UNeXt_Project_Exp1.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "01OdGhmjw3Mv"
      },
      "outputs": [],
      "source": [
        "# %%bash\n",
        "# git clone https://github.com/jeya-maria-jose/UNeXt-pytorch\n",
        "# cd UNeXt-pytorch\n",
        "\n",
        "# pip install -q condacolab\n",
        "\n",
        "!cp -r /content/drive/MyDrive/UNeXt-pytorch ./"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N97hG6x_EQ5T",
        "outputId": "1d723b99-d731-440f-f83f-33733a42450a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "mkdir: cannot create directory ‘./drive/MyDrive/dataset’: File exists\n"
          ]
        }
      ],
      "source": [
        "!mkdir ./drive/MyDrive/dataset\n",
        "!mkdir ./Dataset\n",
        "!tar -xf /content/drive/MyDrive/Task02_Heart.tar -C ./Dataset\n",
        "!mkdir data\n",
        "!mkdir data/train\n",
        "!mkdir data/train/image\n",
        "!mkdir data/train/label\n",
        "!mkdir data/test\n",
        "!mkdir data/test/image\n",
        "!mkdir data/test/label\n",
        "\n",
        "import os\n",
        "\n",
        "import nibabel as nib\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib\n",
        "import numpy as np\n",
        "from glob import glob\n",
        "\n",
        "# Use 24, 26, 29, 30 dataset as test\n",
        "input_path = \"./Dataset/Task02_Heart/*Tr/\"\n",
        "\n",
        "testset_name = [\"la_024.nii.gz\", \"la_026.nii.gz\", \"la_029.nii.gz\", \"la_030.nii.gz\"]\n",
        "testset = [\"./Dataset/Task02_Heart/imagesTr/\" + data for data in testset_name]\n",
        "testset += [\"./Dataset/Task02_Heart/labelsTr/\" + data for data in testset_name]\n",
        "trainset = [data for data in glob(input_path + \"*.nii.gz\") if data not in testset]\n",
        "\n",
        "def save_png_from_nii(dataPath:list, train:bool):\n",
        "  path = \"./data/train/\" if train else \"./data/test/\"\n",
        "  for data in dataPath:\n",
        "    image = \"image/\" if \"images\" in data else \"label/\"\n",
        "    proxy = nib.load(data)\n",
        "    arr = proxy.get_fdata()\n",
        "    nii_image = nib.Nifti1Image(arr, proxy.affine, proxy.header)\n",
        "    data_number = int(data.split(\".\")[1][-3:])\n",
        "    count = arr.shape[2]\n",
        "    for i in range(count):\n",
        "      matplotlib.image.imsave(path + image + \"%d_%d.png\"%(data_number, i), arr[:,:,i])\n",
        "\n",
        "\n",
        "import os\n",
        "if not os.path.exists(\"./data/train/image/19_0.png\"):\n",
        "  save_png_from_nii(trainset, True)\n",
        "  save_png_from_nii(testset, False)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!mv ./data/train/label/* ./inputs/train/masks/0/"
      ],
      "metadata": {
        "id": "qodmv_mtQOpr",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f23f3a8e-a971-47bf-dce3-0b7013b2cc36"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "mv: target './inputs/train/masks/0/' is not a directory\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "SBe4b6Hv8zs-",
        "outputId": "201c2a47-e6f9-4c8a-c962-68e9a6607506"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/UNeXt-pytorch\n",
            "Collecting addict (from -r requirements.txt (line 1))\n",
            "  Downloading addict-2.4.0-py3-none-any.whl (3.8 kB)\n",
            "Collecting dataclasses (from -r requirements.txt (line 2))\n",
            "  Downloading dataclasses-0.6-py3-none-any.whl (14 kB)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 3)) (1.25.2)\n",
            "Requirement already satisfied: opencv-python in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 4)) (4.8.0.76)\n",
            "Collecting perceptual (from -r requirements.txt (line 5))\n",
            "  Downloading perceptual-0.1.tar.gz (5.6 kB)\n",
            "  Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "Requirement already satisfied: pillow in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 6)) (9.4.0)\n",
            "Requirement already satisfied: scikit-image in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 7)) (0.19.3)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 8)) (1.11.4)\n",
            "Requirement already satisfied: tifffile in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 9)) (2024.5.10)\n",
            "Collecting timm (from -r requirements.txt (line 10))\n",
            "  Downloading timm-1.0.3-py3-none-any.whl (2.3 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.3/2.3 MB\u001b[0m \u001b[31m56.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: torch in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 11)) (2.3.0+cu121)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 12)) (0.18.0+cu121)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from -r requirements.txt (line 13)) (4.11.0)\n",
            "Collecting yapf (from -r requirements.txt (line 14))\n",
            "  Downloading yapf-0.40.2-py3-none-any.whl (254 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m254.7/254.7 kB\u001b[0m \u001b[31m35.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: networkx>=2.2 in /usr/local/lib/python3.10/dist-packages (from scikit-image->-r requirements.txt (line 7)) (3.3)\n",
            "Requirement already satisfied: imageio>=2.4.1 in /usr/local/lib/python3.10/dist-packages (from scikit-image->-r requirements.txt (line 7)) (2.31.6)\n",
            "Requirement already satisfied: PyWavelets>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-image->-r requirements.txt (line 7)) (1.6.0)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from scikit-image->-r requirements.txt (line 7)) (24.0)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.10/dist-packages (from timm->-r requirements.txt (line 10)) (6.0.1)\n",
            "Requirement already satisfied: huggingface_hub in /usr/local/lib/python3.10/dist-packages (from timm->-r requirements.txt (line 10)) (0.23.1)\n",
            "Requirement already satisfied: safetensors in /usr/local/lib/python3.10/dist-packages (from timm->-r requirements.txt (line 10)) (0.4.3)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch->-r requirements.txt (line 11)) (3.14.0)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch->-r requirements.txt (line 11)) (1.12)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch->-r requirements.txt (line 11)) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch->-r requirements.txt (line 11)) (2023.6.0)\n",
            "Collecting nvidia-cuda-nvrtc-cu12==12.1.105 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cuda_nvrtc_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (23.7 MB)\n",
            "Collecting nvidia-cuda-runtime-cu12==12.1.105 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cuda_runtime_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (823 kB)\n",
            "Collecting nvidia-cuda-cupti-cu12==12.1.105 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cuda_cupti_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (14.1 MB)\n",
            "Collecting nvidia-cudnn-cu12==8.9.2.26 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cudnn_cu12-8.9.2.26-py3-none-manylinux1_x86_64.whl (731.7 MB)\n",
            "Collecting nvidia-cublas-cu12==12.1.3.1 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cublas_cu12-12.1.3.1-py3-none-manylinux1_x86_64.whl (410.6 MB)\n",
            "Collecting nvidia-cufft-cu12==11.0.2.54 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cufft_cu12-11.0.2.54-py3-none-manylinux1_x86_64.whl (121.6 MB)\n",
            "Collecting nvidia-curand-cu12==10.3.2.106 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_curand_cu12-10.3.2.106-py3-none-manylinux1_x86_64.whl (56.5 MB)\n",
            "Collecting nvidia-cusolver-cu12==11.4.5.107 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cusolver_cu12-11.4.5.107-py3-none-manylinux1_x86_64.whl (124.2 MB)\n",
            "Collecting nvidia-cusparse-cu12==12.1.0.106 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_cusparse_cu12-12.1.0.106-py3-none-manylinux1_x86_64.whl (196.0 MB)\n",
            "Collecting nvidia-nccl-cu12==2.20.5 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_nccl_cu12-2.20.5-py3-none-manylinux2014_x86_64.whl (176.2 MB)\n",
            "Collecting nvidia-nvtx-cu12==12.1.105 (from torch->-r requirements.txt (line 11))\n",
            "  Using cached nvidia_nvtx_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (99 kB)\n",
            "Requirement already satisfied: triton==2.3.0 in /usr/local/lib/python3.10/dist-packages (from torch->-r requirements.txt (line 11)) (2.3.0)\n",
            "Collecting nvidia-nvjitlink-cu12 (from nvidia-cusolver-cu12==11.4.5.107->torch->-r requirements.txt (line 11))\n",
            "  Downloading nvidia_nvjitlink_cu12-12.5.40-py3-none-manylinux2014_x86_64.whl (21.3 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m21.3/21.3 MB\u001b[0m \u001b[31m32.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: importlib-metadata>=6.6.0 in /usr/local/lib/python3.10/dist-packages (from yapf->-r requirements.txt (line 14)) (7.1.0)\n",
            "Requirement already satisfied: platformdirs>=3.5.1 in /usr/local/lib/python3.10/dist-packages (from yapf->-r requirements.txt (line 14)) (4.2.2)\n",
            "Requirement already satisfied: tomli>=2.0.1 in /usr/local/lib/python3.10/dist-packages (from yapf->-r requirements.txt (line 14)) (2.0.1)\n",
            "Requirement already satisfied: zipp>=0.5 in /usr/local/lib/python3.10/dist-packages (from importlib-metadata>=6.6.0->yapf->-r requirements.txt (line 14)) (3.18.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface_hub->timm->-r requirements.txt (line 10)) (2.31.0)\n",
            "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/dist-packages (from huggingface_hub->timm->-r requirements.txt (line 10)) (4.66.4)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch->-r requirements.txt (line 11)) (2.1.5)\n",
            "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch->-r requirements.txt (line 11)) (1.3.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface_hub->timm->-r requirements.txt (line 10)) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface_hub->timm->-r requirements.txt (line 10)) (3.7)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface_hub->timm->-r requirements.txt (line 10)) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface_hub->timm->-r requirements.txt (line 10)) (2024.2.2)\n",
            "Building wheels for collected packages: perceptual\n",
            "  Building wheel for perceptual (setup.py) ... \u001b[?25l\u001b[?25hdone\n",
            "  Created wheel for perceptual: filename=perceptual-0.1-py2.py3-none-any.whl size=5620 sha256=de8d7b176c5756ac5046d1f811a665a8d5cb21bd229d46816a565fba9201004b\n",
            "  Stored in directory: /root/.cache/pip/wheels/8d/2c/4c/7fbe9795a820e3602e020192dc725bf676573e2bd70467e149\n",
            "Successfully built perceptual\n",
            "Installing collected packages: perceptual, dataclasses, addict, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufft-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, yapf, nvidia-cusparse-cu12, nvidia-cudnn-cu12, nvidia-cusolver-cu12, timm\n",
            "Successfully installed addict-2.4.0 dataclasses-0.6 nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-nccl-cu12-2.20.5 nvidia-nvjitlink-cu12-12.5.40 nvidia-nvtx-cu12-12.1.105 perceptual-0.1 timm-1.0.3 yapf-0.40.2\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "dataclasses"
                ]
              },
              "id": "abd71d2cc4df458a93b560d63cbe9af0"
            }
          },
          "metadata": {}
        }
      ],
      "source": [
        "%cd UNeXt-pytorch\n",
        "\n",
        "test= \"addict\\n\\\n",
        "dataclasses\\n\\\n",
        "numpy\\n\\\n",
        "opencv-python\\n\\\n",
        "perceptual\\n\\\n",
        "pillow\\n\\\n",
        "scikit-image\\n\\\n",
        "scipy\\n\\\n",
        "tifffile\\n\\\n",
        "timm\\n\\\n",
        "torch\\n\\\n",
        "torchvision\\n\\\n",
        "typing-extensions\\n\\\n",
        "yapf\"\n",
        "\n",
        "open(\"./requirements.txt\", \"w\").write(test)\n",
        "\n",
        "\n",
        "\n",
        "!pip install -r requirements.txt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dpGPEbT93dYl",
        "outputId": "a0f59fec-e6f4-4a0d-b8a8-98bc89582666"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Errno 2] No such file or directory: 'UNeXt-pytorch'\n",
            "/content/UNeXt-pytorch\n",
            "mkdir: cannot create directory ‘/content/data/train/masks’: No such file or directory\n",
            "mkdir: cannot create directory ‘/content/data/train/masks/0’: No such file or directory\n",
            "mkdir: cannot create directory ‘/content/data/test/masks’: No such file or directory\n",
            "mkdir: cannot create directory ‘/content/data/test/masks/0’: No such file or directory\n",
            "mv: cannot stat '/content/data/train/image': No such file or directory\n",
            "mv: cannot stat '/content/data/train/label/*': No such file or directory\n",
            "mv: cannot stat '/content/data/test/image': No such file or directory\n",
            "mv: cannot stat '/content/data/test/label/*': No such file or directory\n",
            "mv: cannot stat '/content/data': No such file or directory\n"
          ]
        }
      ],
      "source": [
        "%cd UNeXt-pytorch\n",
        "!rm -rf ./inputs\n",
        "!mkdir /content/data/train/masks /content/data/train/masks/0\n",
        "!mkdir /content/data/test/masks /content/data/test/masks/0\n",
        "!mv /content/data/train/image /content/data/train/images\n",
        "!mv /content/data/train/label/* /content/data/train/masks/0\n",
        "!mv /content/data/test/image /content/data/test/images\n",
        "!mv /content/data/test/label/* /content/data/test/masks/0\n",
        "!mv /content/data /content/UNeXt-pytorch\n",
        "!mkdir ./inputs\n",
        "!mv ./data/train ./inputs\n",
        "!mv ./data/test ./inputs"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "E-R9So5AqomU",
        "outputId": "610538fc-1c2c-452f-b94a-e308144fec0c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "mv: cannot stat './inputs/train/masks/0/label/*': No such file or directory\n",
            "mv: cannot stat './inputs/test/masks/0/*': No such file or directory\n"
          ]
        }
      ],
      "source": [
        "# !mv ./inputs/train/masks/0/label/* ./inputs/train/masks/0/\n",
        "# !mv ./inputs/test/masks/0/* ./inputs/train/masks/0/\n",
        "# !mv ./inputs/test/images/* ./inputs/train/images/"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WTCAqdE94YNM",
        "outputId": "1567a89c-5b87-43fb-f1d3-26a8e21be57f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Errno 2] No such file or directory: 'UNeXt-pytorch'\n",
            "/content/UNeXt-pytorch\n",
            "--------------------\n",
            "name: segment_unet\n",
            "epochs: 50\n",
            "batch_size: 16\n",
            "arch: UNet\n",
            "deep_supervision: False\n",
            "input_channels: 3\n",
            "num_classes: 1\n",
            "input_w: 256\n",
            "input_h: 256\n",
            "loss: BCEDiceLoss\n",
            "dataset: train\n",
            "img_ext: .png\n",
            "mask_ext: .png\n",
            "optimizer: Adam\n",
            "lr: 0.0001\n",
            "momentum: 0.9\n",
            "weight_decay: 0.0001\n",
            "nesterov: False\n",
            "scheduler: CosineAnnealingLR\n",
            "min_lr: 1e-05\n",
            "factor: 0.1\n",
            "patience: 2\n",
            "milestones: 1,2\n",
            "gamma: 0.6666666666666666\n",
            "early_stopping: -1\n",
            "cfg: None\n",
            "num_workers: 4\n",
            "--------------------\n",
            "--------------------\n",
            "Train dataset size : 1449, Valid dataset size : 363\n",
            "--------------------\n",
            "Epoch [0/50]\n",
            "100% 90/90 [00:37<00:00,  2.40it/s, loss=1.07, iou=0.196]\n",
            "100% 23/23 [00:04<00:00,  4.91it/s, loss=1.05, iou=2.53e-9, dice=5.07e-9]\n",
            "loss 1.0670 - iou 0.1958 - val_loss 1.0541 - val_iou 0.0000\n",
            "=> saved best model\n",
            "Epoch [1/50]\n",
            "100% 90/90 [00:28<00:00,  3.12it/s, loss=1.05, iou=0.374]\n",
            "100% 23/23 [00:02<00:00,  8.81it/s, loss=1.05, iou=0.337, dice=0.492]\n",
            "loss 1.0462 - iou 0.3740 - val_loss 1.0450 - val_iou 0.3374\n",
            "=> saved best model\n",
            "Epoch [2/50]\n",
            "100% 90/90 [00:29<00:00,  3.02it/s, loss=1.04, iou=0.456]\n",
            "100% 23/23 [00:02<00:00,  8.96it/s, loss=1.04, iou=0.539, dice=0.699]\n",
            "loss 1.0399 - iou 0.4562 - val_loss 1.0366 - val_iou 0.5388\n",
            "=> saved best model\n",
            "Epoch [3/50]\n",
            "100% 90/90 [00:29<00:00,  3.09it/s, loss=1.04, iou=0.506]\n",
            "100% 23/23 [00:02<00:00,  9.00it/s, loss=1.04, iou=0.301, dice=0.459]\n",
            "loss 1.0363 - iou 0.5063 - val_loss 1.0380 - val_iou 0.3005\n",
            "Epoch [4/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.543]\n",
            "100% 23/23 [00:02<00:00,  8.81it/s, loss=1.03, iou=0.355, dice=0.523]\n",
            "loss 1.0340 - iou 0.5434 - val_loss 1.0338 - val_iou 0.3553\n",
            "Epoch [5/50]\n",
            "100% 90/90 [00:29<00:00,  3.07it/s, loss=1.03, iou=0.587]\n",
            "100% 23/23 [00:02<00:00,  8.95it/s, loss=1.03, iou=0.727, dice=0.842]\n",
            "loss 1.0324 - iou 0.5871 - val_loss 1.0305 - val_iou 0.7274\n",
            "=> saved best model\n",
            "Epoch [6/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.598]\n",
            "100% 23/23 [00:02<00:00,  8.83it/s, loss=1.03, iou=0.675, dice=0.805]\n",
            "loss 1.0316 - iou 0.5980 - val_loss 1.0304 - val_iou 0.6754\n",
            "Epoch [7/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.619]\n",
            "100% 23/23 [00:02<00:00,  8.90it/s, loss=1.03, iou=0.724, dice=0.839]\n",
            "loss 1.0310 - iou 0.6193 - val_loss 1.0302 - val_iou 0.7239\n",
            "Epoch [8/50]\n",
            "100% 90/90 [00:29<00:00,  3.07it/s, loss=1.03, iou=0.625]\n",
            "100% 23/23 [00:02<00:00,  8.88it/s, loss=1.03, iou=0.651, dice=0.787]\n",
            "loss 1.0309 - iou 0.6251 - val_loss 1.0297 - val_iou 0.6505\n",
            "Epoch [9/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.637]\n",
            "100% 23/23 [00:02<00:00,  8.95it/s, loss=1.03, iou=0.753, dice=0.859]\n",
            "loss 1.0305 - iou 0.6371 - val_loss 1.0302 - val_iou 0.7534\n",
            "=> saved best model\n",
            "Epoch [10/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.664]\n",
            "100% 23/23 [00:02<00:00,  8.92it/s, loss=1.03, iou=0.727, dice=0.841]\n",
            "loss 1.0303 - iou 0.6639 - val_loss 1.0295 - val_iou 0.7268\n",
            "Epoch [11/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.688]\n",
            "100% 23/23 [00:02<00:00,  8.80it/s, loss=1.03, iou=0.743, dice=0.852]\n",
            "loss 1.0300 - iou 0.6881 - val_loss 1.0289 - val_iou 0.7428\n",
            "Epoch [12/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.694]\n",
            "100% 23/23 [00:02<00:00,  8.87it/s, loss=1.03, iou=0.764, dice=0.865]\n",
            "loss 1.0300 - iou 0.6937 - val_loss 1.0299 - val_iou 0.7637\n",
            "=> saved best model\n",
            "Epoch [13/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.682]\n",
            "100% 23/23 [00:02<00:00,  8.87it/s, loss=1.03, iou=0.746, dice=0.854]\n",
            "loss 1.0301 - iou 0.6825 - val_loss 1.0288 - val_iou 0.7462\n",
            "Epoch [14/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.707]\n",
            "100% 23/23 [00:02<00:00,  8.78it/s, loss=1.03, iou=0.639, dice=0.779]\n",
            "loss 1.0298 - iou 0.7069 - val_loss 1.0290 - val_iou 0.6393\n",
            "Epoch [15/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.708]\n",
            "100% 23/23 [00:02<00:00,  8.89it/s, loss=1.03, iou=0.811, dice=0.896]\n",
            "loss 1.0297 - iou 0.7078 - val_loss 1.0301 - val_iou 0.8115\n",
            "=> saved best model\n",
            "Epoch [16/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.714]\n",
            "100% 23/23 [00:02<00:00,  8.88it/s, loss=1.03, iou=0.785, dice=0.878]\n",
            "loss 1.0298 - iou 0.7141 - val_loss 1.0299 - val_iou 0.7849\n",
            "Epoch [17/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.687]\n",
            "100% 23/23 [00:02<00:00,  8.86it/s, loss=1.03, iou=0.781, dice=0.877]\n",
            "loss 1.0301 - iou 0.6875 - val_loss 1.0289 - val_iou 0.7813\n",
            "Epoch [18/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.706]\n",
            "100% 23/23 [00:02<00:00,  8.90it/s, loss=1.03, iou=0.783, dice=0.878]\n",
            "loss 1.0298 - iou 0.7056 - val_loss 1.0288 - val_iou 0.7830\n",
            "Epoch [19/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.729]\n",
            "100% 23/23 [00:02<00:00,  8.92it/s, loss=1.03, iou=0.803, dice=0.891]\n",
            "loss 1.0296 - iou 0.7294 - val_loss 1.0288 - val_iou 0.8034\n",
            "Epoch [20/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.743]\n",
            "100% 23/23 [00:02<00:00,  8.85it/s, loss=1.03, iou=0.787, dice=0.88]\n",
            "loss 1.0295 - iou 0.7426 - val_loss 1.0285 - val_iou 0.7868\n",
            "Epoch [21/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.748]\n",
            "100% 23/23 [00:02<00:00,  8.85it/s, loss=1.03, iou=0.77, dice=0.87]\n",
            "loss 1.0295 - iou 0.7479 - val_loss 1.0284 - val_iou 0.7703\n",
            "Epoch [22/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.751]\n",
            "100% 23/23 [00:02<00:00,  8.94it/s, loss=1.03, iou=0.764, dice=0.866]\n",
            "loss 1.0294 - iou 0.7507 - val_loss 1.0285 - val_iou 0.7644\n",
            "Epoch [23/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.755]\n",
            "100% 23/23 [00:02<00:00,  8.73it/s, loss=1.03, iou=0.773, dice=0.872]\n",
            "loss 1.0293 - iou 0.7552 - val_loss 1.0284 - val_iou 0.7730\n",
            "Epoch [24/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.749]\n",
            "100% 23/23 [00:02<00:00,  8.90it/s, loss=1.03, iou=0.79, dice=0.882]\n",
            "loss 1.0293 - iou 0.7493 - val_loss 1.0291 - val_iou 0.7905\n",
            "Epoch [25/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.743]\n",
            "100% 23/23 [00:02<00:00,  8.65it/s, loss=1.03, iou=0.79, dice=0.882]\n",
            "loss 1.0294 - iou 0.7428 - val_loss 1.0285 - val_iou 0.7902\n",
            "Epoch [26/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.772]\n",
            "100% 23/23 [00:02<00:00,  8.89it/s, loss=1.03, iou=0.81, dice=0.894]\n",
            "loss 1.0292 - iou 0.7724 - val_loss 1.0282 - val_iou 0.8097\n",
            "Epoch [27/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.776]\n",
            "100% 23/23 [00:02<00:00,  8.86it/s, loss=1.03, iou=0.781, dice=0.877]\n",
            "loss 1.0291 - iou 0.7759 - val_loss 1.0282 - val_iou 0.7814\n",
            "Epoch [28/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.775]\n",
            "100% 23/23 [00:02<00:00,  8.96it/s, loss=1.03, iou=0.807, dice=0.893]\n",
            "loss 1.0291 - iou 0.7752 - val_loss 1.0281 - val_iou 0.8071\n",
            "Epoch [29/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.785]\n",
            "100% 23/23 [00:02<00:00,  8.63it/s, loss=1.03, iou=0.817, dice=0.899]\n",
            "loss 1.0290 - iou 0.7846 - val_loss 1.0281 - val_iou 0.8174\n",
            "=> saved best model\n",
            "Epoch [30/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.79]\n",
            "100% 23/23 [00:02<00:00,  8.82it/s, loss=1.03, iou=0.837, dice=0.911]\n",
            "loss 1.0290 - iou 0.7896 - val_loss 1.0281 - val_iou 0.8373\n",
            "=> saved best model\n",
            "Epoch [31/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.78]\n",
            "100% 23/23 [00:02<00:00,  8.88it/s, loss=1.03, iou=0.824, dice=0.903]\n",
            "loss 1.0291 - iou 0.7804 - val_loss 1.0280 - val_iou 0.8242\n",
            "Epoch [32/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.793]\n",
            "100% 23/23 [00:02<00:00,  8.75it/s, loss=1.03, iou=0.794, dice=0.885]\n",
            "loss 1.0289 - iou 0.7929 - val_loss 1.0280 - val_iou 0.7944\n",
            "Epoch [33/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.794]\n",
            "100% 23/23 [00:02<00:00,  8.91it/s, loss=1.03, iou=0.862, dice=0.925]\n",
            "loss 1.0289 - iou 0.7941 - val_loss 1.0281 - val_iou 0.8615\n",
            "=> saved best model\n",
            "Epoch [34/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.795]\n",
            "100% 23/23 [00:02<00:00,  8.90it/s, loss=1.03, iou=0.833, dice=0.909]\n",
            "loss 1.0289 - iou 0.7954 - val_loss 1.0281 - val_iou 0.8334\n",
            "Epoch [35/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.796]\n",
            "100% 23/23 [00:02<00:00,  8.68it/s, loss=1.03, iou=0.847, dice=0.917]\n",
            "loss 1.0289 - iou 0.7962 - val_loss 1.0280 - val_iou 0.8469\n",
            "Epoch [36/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.809]\n",
            "100% 23/23 [00:02<00:00,  8.85it/s, loss=1.03, iou=0.84, dice=0.913]\n",
            "loss 1.0287 - iou 0.8095 - val_loss 1.0279 - val_iou 0.8399\n",
            "Epoch [37/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.809]\n",
            "100% 23/23 [00:02<00:00,  8.79it/s, loss=1.03, iou=0.782, dice=0.877]\n",
            "loss 1.0288 - iou 0.8089 - val_loss 1.0280 - val_iou 0.7815\n",
            "Epoch [38/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.814]\n",
            "100% 23/23 [00:02<00:00,  8.86it/s, loss=1.03, iou=0.82, dice=0.901]\n",
            "loss 1.0288 - iou 0.8138 - val_loss 1.0278 - val_iou 0.8197\n",
            "Epoch [39/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.812]\n",
            "100% 23/23 [00:02<00:00,  8.86it/s, loss=1.03, iou=0.857, dice=0.923]\n",
            "loss 1.0287 - iou 0.8117 - val_loss 1.0278 - val_iou 0.8569\n",
            "Epoch [40/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.817]\n",
            "100% 23/23 [00:02<00:00,  8.79it/s, loss=1.03, iou=0.838, dice=0.912]\n",
            "loss 1.0287 - iou 0.8168 - val_loss 1.0278 - val_iou 0.8385\n",
            "Epoch [41/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.821]\n",
            "100% 23/23 [00:02<00:00,  8.87it/s, loss=1.03, iou=0.843, dice=0.914]\n",
            "loss 1.0287 - iou 0.8209 - val_loss 1.0278 - val_iou 0.8425\n",
            "Epoch [42/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.827]\n",
            "100% 23/23 [00:02<00:00,  8.89it/s, loss=1.03, iou=0.839, dice=0.912]\n",
            "loss 1.0287 - iou 0.8268 - val_loss 1.0277 - val_iou 0.8390\n",
            "Epoch [43/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.829]\n",
            "100% 23/23 [00:02<00:00,  8.84it/s, loss=1.03, iou=0.859, dice=0.924]\n",
            "loss 1.0287 - iou 0.8288 - val_loss 1.0277 - val_iou 0.8589\n",
            "Epoch [44/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.833]\n",
            "100% 23/23 [00:02<00:00,  8.87it/s, loss=1.03, iou=0.865, dice=0.927]\n",
            "loss 1.0286 - iou 0.8332 - val_loss 1.0277 - val_iou 0.8648\n",
            "=> saved best model\n",
            "Epoch [45/50]\n",
            "100% 90/90 [00:29<00:00,  3.06it/s, loss=1.03, iou=0.837]\n",
            "100% 23/23 [00:02<00:00,  8.88it/s, loss=1.03, iou=0.843, dice=0.915]\n",
            "loss 1.0286 - iou 0.8370 - val_loss 1.0276 - val_iou 0.8432\n",
            "Epoch [46/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.839]\n",
            "100% 23/23 [00:02<00:00,  8.96it/s, loss=1.03, iou=0.861, dice=0.925]\n",
            "loss 1.0286 - iou 0.8387 - val_loss 1.0277 - val_iou 0.8613\n",
            "Epoch [47/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.837]\n",
            "100% 23/23 [00:02<00:00,  8.75it/s, loss=1.03, iou=0.873, dice=0.932]\n",
            "loss 1.0286 - iou 0.8366 - val_loss 1.0277 - val_iou 0.8730\n",
            "=> saved best model\n",
            "Epoch [48/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.843]\n",
            "100% 23/23 [00:02<00:00,  8.90it/s, loss=1.03, iou=0.833, dice=0.909]\n",
            "loss 1.0285 - iou 0.8433 - val_loss 1.0276 - val_iou 0.8328\n",
            "Epoch [49/50]\n",
            "100% 90/90 [00:29<00:00,  3.05it/s, loss=1.03, iou=0.841]\n",
            "100% 23/23 [00:02<00:00,  8.74it/s, loss=1.03, iou=0.879, dice=0.935]\n",
            "loss 1.0285 - iou 0.8413 - val_loss 1.0278 - val_iou 0.8789\n",
            "=> saved best model\n",
            "Total Train time : 1614.468410730362s.\n"
          ]
        }
      ],
      "source": [
        "%cd UNeXt-pytorch\n",
        "!python train.py --dataset train --arch UNet --name segment_unet --img_ext .png --mask_ext .png --lr 0.0001 --epochs 50 --input_w 256 --input_h 256 --b 16"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hrMCsl0Ntt6X",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4e389f4c-a50f-4652-d903-625bdeb593ec"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[Errno 2] No such file or directory: 'UNeXt-pytorch'\n",
            "/content/UNeXt-pytorch\n",
            "--------------------\n",
            "arch: UNet\n",
            "batch_size: 16\n",
            "cfg: None\n",
            "dataset: test\n",
            "deep_supervision: False\n",
            "early_stopping: -1\n",
            "epochs: 50\n",
            "factor: 0.1\n",
            "gamma: 0.6666666666666666\n",
            "img_ext: .png\n",
            "input_channels: 3\n",
            "input_h: 256\n",
            "input_w: 256\n",
            "loss: BCEDiceLoss\n",
            "lr: 0.0001\n",
            "mask_ext: .png\n",
            "milestones: 1,2\n",
            "min_lr: 1e-05\n",
            "momentum: 0.9\n",
            "name: segment_unet\n",
            "nesterov: False\n",
            "num_classes: 1\n",
            "num_workers: 4\n",
            "optimizer: Adam\n",
            "patience: 2\n",
            "scheduler: CosineAnnealingLR\n",
            "weight_decay: 0.0001\n",
            "--------------------\n",
            "=> creating model UNet\n",
            "100% 29/29 [00:08<00:00,  3.44it/s]\n",
            "IoU: 0.8052\n",
            "Dice: 0.8915\n"
          ]
        }
      ],
      "source": [
        "%cd UNeXt-pytorch\n",
        "!python val.py --name segment_unet"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import csv\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "unext = pd.read_csv(\"./models/segment/log.csv\")\n",
        "unet = pd.read_csv(\"./models/segment_unet/log.csv\")\n",
        "\n",
        "unext_training = unext['loss']\n",
        "unet_training = unet['loss']\n",
        "unext_val = unext['val_loss']\n",
        "unet_val = unet['val_loss']\n",
        "\n",
        "x = np.linspace(0, 50, 50)\n",
        "\n",
        "fig = plt.figure()\n",
        "plt.title(\"Training Losses for Segmentation\")\n",
        "plt.plot(x, unext_training)\n",
        "plt.plot(x, unet_training)\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend(('UNeXt - Train','UNet - Train'))\n",
        "\n",
        "\n",
        "fig2 = plt.figure()\n",
        "plt.title(\"Validation Losses for Segmentation\")\n",
        "plt.plot(x, unext_val)\n",
        "plt.plot(x, unet_val)\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend(('UNeXt - Val','UNet - Val'))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 945
        },
        "id": "AR5wPry4cMXb",
        "outputId": "ece0ee2b-8e02-41ac-ce4d-b2686316de94"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7d78133e0940>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from PIL import Image\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "img = Image.open(\"./outputs/segment/0/24_83.jpg\").convert(\"L\")\n",
        "plt.imshow(img)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 453
        },
        "id": "K7Zk8cCbfnqh",
        "outputId": "034d1749-0351-43ce-f8ef-cd750dfdb82b"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7d7826ec2a10>"
            ]
          },
          "metadata": {},
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ls ./inputs/test/masks/0/ | wc -l"
      ],
      "metadata": {
        "id": "8f86rJ9IOfGf",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8f630890-9f42-41f1-d66b-ad607c85cf00"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "459\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qL2c0PjZs8fY"
      },
      "outputs": [],
      "source": [
        "!mkdir /content/drive/MyDrive/UNeXt-pytorch\n",
        "!cp -r ./models/segment /content/drive/MyDrive/UNeXt-pytorch/models"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DixzQpgcEh5g"
      },
      "outputs": [],
      "source": [
        "!cp -r ./ /content/drive/MyDrive/UNeXt-pytorch/"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "machine_shape": "hm",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}