Yolov3/Yolov3_SPP_improved.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Copy of SIMS_Yolov3_SPP.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p3HE4j_2wQru",
        "colab_type": "text"
      },
      "source": [
        "**Copy images and annotation from drive**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OkX9da7Bv6qI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%cp '/content/drive/My Drive/SIMS_Dataset/images.zip' /content/\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/training.txt' /content/\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/test.txt' /content/\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/validation.txt' /content/\n",
        "!unzip images.zip -d /content/"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "773gpD8uyDui",
        "colab_type": "text"
      },
      "source": [
        "**Preprocessing** Did once and copied processed categories into drive for later usage"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n-QXVSNIy-E3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#making directories\n",
        "\n",
        "!mkdir yolo_train/\n",
        "!mkdir yolo_train/images\n",
        "!mkdir yolo_train/labels\n",
        "\n",
        "!mkdir yolo_test/\n",
        "!mkdir yolo_test/images\n",
        "!mkdir yolo_test/labels\n",
        "\n",
        "!mkdir yolo_val/\n",
        "!mkdir yolo_val/images\n",
        "!mkdir yolo_val/labels"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oFs5UhxKwkLq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#separate images based on train,val and test and put corresponding files into \n",
        "#respective folder having images and labels as subfolders in it\n",
        "import shutil\n",
        "import os\n",
        "\n",
        "def copy(filepath,dest_dir):\n",
        "    with open(filepath) as fp:\n",
        "        for line in fp:\n",
        "            os.makedirs(dest_dir, exist_ok=True)  # succeeds even if directory exists.\n",
        "            shutil.copy(line.replace('\\n', ''), dest_dir+\"images\") #image\n",
        "            shutil.copy(line.replace('.jpg\\n', '.txt'), dest_dir+\"labels\") ##label\n",
        "    print(\"Done: \"+ filepath)\n",
        "            \n",
        "copy('./training.txt','yolo_train/')\n",
        "copy('./validation.txt','yolo_val/')\n",
        "copy('./test.txt','yolo_test/')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cZZSt5sf26io",
        "colab_type": "text"
      },
      "source": [
        "**copy pre-processed data for yolov3 work, clone and install install requirements**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Oomm6rmFPgl9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# #update paths of train,test,val\n",
        "\n",
        "def copy(filepath,newfile,binder):\n",
        "    with open(filepath) as fp:\n",
        "      print(filepath)\n",
        "      for line in fp:\n",
        "        in_file = line[2:]\n",
        "        new = binder+in_file\n",
        "        with open(newfile, \"a\") as f:\n",
        "          f.write(new)\n",
        "\n",
        "copy('training.txt','Dataset/yolo_train/yolo_train.txt',\n",
        "     \"/content/yolov3/Dataset/yolo_train/\")\n",
        "\n",
        "copy('test.txt','Dataset/yolo_test/yolo_test.txt',\n",
        "     \"/content/yolov3/Dataset/yolo_test/\")\n",
        "\n",
        "copy('validation.txt','Dataset/yolo_val/yolo_val.txt',\n",
        "     \"/content/yolov3/Dataset/yolo_val/\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8WvBMblVFelr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%cp '/content/drive/My Drive/SIMS_Dataset/training.txt' /content/yolov3/\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/test.txt' /content/yolov3/\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/validation.txt' /content/yolov3/"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XFmmeGG41F9V",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Copied dataset and relevant files to drive\n",
        "!zip -r SIMS_Dataset_yolo.zip /content/yolov3/Dataset/\n",
        "!cp /content/yolov3/SIMS_Dataset_yolo.zip '/content/drive/My Drive/SIMS_Dataset'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zz44GWdLOQAw",
        "colab_type": "text"
      },
      "source": [
        "**Preprocessing ended here and data is saved in drive**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_p7tHeo622UK",
        "colab_type": "code",
        "outputId": "bb76ae82-2921-4639-fa17-3a435033b5df",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "!git clone https://github.com/ultralytics/yolov3\n",
        "%cd yolov3/\n",
        "!pip install -U -r requirements.txt"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'yolov3'...\n",
            "remote: Enumerating objects: 12, done.\u001b[K\n",
            "remote: Counting objects: 100% (12/12), done.\u001b[K\n",
            "remote: Compressing objects: 100% (11/11), done.\u001b[K\n",
            "remote: Total 9094 (delta 2), reused 7 (delta 1), pack-reused 9082\u001b[K\n",
            "Receiving objects: 100% (9094/9094), 7.07 MiB | 19.61 MiB/s, done.\n",
            "Resolving deltas: 100% (6262/6262), done.\n",
            "/content/yolov3\n",
            "Collecting numpy\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/03/27/e35e7c6e6a52fab9fcc64fc2b20c6b516eba930bb02b10ace3b38200d3ab/numpy-1.18.4-cp36-cp36m-manylinux1_x86_64.whl (20.2MB)\n",
            "\u001b[K     |████████████████████████████████| 20.2MB 38.5MB/s \n",
            "\u001b[?25hCollecting opencv-python>=4.1\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/72/c2/e9cf54ae5b1102020ef895866a67cb2e1aef72f16dd1fde5b5fb1495ad9c/opencv_python-4.2.0.34-cp36-cp36m-manylinux1_x86_64.whl (28.2MB)\n",
            "\u001b[K     |████████████████████████████████| 28.2MB 79kB/s \n",
            "\u001b[?25hRequirement already up-to-date: torch>=1.5 in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 4)) (1.5.0+cu101)\n",
            "Requirement already up-to-date: matplotlib in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 5)) (3.2.1)\n",
            "Requirement already up-to-date: pycocotools in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 6)) (2.0.0)\n",
            "Collecting tqdm\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/c9/40/058b12e8ba10e35f89c9b1fdfc2d4c7f8c05947df2d5eb3c7b258019fda0/tqdm-4.46.0-py2.py3-none-any.whl (63kB)\n",
            "\u001b[K     |████████████████████████████████| 71kB 11.8MB/s \n",
            "\u001b[?25hCollecting pillow\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/e0/50/8e78e6f62ffa50d6ca95c281d5a2819bef66d023ac1b723e253de5bda9c5/Pillow-7.1.2-cp36-cp36m-manylinux1_x86_64.whl (2.1MB)\n",
            "\u001b[K     |████████████████████████████████| 2.1MB 43.8MB/s \n",
            "\u001b[?25hRequirement already up-to-date: tensorboard>=1.14 in /usr/local/lib/python3.6/dist-packages (from -r requirements.txt (line 9)) (2.2.1)\n",
            "Requirement already satisfied, skipping upgrade: future in /usr/local/lib/python3.6/dist-packages (from torch>=1.5->-r requirements.txt (line 4)) (0.16.0)\n",
            "Requirement already satisfied, skipping upgrade: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->-r requirements.txt (line 5)) (2.4.7)\n",
            "Requirement already satisfied, skipping upgrade: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->-r requirements.txt (line 5)) (2.8.1)\n",
            "Requirement already satisfied, skipping upgrade: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib->-r requirements.txt (line 5)) (0.10.0)\n",
            "Requirement already satisfied, skipping upgrade: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib->-r requirements.txt (line 5)) (1.2.0)\n",
            "Requirement already satisfied, skipping upgrade: google-auth-oauthlib<0.5,>=0.4.1 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (0.4.1)\n",
            "Requirement already satisfied, skipping upgrade: setuptools>=41.0.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (46.1.3)\n",
            "Requirement already satisfied, skipping upgrade: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (1.0.1)\n",
            "Requirement already satisfied, skipping upgrade: absl-py>=0.4 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (0.9.0)\n",
            "Requirement already satisfied, skipping upgrade: grpcio>=1.24.3 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (1.28.1)\n",
            "Requirement already satisfied, skipping upgrade: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (1.12.0)\n",
            "Requirement already satisfied, skipping upgrade: protobuf>=3.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (3.10.0)\n",
            "Requirement already satisfied, skipping upgrade: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (3.2.1)\n",
            "Requirement already satisfied, skipping upgrade: google-auth<2,>=1.6.3 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (1.7.2)\n",
            "Requirement already satisfied, skipping upgrade: wheel>=0.26; python_version >= \"3\" in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (0.34.2)\n",
            "Requirement already satisfied, skipping upgrade: requests<3,>=2.21.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (2.23.0)\n",
            "Requirement already satisfied, skipping upgrade: tensorboard-plugin-wit>=1.6.0 in /usr/local/lib/python3.6/dist-packages (from tensorboard>=1.14->-r requirements.txt (line 9)) (1.6.0.post3)\n",
            "Requirement already satisfied, skipping upgrade: requests-oauthlib>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=1.14->-r requirements.txt (line 9)) (1.3.0)\n",
            "Requirement already satisfied, skipping upgrade: pyasn1-modules>=0.2.1 in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->-r requirements.txt (line 9)) (0.2.8)\n",
            "Requirement already satisfied, skipping upgrade: rsa<4.1,>=3.1.4 in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->-r requirements.txt (line 9)) (4.0)\n",
            "Requirement already satisfied, skipping upgrade: cachetools<3.2,>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from google-auth<2,>=1.6.3->tensorboard>=1.14->-r requirements.txt (line 9)) (3.1.1)\n",
            "Requirement already satisfied, skipping upgrade: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests<3,>=2.21.0->tensorboard>=1.14->-r requirements.txt (line 9)) (1.24.3)\n",
            "Requirement already satisfied, skipping upgrade: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests<3,>=2.21.0->tensorboard>=1.14->-r requirements.txt (line 9)) (3.0.4)\n",
            "Requirement already satisfied, skipping upgrade: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests<3,>=2.21.0->tensorboard>=1.14->-r requirements.txt (line 9)) (2.9)\n",
            "Requirement already satisfied, skipping upgrade: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests<3,>=2.21.0->tensorboard>=1.14->-r requirements.txt (line 9)) (2020.4.5.1)\n",
            "Requirement already satisfied, skipping upgrade: oauthlib>=3.0.0 in /usr/local/lib/python3.6/dist-packages (from requests-oauthlib>=0.7.0->google-auth-oauthlib<0.5,>=0.4.1->tensorboard>=1.14->-r requirements.txt (line 9)) (3.1.0)\n",
            "Requirement already satisfied, skipping upgrade: pyasn1<0.5.0,>=0.4.6 in /usr/local/lib/python3.6/dist-packages (from pyasn1-modules>=0.2.1->google-auth<2,>=1.6.3->tensorboard>=1.14->-r requirements.txt (line 9)) (0.4.8)\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",
            "\u001b[31mERROR: albumentations 0.1.12 has requirement imgaug<0.2.7,>=0.2.5, but you'll have imgaug 0.2.9 which is incompatible.\u001b[0m\n",
            "Installing collected packages: numpy, opencv-python, tqdm, pillow\n",
            "  Found existing installation: numpy 1.18.3\n",
            "    Uninstalling numpy-1.18.3:\n",
            "      Successfully uninstalled numpy-1.18.3\n",
            "  Found existing installation: opencv-python 4.1.2.30\n",
            "    Uninstalling opencv-python-4.1.2.30:\n",
            "      Successfully uninstalled opencv-python-4.1.2.30\n",
            "  Found existing installation: tqdm 4.38.0\n",
            "    Uninstalling tqdm-4.38.0:\n",
            "      Successfully uninstalled tqdm-4.38.0\n",
            "  Found existing installation: Pillow 7.0.0\n",
            "    Uninstalling Pillow-7.0.0:\n",
            "      Successfully uninstalled Pillow-7.0.0\n",
            "Successfully installed numpy-1.18.4 opencv-python-4.2.0.34 pillow-7.1.2 tqdm-4.46.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "PIL",
                  "numpy",
                  "tqdm"
                ]
              }
            }
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EqOwJUq3-eV_",
        "colab_type": "text"
      },
      "source": [
        "**created sims_classes.names and sims.data files**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zaN1KBtzJnDU",
        "colab_type": "code",
        "outputId": "4eac937f-8c01-4cd6-bf12-8ad3accea752",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "%cd yolov3"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/yolov3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9hY9QBVkOIpK",
        "colab_type": "text"
      },
      "source": [
        "**Load Already preprocessed Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9BxPfot1Zkxc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# #using copied/preprocessed data now\n",
        "# #!mkdir /content/yolov3/Dataset\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/SIMS_Dataset_yolo.zip' /content/yolov3/\n",
        "!unzip SIMS_Dataset_yolo.zip -d /content/yolov3/"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7OUnd4rDBQw0",
        "colab_type": "code",
        "outputId": "f945c4bb-0afa-493a-f15e-2f9202d2c07f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "#download pretrained weights\n",
        "!. weights/download_yolov3_weights.sh"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100   408    0   408    0     0   1924      0 --:--:-- --:--:-- --:--:--  1924\n",
            "  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0\n",
            "  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0\n",
            "100  236M    0  236M    0     0  73.3M      0 --:--:--  0:00:03 --:--:--  100M\n",
            "Downloading https://drive.google.com/uc?export=download&id=1SHNFyoe5Ni8DajDNEqgB2oVKBb_NoEad as weights/yolov3.pt... Done (3.6s)\n",
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100   408    0   408    0     0   2487      0 --:--:-- --:--:-- --:--:--  2487\n",
            "  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0\n",
            "  0     0    0     0    0     0      0      0 --:--:-- --:--:-- --:--:--     0\n",
            "100  240M    0  240M    0     0  70.2M      0 --:--:--  0:00:03 --:--:-- 80.9M\n",
            "Downloading https://drive.google.com/uc?export=download&id=1f6Ovy3BSq2wYq4UfvFUpxJFNDFfrIDcR as weights/yolov3-spp.pt... Done (3.7s)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Liht8KxxOcc_",
        "colab_type": "text"
      },
      "source": [
        "**Start Training**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "omNsjqJ4iE7o",
        "colab_type": "code",
        "outputId": "602b39eb-180c-4e6f-ecbb-f3598026e2f2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#start training\n",
        "!python3 train.py --batch 8 --epochs 60 --img-size 512 --data Dataset/sims.txt --cache-images --rec --cfg yolov3-spp.cfg --name from_yolov3_spp --weights weights/yolov3-spp.pt"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Apex recommended for faster mixed precision training: https://github.com/NVIDIA/apex\n",
            "Namespace(adam=False, batch_size=8, bucket='', cache_images=True, cfg='yolov3-spp.cfg', data='Dataset/sims.txt', device='', epochs=60, evolve=False, img_size=[512], multi_scale=False, name='from_yolov3_spp', nosave=False, notest=False, rect=True, resume=False, single_cls=False, weights='weights/yolov3-spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)\n",
            "\n",
            "Start Tensorboard with \"tensorboard --logdir=runs\", view at http://localhost:6006/\n",
            "2020-05-06 22:39:03.493446: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Optimizer groups: 76 .bias, 76 Conv2d.weight, 73 other\n",
            "Caching labels (3232 found, 0 missing, 0 empty, 0 duplicate, for 3233 images): 100% 3233/3233 [00:00<00:00, 7908.87it/s]\n",
            "Caching images (1.9GB): 100% 3233/3233 [00:40<00:00, 80.49it/s]\n",
            "Caching labels (1740 found, 0 missing, 0 empty, 0 duplicate, for 1741 images): 100% 1741/1741 [00:00<00:00, 7560.49it/s]\n",
            "Caching images (1.0GB): 100% 1741/1741 [00:21<00:00, 82.70it/s]\n",
            "Image sizes 512 - 512 train, 512 test\n",
            "Using 2 dataloader workers\n",
            "Starting training for 60 epochs...\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      0/59     4.94G      5.29      4.29       4.6      14.2         0       512: 100% 405/405 [04:33<00:00,  1.48it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:15<00:00,  2.90it/s]\n",
            "                 all  1.74e+03  2.12e+04    0.0672     0.109    0.0646     0.072\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      1/59     4.95G      4.02      2.17      3.81        10         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:16<00:00,  2.85it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.196      0.31     0.175     0.172\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      2/59     4.95G       3.5      1.66      3.27      8.43         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:13<00:00,  2.98it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.21     0.362     0.227     0.188\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "  0% 0/405 [00:00<?, ?it/s]\n",
            "Model Bias Summary:    layer        regression        objectness    classification\n",
            "                          89      -0.14+/-0.16      -4.93+/-0.92      -3.11+/-0.35 \n",
            "                         101      -0.08+/-0.23      -4.62+/-0.20      -3.10+/-0.36 \n",
            "                         113       0.09+/-0.31      -5.49+/-0.67      -3.14+/-0.53 \n",
            "      3/59     4.95G      3.15      1.41      2.84       7.4         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:09<00:00,  3.15it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.18     0.467     0.254     0.242\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      4/59     4.95G      2.82      1.26      2.54      6.62         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:04<00:00,  3.36it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.211     0.467     0.284     0.265\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      5/59     4.95G      2.67      1.17      2.24      6.08         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:02<00:00,  3.50it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.223     0.517     0.366     0.301\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      6/59     4.95G      2.52      1.08      1.96      5.56         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:02<00:00,  3.49it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.311     0.534     0.392     0.329\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      7/59     4.95G       2.5      1.03      1.73      5.26         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:00<00:00,  3.59it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.254     0.576     0.405      0.31\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      8/59     4.95G      2.37     0.987      1.52      4.88         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:59<00:00,  3.65it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.271     0.577     0.445     0.331\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "      9/59     4.95G      2.39     0.939      1.34      4.67         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [01:00<00:00,  3.61it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.249     0.591     0.413     0.326\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     10/59     4.95G      2.37     0.921      1.24      4.54         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:59<00:00,  3.67it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.27     0.589     0.446     0.348\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     11/59     4.95G      2.26     0.878       1.1      4.23         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:58<00:00,  3.72it/s]\n",
            "                 all  1.74e+03  2.12e+04       0.3     0.606     0.455     0.373\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     12/59     4.95G      2.24     0.852      1.03      4.13         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:58<00:00,  3.72it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.301     0.598     0.462     0.382\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     13/59     4.95G      2.13     0.819     0.956       3.9         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:57<00:00,  3.76it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.329     0.591     0.479     0.409\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     14/59     4.95G       2.1     0.793     0.874      3.77         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:58<00:00,  3.74it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.323     0.564     0.453     0.396\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     15/59     4.95G      2.09     0.786     0.851      3.72         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:57<00:00,  3.80it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.395     0.581      0.47     0.415\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     16/59     4.95G      2.04     0.751     0.779      3.57         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:57<00:00,  3.80it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.409      0.59     0.496     0.418\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     17/59     4.95G      2.03     0.739     0.762      3.53         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:58<00:00,  3.73it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.408     0.597     0.486     0.415\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     18/59     4.95G      1.94     0.718     0.702      3.36         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:57<00:00,  3.82it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.422     0.596     0.502     0.434\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     19/59     4.95G      1.92       0.7     0.663      3.29         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.85it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.415     0.589      0.49     0.434\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     20/59     4.95G      1.88     0.693     0.635       3.2         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:57<00:00,  3.82it/s]\n",
            "                 all  1.74e+03  2.12e+04       0.4      0.63     0.488     0.426\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     21/59     4.95G      1.82     0.665      0.59      3.08         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.84it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.438     0.609      0.52     0.461\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     22/59     4.95G      1.75     0.642     0.532      2.92         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.86it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.425     0.617     0.507     0.452\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     23/59     4.95G      1.77      0.63     0.509      2.91         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.89it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.433     0.632     0.503     0.459\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     24/59     4.95G      1.71      0.61     0.486       2.8         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.89it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.458     0.623     0.523      0.48\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     25/59     4.95G      1.67     0.595     0.453      2.72         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.88it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.438     0.616     0.502     0.457\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     26/59     4.95G      1.68     0.594     0.436      2.71         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.90it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.457     0.638     0.538     0.501\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     27/59     4.95G      1.62     0.574      0.42      2.61         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.88it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.443     0.634     0.504     0.482\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     28/59     4.95G       1.6     0.562     0.394      2.55         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.88it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.458     0.653     0.543     0.504\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     29/59     4.95G      1.59     0.549      0.38      2.52         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.88it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.442     0.666     0.532     0.491\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     30/59     4.95G      1.51     0.531     0.352      2.39         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.88it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.467     0.667     0.575     0.505\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     31/59     4.95G       1.5     0.524     0.356      2.38         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.89it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.468     0.672      0.56     0.509\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     32/59     4.95G      1.44     0.503     0.323      2.26         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.91it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.495     0.668     0.576     0.533\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     33/59     4.95G      1.43      0.49     0.315      2.23         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.90it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.448     0.685     0.574     0.511\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     34/59     4.95G       1.4     0.479     0.302      2.19         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.91it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.534     0.672     0.574     0.543\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     35/59     4.95G      1.35     0.469     0.298      2.12         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.95it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.49     0.667     0.599     0.551\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     36/59     4.95G      1.36     0.464     0.275      2.09         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.96it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.475     0.672     0.602     0.545\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     37/59     4.95G      1.33     0.455     0.274      2.05         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.98it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.488     0.658     0.574     0.528\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     38/59     4.95G      1.32     0.454     0.258      2.03         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.96it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.564     0.659     0.613      0.58\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     39/59     4.95G      1.28     0.446     0.253      1.98         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.99it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.504     0.673     0.609     0.557\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     40/59     4.95G      1.25     0.438     0.237      1.92         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.99it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.544     0.674     0.627     0.585\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     41/59     4.95G      1.22     0.429     0.218      1.87         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.99it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.527     0.684      0.62     0.579\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     42/59     4.95G      1.25     0.426     0.223       1.9         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.98it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.495     0.678      0.61     0.552\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     43/59     4.95G      1.21     0.422     0.213      1.84         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.98it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.565     0.692     0.644     0.604\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     44/59     4.95G      1.22     0.423     0.203      1.85         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.98it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.502     0.702     0.622     0.554\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     45/59     4.95G      1.25     0.427     0.206      1.88         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:54<00:00,  3.96it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.528     0.705     0.638     0.591\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     46/59     4.95G      1.27     0.428     0.201       1.9         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.87it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.48     0.688     0.608     0.554\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     47/59     4.95G      1.28     0.429     0.204      1.92         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.94it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.51     0.688     0.624     0.572\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     48/59     4.95G      1.22     0.418     0.201      1.84         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.93it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.55     0.695     0.642      0.59\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     49/59     4.95G      1.18     0.415     0.192      1.79         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.89it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.532     0.712     0.645     0.595\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     50/59     4.95G       1.2     0.417     0.192      1.81         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.90it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.527     0.706     0.644     0.593\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     51/59     4.95G      1.24     0.414     0.192      1.85         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.91it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.537       0.7     0.647     0.596\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     52/59     4.95G      1.25      0.41     0.195      1.86         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.91it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.554     0.698     0.643     0.599\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     53/59     4.95G       1.2     0.401     0.194       1.8         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.90it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.55     0.697     0.649     0.602\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     54/59     4.95G      1.16     0.389     0.185      1.73         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.87it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.554     0.697     0.649     0.603\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     55/59     4.95G      1.13     0.384     0.185       1.7         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.87it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.568     0.696     0.651     0.609\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     56/59     4.95G       1.1     0.381      0.18      1.66         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.85it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.568     0.697     0.653     0.609\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     57/59     4.95G       1.1     0.379     0.178      1.66         0       512: 100% 405/405 [04:36<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.89it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.56     0.696     0.648     0.604\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     58/59     4.95G      1.07     0.375     0.175      1.62         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:55<00:00,  3.90it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.563     0.696     0.652     0.609\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     59/59     4.95G      1.03     0.368     0.172      1.57         0       512: 100% 405/405 [04:35<00:00,  1.47it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:56<00:00,  3.87it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.569     0.708     0.648     0.618\n",
            "60 epochs completed in 5.613 hours.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wEIQTBQmOiYt",
        "colab_type": "text"
      },
      "source": [
        "**Performance Graphs**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3NgOyNCXXNgD",
        "colab_type": "code",
        "outputId": "22c70fd3-bb0a-4879-88e7-dbb29c908a58",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 414
        }
      },
      "source": [
        "from utils import utils;\n",
        "utils.plot_results()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qPMrUndWLhVH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#load already trained model for 60 epochs\n",
        "%cp '/content/drive/My Drive/SIMS_Dataset/yolov3 spp/60 epochs/last_from_yolov3_spp.pt' /content/yolov3/weights"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OXrQGWd4P25L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%cp /content/yolov3/weights/last_from_yolov3_spp.pt '/content/drive/My Drive/SIMS_Dataset'\n",
        "%cp /content/yolov3/weights/best_from_yolov3_spp.pt '/content/drive/My Drive/SIMS_Dataset'"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8cDEJRYL3vPM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load the TensorBoard notebook extension\n",
        "%load_ext tensorboard\n",
        "%tensorboard --logdir=runs"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8wDcKSIdOsOS",
        "colab_type": "text"
      },
      "source": [
        "**Test, Detect and Save Images**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8TyL8C4aOpUH",
        "colab_type": "code",
        "outputId": "0be2e61f-417e-4c55-b6da-fe760c5463b7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 175
        }
      },
      "source": [
        "!python3 detect.py --names Dataset/sims_classes.names --cfg yolov3-spp.cfg --img-size 1024 --weights weights/last_from_yolov3_spp.pt"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(agnostic_nms=False, augment=False, cfg='yolov3-spp.cfg', classes=None, conf_thres=0.3, device='', fourcc='mp4v', half=False, img_size=1024, iou_thres=0.6, names='Dataset/sims_classes.names', output='output', save_txt=False, source='data/samples', view_img=False, weights='weights/last_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "image 1/2 data/samples/0073.jpg: 768x1024 1 airliners, Done. (0.093s)\n",
            "image 2/2 data/samples/2644.jpg: 768x1024 2 cars, 1 vans, Done. (0.063s)\n",
            "Results saved to /content/yolov3/output\n",
            "Done. (0.360s)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XiH3Lc5mjVg4",
        "colab_type": "code",
        "outputId": "2f17ab46-a673-4f98-c787-839e5c626611",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        }
      },
      "source": [
        "#test on last model\n",
        "!python3 test.py --data Dataset/sims_test.txt --cfg yolov3-spp.cfg --batch-size 8 --weights weights/last_from_yolov3_spp.pt --save-json --img-size 512"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(augment=False, batch_size=8, cfg='yolov3-spp.cfg', conf_thres=0.001, data='Dataset/sims_test.txt', device='', img_size=512, iou_thres=0.6, save_json=True, single_cls=False, task='test', weights='weights/last_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Fusing layers...\n",
            "Model Summary: 152 layers, 6.26219e+07 parameters, 6.26219e+07 gradients\n",
            "Caching labels (747 found, 0 missing, 0 empty, 0 duplicate, for 748 images): 100% 748/748 [00:00<00:00, 1285.03it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 94/94 [00:22<00:00,  4.18it/s]\n",
            "                 all       748  7.98e+03      0.56     0.705     0.656     0.614\n",
            "                 car       748  3.68e+03     0.776     0.864     0.867     0.818\n",
            "               truck       748       446     0.546      0.78     0.724     0.642\n",
            "                 van       748       874     0.632     0.779     0.747     0.698\n",
            "         longvehicle       748       222     0.472     0.842     0.633     0.605\n",
            "                 bus       748       366     0.762     0.877      0.86     0.816\n",
            "            airliner       748       161     0.926     0.969     0.966     0.947\n",
            "           propeller       748        25     0.517      0.88      0.88     0.651\n",
            "             trainer       748        49     0.543      0.98     0.924     0.699\n",
            "           chartered       748       103     0.636     0.961     0.927     0.765\n",
            "             fighter       748         8     0.787      0.75     0.745     0.768\n",
            "               other       748        85    0.0967     0.247     0.111     0.139\n",
            "          stairtruck       748        83     0.309     0.313     0.243     0.311\n",
            "       pushbacktruck       748        55     0.217     0.145     0.132     0.174\n",
            "          helicopter       748         9     0.303     0.222      0.12     0.256\n",
            "                boat       748  1.82e+03     0.884     0.962     0.957     0.921\n",
            "Speed: 16.7/2.4/19.2 ms inference/NMS/total per 512x512 image at batch-size 8\n",
            "\n",
            "COCO mAP with pycocotools...\n",
            "Traceback (most recent call last):\n",
            "  File \"test.py\", line 257, in <module>\n",
            "    opt.augment)\n",
            "  File \"test.py\", line 211, in test\n",
            "    cocoGt = COCO(glob.glob('../coco/annotations/instances_val*.json')[0])  # initialize COCO ground truth api\n",
            "IndexError: list index out of range\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FrYEM3XvcyLK",
        "colab_type": "code",
        "outputId": "fd83e46a-0230-4d43-8b50-669bd01b8d81",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 469
        }
      },
      "source": [
        "#test on best model\n",
        "!python3 test.py --data Dataset/sims_test.txt --task test --cfg yolov3-spp.cfg --batch-size 8 --weights weights/best_from_yolov3_spp.pt --save-json --img-size 512"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(augment=False, batch_size=8, cfg='yolov3-spp.cfg', conf_thres=0.001, data='Dataset/sims_test.txt', device='', img_size=512, iou_thres=0.6, save_json=True, single_cls=False, task='test', weights='weights/best_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla T4', total_memory=15079MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Fusing layers...\n",
            "Model Summary: 152 layers, 6.26219e+07 parameters, 6.26219e+07 gradients\n",
            "Caching labels (747 found, 0 missing, 0 empty, 0 duplicate, for 748 images): 100% 748/748 [00:00<00:00, 8451.72it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 94/94 [00:23<00:00,  4.04it/s]\n",
            "                 all       748  7.98e+03     0.547      0.69      0.66       0.6\n",
            "                 car       748  3.68e+03     0.773     0.872     0.867      0.82\n",
            "               truck       748       446     0.558     0.776     0.723     0.649\n",
            "                 van       748       874     0.634     0.783     0.746       0.7\n",
            "         longvehicle       748       222     0.477     0.851     0.629     0.611\n",
            "                 bus       748       366     0.778      0.87     0.864     0.821\n",
            "            airliner       748       161     0.906     0.969      0.97     0.936\n",
            "           propeller       748        25     0.474      0.88     0.884     0.616\n",
            "             trainer       748        49     0.573      0.98      0.92     0.723\n",
            "           chartered       748       103      0.62     0.965     0.924     0.755\n",
            "             fighter       748         8     0.912      0.75     0.809     0.823\n",
            "               other       748        85    0.0977     0.235     0.109     0.138\n",
            "          stairtruck       748        83     0.299     0.313     0.248     0.306\n",
            "       pushbacktruck       748        55     0.214     0.145     0.144     0.173\n",
            "          helicopter       748         9         0         0     0.111         0\n",
            "                boat       748  1.82e+03     0.884     0.963     0.957     0.922\n",
            "Speed: 18.5/2.5/21.0 ms inference/NMS/total per 512x512 image at batch-size 8\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r4z0NFFOiGIX",
        "colab_type": "text"
      },
      "source": [
        "**Resume Training after 60 epochs**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xXyP7Hc_7hbX",
        "colab_type": "code",
        "outputId": "8190ae94-6e9b-4a31-91cf-7bcc8232b3cf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "#Resume\n",
        "#start training\n",
        "!python3 train.py --resume --batch 8 --epochs 90 --img-size 512 --data Dataset/sims.txt --cache-images --rec --cfg yolov3-spp.cfg --name from_yolov3_spp --weights weights/last.pt\n",
        "#--img-size 1024 --cache-images --nosave --evolve"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Apex recommended for faster mixed precision training: https://github.com/NVIDIA/apex\n",
            "Namespace(adam=False, batch_size=8, bucket='', cache_images=True, cfg='yolov3-spp.cfg', data='Dataset/sims.txt', device='', epochs=90, evolve=False, img_size=[512], multi_scale=False, name='from_yolov3_spp', nosave=False, notest=False, rect=True, resume=True, single_cls=False, weights='weights/last.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla P100-PCIE-16GB', total_memory=16280MB)\n",
            "\n",
            "Start Tensorboard with \"tensorboard --logdir=runs\", view at http://localhost:6006/\n",
            "2020-05-08 02:54:20.340731: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Optimizer groups: 76 .bias, 76 Conv2d.weight, 73 other\n",
            "Caching labels (3232 found, 0 missing, 0 empty, 0 duplicate, for 3233 images): 100% 3233/3233 [00:00<00:00, 7916.63it/s]\n",
            "Caching images (1.9GB): 100% 3233/3233 [00:37<00:00, 85.44it/s]\n",
            "Caching labels (1740 found, 0 missing, 0 empty, 0 duplicate, for 1741 images): 100% 1741/1741 [00:00<00:00, 7750.66it/s]\n",
            "Caching images (1.0GB): 100% 1741/1741 [00:19<00:00, 88.79it/s]\n",
            "Image sizes 512 - 512 train, 512 test\n",
            "Using 2 dataloader workers\n",
            "Starting training for 90 epochs...\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     60/89     4.94G       1.9     0.518     0.314      2.73         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:42<00:00,  5.09it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.458     0.643     0.543     0.509\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     61/89     4.94G      1.35     0.457     0.319      2.13         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:41<00:00,  5.29it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.516     0.671     0.628     0.561\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     62/89     4.94G      1.13     0.399     0.221      1.75         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.34it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.503     0.687     0.625     0.563\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     63/89     4.94G      1.06     0.376     0.188      1.62         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.527     0.699     0.639     0.582\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     64/89     4.94G     0.985     0.359     0.176      1.52         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.39it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.547       0.7     0.651     0.601\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     65/89     4.94G     0.988     0.359     0.168      1.52         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.38it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.551     0.712     0.659     0.604\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     66/89     4.94G     0.945     0.345     0.166      1.46         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.552     0.708     0.642     0.598\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     67/89     4.94G     0.897     0.339     0.155      1.39         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.35it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.536     0.715     0.656     0.599\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     68/89     4.94G     0.909     0.337     0.149       1.4         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.35it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.521     0.706     0.659     0.591\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     69/89     4.94G     0.882     0.331     0.153      1.37         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.40it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.522     0.719     0.653     0.595\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     70/89     4.94G     0.907     0.336     0.146      1.39         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.519     0.712      0.65     0.592\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     71/89     4.94G     0.925     0.339     0.148      1.41         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.38it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.531     0.701     0.642     0.595\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     72/89     4.94G      0.92     0.337     0.147       1.4         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.40it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.544     0.714     0.651     0.609\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     73/89     4.94G     0.957     0.341     0.145      1.44         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.42it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.546     0.704     0.641     0.595\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     74/89     4.94G     0.989     0.344     0.151      1.48         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.44it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.568     0.699     0.652      0.61\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     75/89     4.94G      1.03     0.355     0.153      1.54         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.42it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.552     0.698     0.633     0.603\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     76/89     4.94G      1.11     0.368     0.157      1.63         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.587     0.693     0.644     0.611\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     77/89     4.94G      1.06     0.358     0.157      1.58         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.573     0.706     0.652     0.617\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     78/89     4.94G      1.04     0.354     0.155      1.55         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.604     0.702     0.662     0.629\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     79/89     4.94G      1.04     0.355     0.156      1.55         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.35it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.569     0.716     0.663     0.622\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     80/89     4.94G      1.04     0.354     0.158      1.55         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.36it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.608     0.705     0.666     0.636\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     81/89     4.94G      1.01     0.348     0.154      1.51         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.39it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.571     0.708     0.659     0.618\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     82/89     4.94G         1     0.344     0.154       1.5         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.37it/s]\n",
            "                 all  1.74e+03  2.12e+04      0.61     0.713     0.669     0.643\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     83/89     4.94G     0.996     0.345     0.152      1.49         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.40it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.583     0.707     0.661     0.627\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     84/89     4.94G     0.986     0.342     0.151      1.48         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.38it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.592     0.704     0.665      0.63\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     85/89     4.94G      0.95     0.333     0.149      1.43         0       512: 100% 405/405 [02:53<00:00,  2.34it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.42it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.574     0.706     0.659     0.621\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     86/89     4.94G     0.916     0.327     0.145      1.39         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.38it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.597     0.709     0.666     0.636\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     87/89     4.94G     0.902     0.324     0.141      1.37         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.38it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.589      0.71     0.662     0.633\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     88/89     4.94G     0.894     0.322     0.139      1.36         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:40<00:00,  5.40it/s]\n",
            "                 all  1.74e+03  2.12e+04       0.6      0.71     0.668     0.639\n",
            "\n",
            "     Epoch   gpu_mem      GIoU       obj       cls     total   targets  img_size\n",
            "     89/89     4.94G     0.873     0.317     0.138      1.33         0       512: 100% 405/405 [02:53<00:00,  2.33it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 218/218 [00:41<00:00,  5.31it/s]\n",
            "                 all  1.74e+03  2.12e+04     0.594     0.712      0.67     0.638\n",
            "30 epochs completed in 1.805 hours.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j_VwsAjticEE",
        "colab_type": "code",
        "outputId": "b3cb5fa5-54fb-4072-b62b-18bc51f0ea3c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 415
        }
      },
      "source": [
        "from utils import utils;\n",
        "utils.plot_results()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x432 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n7ApQCxPnBjx",
        "colab_type": "code",
        "outputId": "780c40f6-04da-4b53-ef61-cb650e82e121",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        }
      },
      "source": [
        "#test on last model\n",
        "!python3 test.py --data Dataset/sims_test.txt --cfg yolov3-spp.cfg --batch-size 8 --weights weights/last_from_yolov3_spp.pt --save-json --img-size 512"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(augment=False, batch_size=8, cfg='yolov3-spp.cfg', conf_thres=0.001, data='Dataset/sims_test.txt', device='', img_size=512, iou_thres=0.6, save_json=True, single_cls=False, task='test', weights='weights/last_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla P100-PCIE-16GB', total_memory=16280MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Fusing layers...\n",
            "Model Summary: 152 layers, 6.26219e+07 parameters, 6.26219e+07 gradients\n",
            "Caching labels (747 found, 0 missing, 0 empty, 0 duplicate, for 748 images): 100% 748/748 [00:00<00:00, 1374.89it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 94/94 [00:17<00:00,  5.24it/s]\n",
            "                 all       748  7.98e+03     0.586     0.712     0.679     0.636\n",
            "                 car       748  3.68e+03     0.785     0.856     0.859     0.819\n",
            "               truck       748       446      0.57     0.767     0.727     0.654\n",
            "                 van       748       874     0.643     0.778     0.743     0.704\n",
            "         longvehicle       748       222     0.488     0.833     0.632     0.615\n",
            "                 bus       748       366     0.757     0.883     0.864     0.815\n",
            "            airliner       748       161     0.904     0.969     0.968     0.935\n",
            "           propeller       748        25     0.658      0.88     0.903     0.753\n",
            "             trainer       748        49     0.687      0.98     0.941     0.807\n",
            "           chartered       748       103     0.635     0.963     0.927     0.765\n",
            "             fighter       748         8      0.75      0.75     0.821      0.75\n",
            "               other       748        85    0.0909     0.235    0.0898     0.131\n",
            "          stairtruck       748        83      0.32     0.283      0.22       0.3\n",
            "       pushbacktruck       748        55      0.17     0.109     0.115     0.133\n",
            "          helicopter       748         9     0.437     0.432     0.421     0.434\n",
            "                boat       748  1.82e+03       0.9     0.955     0.952     0.927\n",
            "Speed: 9.1/2.6/11.8 ms inference/NMS/total per 512x512 image at batch-size 8\n",
            "\n",
            "COCO mAP with pycocotools...\n",
            "Traceback (most recent call last):\n",
            "  File \"test.py\", line 257, in <module>\n",
            "    opt.augment)\n",
            "  File \"test.py\", line 211, in test\n",
            "    cocoGt = COCO(glob.glob('../coco/annotations/instances_val*.json')[0])  # initialize COCO ground truth api\n",
            "IndexError: list index out of range\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i8OW66a7vuqM",
        "colab_type": "code",
        "outputId": "be12d250-9aaa-4b3e-c23c-baa0748b204e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        }
      },
      "source": [
        "#test on best model\n",
        "!python3 test.py --data Dataset/sims_test.txt --cfg yolov3-spp.cfg --batch-size 8 --weights weights/best_from_yolov3_spp.pt --save-json --img-size 512"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(augment=False, batch_size=8, cfg='yolov3-spp.cfg', conf_thres=0.001, data='Dataset/sims_test.txt', device='', img_size=512, iou_thres=0.6, save_json=True, single_cls=False, task='test', weights='weights/best_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla P100-PCIE-16GB', total_memory=16280MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "Fusing layers...\n",
            "Model Summary: 152 layers, 6.26219e+07 parameters, 6.26219e+07 gradients\n",
            "Caching labels (747 found, 0 missing, 0 empty, 0 duplicate, for 748 images): 100% 748/748 [00:00<00:00, 7106.25it/s]\n",
            "               Class    Images   Targets         P         R   mAP@0.5        F1: 100% 94/94 [00:17<00:00,  5.31it/s]\n",
            "                 all       748  7.98e+03     0.603     0.714     0.677     0.644\n",
            "                 car       748  3.68e+03     0.791     0.854     0.856     0.821\n",
            "               truck       748       446     0.566      0.76     0.724     0.649\n",
            "                 van       748       874     0.641     0.785      0.74     0.706\n",
            "         longvehicle       748       222     0.501     0.838     0.641     0.627\n",
            "                 bus       748       366     0.766     0.874     0.867     0.817\n",
            "            airliner       748       161     0.913     0.969     0.969      0.94\n",
            "           propeller       748        25      0.62      0.88     0.893     0.727\n",
            "             trainer       748        49     0.691      0.98     0.929      0.81\n",
            "           chartered       748       103     0.612     0.981     0.918     0.753\n",
            "             fighter       748         8         1     0.848     0.911     0.918\n",
            "               other       748        85    0.0841     0.224    0.0853     0.122\n",
            "          stairtruck       748        83     0.346     0.301     0.225     0.322\n",
            "       pushbacktruck       748        55     0.179     0.127     0.124     0.149\n",
            "          helicopter       748         9     0.442     0.333     0.314      0.38\n",
            "                boat       748  1.82e+03     0.898     0.953     0.952     0.924\n",
            "Speed: 9.2/2.6/11.8 ms inference/NMS/total per 512x512 image at batch-size 8\n",
            "\n",
            "COCO mAP with pycocotools...\n",
            "Traceback (most recent call last):\n",
            "  File \"test.py\", line 257, in <module>\n",
            "    opt.augment)\n",
            "  File \"test.py\", line 211, in test\n",
            "    cocoGt = COCO(glob.glob('../coco/annotations/instances_val*.json')[0])  # initialize COCO ground truth api\n",
            "IndexError: list index out of range\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xNkVgJletUdA",
        "colab_type": "code",
        "outputId": "a1ea1ac6-7b5f-494e-a4bd-56af2103f304",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 244
        }
      },
      "source": [
        "!python3 detect.py --names Dataset/sims_classes.names --cfg yolov3-spp.cfg --img-size 1024 --weights '/content/drive/My Drive/SIMS_Dataset/yolov3 spp/best_from_yolov3_spp.pt'"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Namespace(agnostic_nms=False, augment=False, cfg='yolov3-spp.cfg', classes=None, conf_thres=0.3, device='', fourcc='mp4v', half=False, img_size=1024, iou_thres=0.6, names='Dataset/sims_classes.names', output='output', save_txt=False, source='data/samples', view_img=False, weights='/content/drive/My Drive/SIMS_Dataset/yolov3 spp/best_from_yolov3_spp.pt')\n",
            "Using CUDA device0 _CudaDeviceProperties(name='Tesla P100-PCIE-16GB', total_memory=16280MB)\n",
            "\n",
            "Model Summary: 225 layers, 6.26487e+07 parameters, 6.26487e+07 gradients\n",
            "image 1/6 data/samples/3716.jpg: 768x1024 11 cars, 1 trucks, 1 vans, Done. (0.044s)\n",
            "image 2/6 data/samples/3732.jpg: 768x1024 22 cars, 1 trucks, 6 vans, Done. (0.041s)\n",
            "image 3/6 data/samples/3792.jpg: 768x1024 7 cars, 1 trucks, 2 vans, Done. (0.041s)\n",
            "image 4/6 data/samples/3820.jpg: 768x1024 Done. (0.041s)\n",
            "image 5/6 data/samples/4738.jpg: 768x1024 21 cars, 3 trucks, 11 vans, 1 others, Done. (0.041s)\n",
            "image 6/6 data/samples/4785.jpg: 768x1024 1 longvehicles, 2 buss, Done. (0.041s)\n",
            "Results saved to /content/yolov3/output\n",
            "Done. (0.514s)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2VxJ6qh03V7c",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}