baseLineModel.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "private_outputs": true,
      "provenance": [],
      "authorship_tag": "ABX9TyPyUHVlUzGh2/rL13WilfEc",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/eichi7/internshipHachi-x/blob/main/baseLineModel.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Baseline System is a model which can detect and recognize Ivorian dishes and estimate calorie\n",
        "It is a combination of Yolov5 and database ."
      ],
      "metadata": {
        "id": "n7fohaMft05m"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **I- Data labeling**\n",
        "\n",
        "\n",
        "How data labeling was doing on robowflow we obtained 887 images and 14 classes\n"
      ],
      "metadata": {
        "id": "r3fMvVF9RlG8"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ie5uLDH4uzAp"
      },
      "outputs": [],
      "source": [
        "# clone YOLOv5 repository\n",
        "!git clone https://github.com/ultralytics/yolov5  # clone\n",
        "%cd yolov5\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install -r requirements.txt  # install"
      ],
      "metadata": {
        "id": "Ma9PpGp1dTVz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wbvMlHd_QwMG"
      },
      "outputs": [],
      "source": [
        "# install dependencies as necessary\n",
        "!pip install -qr requirements.txt  # install dependencies (ignore errors)\n",
        "import torch\n",
        "\n",
        "from IPython.display import Image, clear_output  # to display images\n",
        "from utils.google_utils import gdrive_download  # to download models/datasets\n",
        "\n",
        "# clear_output()\n",
        "print('Setup complete. Using torch %s %s' % (torch.__version__, torch.cuda.get_device_properties(0) if torch.cuda.is_available() else 'CPU'))"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install roboflow\n",
        "\n",
        "from roboflow import Roboflow\n",
        "rf = Roboflow(api_key=\"FUq44fGv5FGSHq7oA1st\")\n",
        "project = rf.workspace(\"ezoa\").project(\"ivoriandishes-14\")\n",
        "dataset = project.version(1).download(\"yolov5\")"
      ],
      "metadata": {
        "id": "KXkk47ChVxmC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ug_PhK1oqwQA"
      },
      "outputs": [],
      "source": [
        "%cd /content/yolov5\n",
        "#after following the link above, recieve python code with these fields filled in\n",
        "#from roboflow import Roboflow\n",
        "#rf = Roboflow(api_key=\"YOUR API KEY HERE\")\n",
        "#project = rf.workspace().project(\"YOUR PROJECT\")\n",
        "#dataset = project.version(\"YOUR VERSION\").download(\"yolov5\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "ls"
      ],
      "metadata": {
        "id": "-HgecIWjWOBD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ZZ3DmmGQztJj"
      },
      "outputs": [],
      "source": [
        "# this is the YAML file Roboflow wrote for us that we're loading into this notebook with our data\n",
        "%cat {dataset.location}/data.yaml"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UwJx-2NHsYxT"
      },
      "source": [
        "# Define Model Configuration and Architecture\n",
        "\n",
        "We will write a yaml script that defines the parameters for our model like the number of classes, anchors, and each layer.\n",
        "\n",
        "You do not need to edit these cells, but you may."
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "v_zDRagoWgUc"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dOPn9wjOAwwK"
      },
      "outputs": [],
      "source": [
        "# define number of classes based on YAML\n",
        "import yaml\n",
        "with open(dataset.location + \"/data.yaml\", 'r') as stream:\n",
        "    num_classes = str(yaml.safe_load(stream)['nc'])"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Q3LytK19WtNM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1Rvt5wilnDyX"
      },
      "outputs": [],
      "source": [
        "#this is the model configuration we will use for our tutorial \n",
        "%cat /content/yolov5/models/yolov5s.yaml"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "SRkV2qYcW_n2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "t14hhyqdmw6O"
      },
      "outputs": [],
      "source": [
        "#customize iPython writefile so we can write variables\n",
        "from IPython.core.magic import register_line_cell_magic\n",
        "\n",
        "@register_line_cell_magic\n",
        "def writetemplate(line, cell):\n",
        "    with open(line, 'w') as f:\n",
        "        f.write(cell.format(**globals()))"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "M3OynaPoXCSt"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "uDxebz13RdRA"
      },
      "outputs": [],
      "source": [
        "%%writetemplate /content/yolov5/models/custom_yolov5s.yaml\n",
        "\n",
        "# parameters\n",
        "nc: {num_classes}  # number of classes\n",
        "depth_multiple: 0.33  # model depth multiple\n",
        "width_multiple: 0.50  # layer channel multiple\n",
        "\n",
        "# anchors\n",
        "anchors:\n",
        "  - [10,13, 16,30, 33,23]  # P3/8\n",
        "  - [30,61, 62,45, 59,119]  # P4/16\n",
        "  - [116,90, 156,198, 373,326]  # P5/32\n",
        "\n",
        "# YOLOv5 backbone\n",
        "backbone:\n",
        "  # [from, number, module, args]\n",
        "  [[-1, 1, Focus, [64, 3]],  # 0-P1/2\n",
        "   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4\n",
        "   [-1, 3, BottleneckCSP, [128]],\n",
        "   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8\n",
        "   [-1, 9, BottleneckCSP, [256]],\n",
        "   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16\n",
        "   [-1, 9, BottleneckCSP, [512]],\n",
        "   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32\n",
        "   [-1, 1, SPP, [1024, [5, 9, 13]]],\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 9\n",
        "  ]\n",
        "\n",
        "# YOLOv5 head\n",
        "head:\n",
        "  [[-1, 1, Conv, [512, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 6], 1, Concat, [1]],  # cat backbone P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 13\n",
        "\n",
        "   [-1, 1, Conv, [256, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 4], 1, Concat, [1]],  # cat backbone P3\n",
        "   [-1, 3, BottleneckCSP, [256, False]],  # 17 (P3/8-small)\n",
        "\n",
        "   [-1, 1, Conv, [256, 3, 2]],\n",
        "   [[-1, 14], 1, Concat, [1]],  # cat head P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 20 (P4/16-medium)\n",
        "\n",
        "   [-1, 1, Conv, [512, 3, 2]],\n",
        "   [[-1, 10], 1, Concat, [1]],  # cat head P5\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 23 (P5/32-large)\n",
        "\n",
        "   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)\n",
        "  ]"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "i2AIdVJJXNGo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VUOiNLtMP5aG"
      },
      "source": [
        "# Train Custom YOLOv5 Detector\n",
        "\n",
        "### Next, we'll fire off training!\n",
        "\n",
        "\n",
        "Here, we are able to pass a number of arguments:\n",
        "- **img:** define input image size\n",
        "- **batch:** determine batch size\n",
        "- **epochs:** define the number of training epochs. (Note: often, 3000+ are common here!)\n",
        "- **data:** set the path to our yaml file\n",
        "- **cfg:** specify our model configuration\n",
        "- **weights:** specify a custom path to weights. (Note: you can download weights from the Ultralytics Google Drive [folder](https://drive.google.com/open?id=1Drs_Aiu7xx6S-ix95f9kNsA6ueKRpN2J))\n",
        "- **name:** result names\n",
        "- **nosave:** only save the final checkpoint\n",
        "- **cache:** cache images for faster training"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "lVWdS-UPXUIX"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1NcFxRcFdJ_O"
      },
      "outputs": [],
      "source": [
        "# train yolov5s on custom data for 0 epochs\n",
        "# time its performance\n",
        "%%time\n",
        "%cd /content/yolov5/\n",
        "!python train.py --img 416 --batch 16 --epochs 101 --data {dataset.location}/data.yaml --cfg ./models/custom_yolov5s.yaml --weights '' --name yolov5s_results  --cache"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "qzzCBxRroEyh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bOy5KI2ncnWd"
      },
      "outputs": [],
      "source": [
        "# Start tensorboard\n",
        "# Launch after you have started training\n",
        "# logs save in the folder \"runs\"\n",
        "%load_ext tensorboard\n",
        "%tensorboard --logdir runs"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#display inference on ALL test images\n",
        "#this looks much better with longer training above\n",
        "#display inference on ALL test images\n",
        "#this looks much better with longer training above\n"
      ],
      "metadata": {
        "id": "z0tiQUTro6iC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PF9MLHDb7tB6"
      },
      "outputs": [],
      "source": [
        "# first, display our ground truth data\n",
        "print(\"GROUND TRUTH TRAINING DATA:\")\n",
        "Image(filename='/content/yolov5/runs/train/yolov5s_results/test_batch0_labels.jpg', width=900)"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "aaAKIbm5t3br"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9nmZZnWOgJ2S"
      },
      "outputs": [],
      "source": [
        "# when we ran this, we saw .007 second inference time. That is 140 FPS on a TESLA P100!\n",
        "# use the best weights!\n",
        "%cd /content/yolov5/\n",
        "!python detect.py --weights runs/train/yolov5s_results/weights/best.pt --img 416 --conf 0.4 --source ../yolov5/ivoriandishes-14-1/test/images"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "odKEqYtTgbRc"
      },
      "outputs": [],
      "source": [
        "#display inference on ALL test images\n",
        "#this looks much better with longer training above\n",
        "\n",
        "import glob\n",
        "from IPython.display import Image, display\n",
        "\n",
        "for imageName in glob.glob('/content/yolov5/runs/detect/exp/*.jpg'): #assuming JPG\n",
        "    display(Image(filename=imageName))\n",
        "    print(\"\\n\")"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Training with data augmentation\n"
      ],
      "metadata": {
        "id": "DsrGWRL1wjBb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install roboflow\n",
        "\n",
        "from roboflow import Roboflow\n",
        "rf = Roboflow(api_key=\"FUq44fGv5FGSHq7oA1st\")\n",
        "project = rf.workspace(\"ezoa\").project(\"ivoriandishes-14\")\n",
        "dataset = project.version(3).download(\"yolov5\")"
      ],
      "metadata": {
        "id": "sZoFuPiUwnmU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ls"
      ],
      "metadata": {
        "id": "tCX1E7lnxO2H"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "HUwj7Wv-xqt-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HmQWHWsEyMxA"
      },
      "outputs": [],
      "source": [
        "# this is the YAML file Roboflow wrote for us that we're loading into this notebook with our data\n",
        "%cat {dataset.location}/data.yaml"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "t6lcutm2ycdJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Jc336AeLyc97"
      },
      "outputs": [],
      "source": [
        "# define number of classes based on YAML\n",
        "import yaml\n",
        "with open(dataset.location + \"/data.yaml\", 'r') as stream:\n",
        "    num_classes = str(yaml.safe_load(stream)['nc'])"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "4L_-M8Ynyl3s"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cYXNCeOaymWC"
      },
      "outputs": [],
      "source": [
        "#this is the model configuration we will use for our tutorial \n",
        "%cat /content/yolov5/models/yolov5s.yaml"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#customizemize iPython writefile so we can write variables\n",
        "from IPython.core.magic import register_line_cell_magic\n",
        "\n",
        "@register_line_cell_magic\n",
        "def writetemplate(line, cell):\n",
        "    with open(line, 'w') as f:\n",
        "        f.write(cell.format(**globals()))"
      ],
      "metadata": {
        "id": "pzTAD4Oay0zI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "FPCN01jA0Nrs"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DPGerSuY0OHC"
      },
      "outputs": [],
      "source": [
        "%%writetemplate /content/yolov5/models/custom_yolov5s.yaml\n",
        "\n",
        "# parameters\n",
        "nc: {num_classes}  # number of classes\n",
        "depth_multiple: 0.33  # model depth multiple\n",
        "width_multiple: 0.50  # layer channel multiple\n",
        "\n",
        "# anchors\n",
        "anchors:\n",
        "  - [10,13, 16,30, 33,23]  # P3/8\n",
        "  - [30,61, 62,45, 59,119]  # P4/16\n",
        "  - [116,90, 156,198, 373,326]  # P5/32\n",
        "\n",
        "# YOLOv5 backbone\n",
        "backbone:\n",
        "  # [from, number, module, args]\n",
        "  [[-1, 1, Focus, [64, 3]],  # 0-P1/2\n",
        "   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4\n",
        "   [-1, 3, BottleneckCSP, [128]],\n",
        "   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8\n",
        "   [-1, 9, BottleneckCSP, [256]],\n",
        "   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16\n",
        "   [-1, 9, BottleneckCSP, [512]],\n",
        "   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32\n",
        "   [-1, 1, SPP, [1024, [5, 9, 13]]],\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 9\n",
        "  ]\n",
        "\n",
        "# YOLOv5 head\n",
        "head:\n",
        "  [[-1, 1, Conv, [512, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 6], 1, Concat, [1]],  # cat backbone P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 13\n",
        "\n",
        "   [-1, 1, Conv, [256, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 4], 1, Concat, [1]],  # cat backbone P3\n",
        "   [-1, 3, BottleneckCSP, [256, False]],  # 17 (P3/8-small)\n",
        "\n",
        "   [-1, 1, Conv, [256, 3, 2]],\n",
        "   [[-1, 14], 1, Concat, [1]],  # cat head P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 20 (P4/16-medium)\n",
        "\n",
        "   [-1, 1, Conv, [512, 3, 2]],\n",
        "   [[-1, 10], 1, Concat, [1]],  # cat head P5\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 23 (P5/32-large)\n",
        "\n",
        "   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)\n",
        "  ]"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "RBhUPfKx1FKf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TcqRy4rp1Fpt"
      },
      "outputs": [],
      "source": [
        "# train yolov5s on custom data for 0 epochs\n",
        "# time its performance\n",
        "%%time\n",
        "%cd /content/yolov5/\n",
        "!python train.py --img 416 --batch 16 --epochs 101 --data {dataset.location}/data.yaml --cfg ./models/custom_yolov5s.yaml --weights '' --name yolov5s_results_1  --cache"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "C_5EfnA-MYYM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "LhjS5hdy45Qw"
      },
      "outputs": [],
      "source": [
        "# Start tensorboard\n",
        "# Launch after you have started training\n",
        "# logs save in the folder \"runs\"\n",
        "%load_ext tensorboard\n",
        "%tensorboard --logdir runs"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "New Test on Ivorian with mix data augmentation "
      ],
      "metadata": {
        "id": "ReUf5liaMgur"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install roboflow\n",
        "\n",
        "from roboflow import Roboflow\n",
        "rf = Roboflow(api_key=\"FUq44fGv5FGSHq7oA1st\")\n",
        "project = rf.workspace(\"ezoa\").project(\"ivoriandishes-14\")\n",
        "dataset = project.version(6).download(\"yolov5\")"
      ],
      "metadata": {
        "id": "RD48f9xahPge"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "DiirrWUNiHsD"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QIIY6PjdibCG"
      },
      "outputs": [],
      "source": [
        "# this is the YAML file Roboflow wrote for us that we're loading into this notebook with our data\n",
        "%cat {dataset.location}/data.yaml"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "8jHY0ADlijoe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "gRSv9XGyikKf"
      },
      "outputs": [],
      "source": [
        "# define number of classes based on YAML\n",
        "import yaml\n",
        "with open(dataset.location + \"/data.yaml\", 'r') as stream:\n",
        "    num_classes = str(yaml.safe_load(stream)['nc'])"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "HDgyIA04itzP"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "RsIescKmiueT"
      },
      "outputs": [],
      "source": [
        "#this is the model configuration we will use for our tutorial \n",
        "%cat /content/yolov5/models/yolov5s.yaml"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tbKH901Bi-0T"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#customizemize iPython writefile so we can write variables\n",
        "from IPython.core.magic import register_line_cell_magic\n",
        "\n",
        "@register_line_cell_magic\n",
        "def writetemplate(line, cell):\n",
        "    with open(line, 'w') as f:\n",
        "        f.write(cell.format(**globals()))"
      ],
      "metadata": {
        "id": "RAaHQ_Jki_TU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "5Grn3JeFjIS0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "mDTtVemqjIyD"
      },
      "outputs": [],
      "source": [
        "%%writetemplate /content/yolov5/models/custom_yolov5s.yaml\n",
        "\n",
        "# parameters\n",
        "nc: {num_classes}  # number of classes\n",
        "depth_multiple: 0.33  # model depth multiple\n",
        "width_multiple: 0.50  # layer channel multiple\n",
        "\n",
        "# anchors\n",
        "anchors:\n",
        "  - [10,13, 16,30, 33,23]  # P3/8\n",
        "  - [30,61, 62,45, 59,119]  # P4/16\n",
        "  - [116,90, 156,198, 373,326]  # P5/32\n",
        "\n",
        "# YOLOv5 backbone\n",
        "backbone:\n",
        "  # [from, number, module, args]\n",
        "  [[-1, 1, Focus, [64, 3]],  # 0-P1/2\n",
        "   [-1, 1, Conv, [128, 3, 2]],  # 1-P2/4\n",
        "   [-1, 3, BottleneckCSP, [128]],\n",
        "   [-1, 1, Conv, [256, 3, 2]],  # 3-P3/8\n",
        "   [-1, 9, BottleneckCSP, [256]],\n",
        "   [-1, 1, Conv, [512, 3, 2]],  # 5-P4/16\n",
        "   [-1, 9, BottleneckCSP, [512]],\n",
        "   [-1, 1, Conv, [1024, 3, 2]],  # 7-P5/32\n",
        "   [-1, 1, SPP, [1024, [5, 9, 13]]],\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 9\n",
        "  ]\n",
        "\n",
        "# YOLOv5 head\n",
        "head:\n",
        "  [[-1, 1, Conv, [512, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 6], 1, Concat, [1]],  # cat backbone P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 13\n",
        "\n",
        "   [-1, 1, Conv, [256, 1, 1]],\n",
        "   [-1, 1, nn.Upsample, [None, 2, 'nearest']],\n",
        "   [[-1, 4], 1, Concat, [1]],  # cat backbone P3\n",
        "   [-1, 3, BottleneckCSP, [256, False]],  # 17 (P3/8-small)\n",
        "\n",
        "   [-1, 1, Conv, [256, 3, 2]],\n",
        "   [[-1, 14], 1, Concat, [1]],  # cat head P4\n",
        "   [-1, 3, BottleneckCSP, [512, False]],  # 20 (P4/16-medium)\n",
        "\n",
        "   [-1, 1, Conv, [512, 3, 2]],\n",
        "   [[-1, 10], 1, Concat, [1]],  # cat head P5\n",
        "   [-1, 3, BottleneckCSP, [1024, False]],  # 23 (P5/32-large)\n",
        "\n",
        "   [[17, 20, 23], 1, Detect, [nc, anchors]],  # Detect(P3, P4, P5)\n",
        "  ]"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "9xboICSXjV5F"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jr6U1z_7jWaA"
      },
      "outputs": [],
      "source": [
        "# train yolov5s on custom data for 0 epochs\n",
        "# time its performance\n",
        "%%time\n",
        "%cd /content/yolov5/\n",
        "!python train.py --img 416 --batch 16 --epochs 101 --data {dataset.location}/data.yaml --cfg ./models/custom_yolov5s.yaml --weights '' --name yolov5s_results_2  --cache"
      ]
    }
  ]
}