                self.alpha_width_la.layout.visibility = \"hidden\"\n",
    "            elif self.e_ta.value.count(\"mobilenet_v1\"):\n",
    "                self.k_ta.max = 86\n",
    "                self.k_ta.value = 40\n",
    "                self.m_ta.layout.visibility = \"visible\"\n",
    "                self.alpha_width_la.layout.visibility = \"visible\"\n",
    "            elif self.e_ta.value.count(\"mobilenet_v2\"):\n",
    "                self.k_ta.max = 154\n",
    "                self.k_ta.value = 80\n",
    "                self.m_ta.layout.visibility = \"visible\"\n",
    "                self.alpha_width_la.layout.visibility = \"visible\"\n",
    "            elif self.e_ta.value.count(\"mobilenet_v3\"):\n",
    "                self.k_ta.max = 228\n",
    "                self.k_ta.value = 120\n",
    "                self.m_ta.layout.visibility = \"visible\"\n",
    "                self.alpha_width_la.layout.visibility = \"visible\"\n",
    "            elif self.e_ta.value.count(\"mobilenet_v3_mini\"):\n",
    "                self.k_ta.max = 102\n",
    "                self.k_ta.value = 50\n",
    "                self.m_ta.layout.visibility = \"visible\"\n",
    "                self.alpha_width_la.layout.visibility = \"visible\"\n",
    "            elif self.e_ta.value.count(\"efficientnetB0\"):\n",
    "                self.k_ta.max = 238\n",
    "                self.k_ta.value = 120\n",
    "                self.m_ta.layout.visibility = \"hidden\"\n",
    "                self.alpha_width_la.layout.visibility = \"hidden\"\n",
    "            elif self.e_ta.value.count(\"efficientnetv2B0\"):\n",
    "                self.k_ta.max = 170\n",
    "                self.k_ta.value = 150\n",
    "                self.m_ta.layout.visibility = \"hidden\"\n",
    "                self.alpha_width_la.layout.visibility = \"hidden\"\n",
    "\n",
    "        self.e_ta.observe(update)\n",
    "\n",
    "        def act_para(*, data_exist, url, zip_n, dataset_n, a_ta, b_ta, c_ta, d_ta, e_ta, f_ta, g_ta, h_ta, i_ta, j_ta, k_ta, a_cm, b_cm, c_cm, f_cm, g_cm, b_tt, m_ta, o_ta):\n",
    "\n",
    "            # If any value is changed, clear the widgets\n",
    "            with output_widgets:\n",
    "                output_widgets.clear_output()\n",
    "\n",
    "            if data_exist:\n",
    "                self.form_data_prepare_cmd.layout.visibility = \"hidden\"\n",
    "            else:\n",
    "                self.form_data_prepare_cmd.layout.visibility = \"visible\"\n",
    "\n",
    "            if j_ta.count(\"3:\"):\n",
    "                self.k_ta.layout.visibility = \"visible\"\n",
    "                self.fine_tune_la.layout.visibility = \"visible\"\n",
    "            else:\n",
    "                self.k_ta.layout.visibility = \"hidden\"\n",
    "                self.fine_tune_la.layout.visibility = \"hidden\"\n",
    "\n",
    "        # ------------------#\n",
    "        # widgets.Accordion's interactive input with action function `act_para()`\n",
    "        # ------------------#\n",
    "        out_inter = widgets.interactive_output(\n",
    "            act_para,\n",
    "            {\n",
    "                \"data_exist\": self.a_de,\n",
    "                \"url\": self.a_dp,\n",
    "                \"zip_n\": self.b_dp,\n",
    "                \"dataset_n\": self.b_de,\n",
    "                \"a_ta\": self.a_ta,\n",
    "                \"b_ta\": self.b_ta,\n",
    "                \"c_ta\": self.c_ta,\n",
    "                \"d_ta\": self.d_ta,\n",
    "                \"e_ta\": self.e_ta,\n",
    "                \"f_ta\": self.f_ta,\n",
    "                \"g_ta\": self.g_ta,\n",
    "                \"h_ta\": self.h_ta,\n",
    "                \"i_ta\": self.i_ta,\n",
    "                \"j_ta\": self.j_ta,\n",
    "                \"k_ta\": self.k_ta,\n",
    "                \"a_cm\": self.a_cm,\n",
    "                \"b_cm\": self.b_cm,\n",
    "                \"c_cm\": self.c_cm,\n",
    "                \"f_cm\": self.f_cm,\n",
    "                \"g_cm\": self.g_cm,\n",
    "                \"b_tt\": self.b_tt,\n",
    "                \"m_ta\": self.m_ta,\n",
    "                \"o_ta\": self.o_ta,\n",
    "            },\n",
    "        )\n",
    "\n",
    "        display(tab, out_inter)\n",
    "\n",
    "        # ------------------#\n",
    "        # for labelimg cmd, move to outside of act_para to prevent keep trigering\n",
    "        # ------------------#\n",
    "        # output_widgets = widgets.Output(layout=Layout(border = '1px solid green'))\n",
    "        display(output_widgets)\n",
    "\n",
    "        def on_button_clicked_train(b):\n",
    "            with output_widgets:\n",
    "                clear_output()\n",
    "                print(\"Train. . .\")\n",
    "                self.run_train()\n",
    "\n",
    "        self.l_ta.on_click(on_button_clicked_train)\n",
    "\n",
    "        def on_button_clicked_choose_tflite(b):\n",
    "            with output_widgets:\n",
    "                clear_output()\n",
    "                self.choose_tflite()\n",
    "\n",
    "        self.e_cm.on_click(on_button_clicked_choose_tflite)\n",
    "\n",
    "        def on_button_clicked_cpp(b):\n",
    "            with output_widgets:\n",
    "                clear_output()\n",
    "                print(\"Convert to cpp & Vela. . .\")\n",
    "                self.convert_tflu()\n",
    "\n",
    "        self.d_cm.on_click(on_button_clicked_cpp)\n",
    "\n",
    "        def on_button_clicked_choose_tflite_test(b):\n",
    "            with output_widgets:\n",
    "                clear_output()\n",
    "                self.choose_tflite()\n",
    "\n",
    "        self.a_tt.on_click(on_button_clicked_choose_tflite_test)\n",
    "\n",
    "        def on_button_clicked_test(b):\n",
    "            with output_widgets:\n",
    "                clear_output()\n",
    "                self.run_test_tflite()\n",
    "\n",
    "        self.c_tt.on_click(on_button_clicked_test)\n",
    "\n",
    "    def choose_tflite(self):\n",
    "        \"\"\"\n",
    "        Prompts the user to choose a TensorFlow Lite (.tflite) file for conversion.\n",
    "        This method uses a file chooser to allow the user to select a .tflite file from the workspace directory.\n",
    "        It restricts navigation to the current working directory and filters the files to only show .tflite files.\n",
    "        Once a file is selected, it sets the chosen directory and file name to the corresponding attributes and prints them.\n",
    "        Attributes:\n",
    "            a_cm (ipywidgets.Text): Widget to display the chosen directory.\n",
    "            b_cm (ipywidgets.Text): Widget to display the chosen .tflite file name.\n",
    "            c_cm (ipywidgets.Text): Widget to display the path for the 'vela' directory.\n",
    "        \"\"\"\n",
    "\n",
    "        path_ftflite = os.path.join(os.getcwd(), \"workspace\")\n",
    "        f_tflite = FileChooser(path_ftflite)\n",
    "        # Restrict navigation to /Users\n",
    "        f_tflite.sandbox_path = os.getcwd()\n",
    "        f_tflite.filter_pattern = [\"*.tflite\"]\n",
    "        f_tflite.title = \"<b><font color='lightblue'><font size=4>Choose the Tflite for Converting.</b>\"\n",
    "        display(f_tflite)\n",
    "\n",
    "        def act_test():\n",
    "            work_dir_name = os.getcwd().split(\"\\\\\")[-1]\n",
    "            m_src_dir = r\"..\" + f_tflite.selected_path.split(work_dir_name)[-1]\n",
    "            m_src_tflite = f_tflite.selected.split(\"\\\\\")[-1]\n",
    "            print(f\"The chosen dir: {m_src_dir}\")\n",
    "            print(f\"The chosen tflite: {m_src_tflite}\")\n",
    "            self.a_cm.value = m_src_dir\n",
    "            self.b_cm.value = m_src_tflite\n",
    "            self.c_cm.value = os.path.join(m_src_dir, \"vela\")\n",
    "\n",
    "            print(\"Finish!\")\n",
    "\n",
    "        evt = interact_manual(act_test)\n",
    "        evt.widget.children[0].description = \"Set this file\"  # because there are 3 parameter of the evt\n",
    "        evt.widget.children[0].button_style = \"primary\"\n",
    "\n",
    "    def run_train(self):\n",
    "        \"\"\"\n",
    "        Executes the training process for a machine learning model based on the provided parameters.\n",
    "        This method determines the appropriate training script to use and constructs the command to run it\n",
    "        with the specified parameters. It also records the command in a text file for future reference.\n",
    "        \"\"\"\n",
    "\n",
    "        sw_num = int(self.j_ta.value.split(\":\")[0])\n",
    "\n",
    "        if self.o_ta.value.count(\"1:\"):\n",
    "            dl_model_set = 0\n",
    "        elif self.o_ta.value.count(\"2:\"):\n",
    "            dl_model_set = 1\n",
    "        else:\n",
    "            dl_model_set = 2\n",
    "\n",
    "        if self.e_ta.value.count(\"fdmobile\") or self.e_ta.value.count(\"shufflenet\"):\n",
    "            python_file = \"train_tf2cv.py\"\n",
    "            print(\"Need tf2cv to run this model!\")\n",
    "        else:\n",
    "            python_file = \"train.py\"\n",
    "\n",
    "        %run $python_file --data_exist $self.a_de.value --url $self.a_dp.value --zip_name $self.b_dp.value --dataset_name $self.b_de.value \\\n",
    "        --proj_name $self.a_ta.value --BATCH_SIZE $self.b_ta.value --IMG_SIZE $self.c_ta.value --VAL_PCT $self.d_ta.value \\\n",
    "        --MODEL_NAME $self.e_ta.value --TEST_PCT $self.f_ta.value --DATA_AUGM $self.g_ta.value --EPOCHS $self.h_ta.value \\\n",
    "        --LEARNING_RATE $self.i_ta.value --FINE_TUNE_LAYER $self.k_ta.value --switch_mode $sw_num --ALPHA_WIDTH $self.m_ta.value \\\n",
    "        --IMAGENET_MODEL_EN $dl_model_set\n",
    "\n",
    "        path_cmd_record = os.path.join(os.getcwd(), \"workspace\", self.a_ta.value, \"train_cmd_record.txt\")\n",
    "        with open(path_cmd_record, \"w\", encoding='utf-8') as file1:\n",
    "            open_dir = f\"cd {os.getcwd()} \\n\"\n",
    "            cmd = f\"python {python_file} --data_exist {self.a_de.value} --url {self.a_dp.value} --zip_name {self.b_dp.value} --dataset_name {self.b_de.value} \\\n",
    "--proj_name {self.a_ta.value} \\\n",
    "--BATCH_SIZE { self.b_ta.value} --IMG_SIZE {self.c_ta.value} --VAL_PCT {self.d_ta.value} --MODEL_NAME {self.e_ta.value} \\\n",
    "--TEST_PCT {self.f_ta.value} --DATA_AUGM {self.g_ta.value} --EPOCHS {self.h_ta.value} --LEARNING_RATE {self.i_ta.value} \\\n",
    "--FINE_TUNE_LAYER {self.k_ta.value} --switch_mode {sw_num} --ALPHA_WIDTH {self.m_ta.value} \\\n",
    "--IMAGENET_MODEL_EN {dl_model_set}\"\n",
    "            file1.writelines([open_dir, cmd])\n",
    "        print(\"Finish !!\")\n",
    "\n",
    "    def run_test_tflite(self):\n",
    "        \"\"\"\n",
    "        Runs a test on a TensorFlow Lite model.\n",
    "        This method constructs the file path for the TensorFlow Lite model based on the provided configuration values,\n",
    "        determines the appropriate Python script to use for running the model, and prints relevant information about the\n",
    "        test.\n",
    "        \"\"\"\n",
    "\n",
    "        tflite_location = os.path.join((self.a_cm.value).split(\"..\\\\\")[-1], self.b_cm.value)\n",
    "        print(f\"The tflite file: {tflite_location} The number of test batch: {self.b_tt.value}\")\n",
    "\n",
    "        if self.b_cm.value.count(\"fdmobile\") or self.b_cm.value.count(\"shufflenet\"):\n",
    "            python_file = \"train_tf2cv.py\"\n",
    "            print(\"Need tf2cv to run this model!\")\n",
    "        else:\n",
    "            python_file = \"train.py\"\n",
    "\n",
    "        %run $python_file --data_exist $self.a_de.value --url $self.a_dp.value --zip_name $self.b_dp.value --dataset_name $self.b_de.value \\\n",
    "        --proj_name $self.a_ta.value --BATCH_SIZE $self.b_ta.value --IMG_SIZE $self.c_ta.value --VAL_PCT $self.d_ta.value \\\n",
    "        --MODEL_NAME $self.e_ta.value --TEST_PCT $self.f_ta.value --switch_mode 4 --ALPHA_WIDTH $self.m_ta.value\\\n",
    "        --TFLITE_F $tflite_location --TFLITE_TEST_BATCH_N $self.b_tt.value\n",
    "\n",
    "        print(\"Finish !!\")\n",
    "\n",
    "    def convert_tflu(self):\n",
    "        \"\"\"\n",
    "        Converts TensorFlow Lite model and generates necessary C++ source/header files for labels.\n",
    "        Attributes:\n",
    "            self.a_cm.value (str): Source directory path.\n",
    "            self.b_cm.value (str): Source file name.\n",
    "            self.c_cm.value (str): Generated directory path.\n",
    "            self.f_cm.value (str): Labels dataset name.\n",
    "            self.g_cm.value (bool): Flag to indicate whether to create the label C++ source/header files.\n",
    "        \"\"\"\n",
    "\n",
    "        %run exebat.py --SRC_DIR $self.a_cm.value --SRC_FILE $self.b_cm.value --GEN_DIR $self.c_cm.value\n",
    "\n",
    "        if self.g_cm.value:  # create the label C++ source/header files\n",
    "            # Change to dataset folder\n",
    "            old_cwd = os.getcwd()\n",
    "            batch_cwd = os.path.join(old_cwd, \"dataset\")\n",
    "            os.chdir(batch_cwd)\n",
    "\n",
    "            %run gen_labels_cpp.py --labels_dataset_name $self.f_cm.value --source_folder_path $self.a_cm.value \\\n",
    "            --header_folder_path $self.a_cm.value\n",
    "\n",
    "            os.chdir(old_cwd)\n",
    "            print(f'Finish, the label file is at: {(old_cwd + self.a_cm.value.split(\"..\")[1])}')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bc619de2-0d01-4133-a64c-168041666d5e",
   "metadata": {},
   "source": [
    "# Run Section\n",
    "---\n",
    "- The detail description of all the parameters and each step meaning is here [meaning](#id-train_evl_monitor)\n",
    "- In this notebook step, you have alreay finish the dataset prepared. If not, please go to `image_dataset\\create_data.ipynb`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c6b58cfe-a753-4869-a3fb-8a3cc9b76bcb",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "f8537a48f4bb4778a6958767c9d7d247",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HTML(value=\"<b><font color='lightgreen'><font size=6>Please Choose the setting of data prepare & train</b>\")"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<style>.parentstyle > .p-Accordion-child > .p-Collapse-header{background-color:green}</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "987ddb6880484e9aba2bb83112f15756",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Tab(children=(Box(children=(Box(children=(Box(children=(Label(value='Data Exist'), Checkbox(value=True, indent…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "6534128085e544a8a73e50642ebaae81",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output()"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "68558490a6b443469a97991c80d45aa0",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "Output(layout=Layout(border_bottom='1px solid green', border_left='1px solid green', border_right='1px solid g…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "act = TrainConfigAndCmdsWidgets()\n",
    "act.show_main()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "354f0e11-fdcf-4aa2-8e0f-b9fa5e8b5e54",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "NuEdgeWise_env",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}