{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport os","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2023-05-09T02:34:56.013970Z","iopub.execute_input":"2023-05-09T02:34:56.014223Z","iopub.status.idle":"2023-05-09T02:34:56.024601Z","shell.execute_reply.started":"2023-05-09T02:34:56.014200Z","shell.execute_reply":"2023-05-09T02:34:56.023680Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# I used MMDetection framework based on torch, following their tutorial   \n## https://github.com/open-mmlab/mmdetection/blob/master/demo/MMDet_Tutorial.ipynb","metadata":{}},{"cell_type":"markdown","source":"## i wanted to use mmdetection of version 2. like in tutorial and mmcv 1. so i use torch <2.0","metadata":{}},{"cell_type":"code","source":"!pip uninstall torch --yes\n!pip uninstall torch --yes","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:34:56.028992Z","iopub.execute_input":"2023-05-09T02:34:56.029882Z","iopub.status.idle":"2023-05-09T02:35:19.478904Z","shell.execute_reply.started":"2023-05-09T02:34:56.029851Z","shell.execute_reply":"2023-05-09T02:35:19.477589Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# install dependencies: (use cu113 because has problems with cu12)\n!pip install torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113\n# install mmcv-full thus we could use CUDA operators\n!pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.12.0/index.html\n\n# Install mmdetection\n!rm -rf mmdetection\n!git clone https://github.com/open-mmlab/mmdetection --branch 2.x\n%cd mmdetection\n\n!pip install -e .\n\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:35:19.486043Z","iopub.execute_input":"2023-05-09T02:35:19.489806Z","iopub.status.idle":"2023-05-09T02:38:01.748372Z","shell.execute_reply.started":"2023-05-09T02:35:19.489770Z","shell.execute_reply":"2023-05-09T02:38:01.747111Z"},"collapsed":true,"jupyter":{"outputs_hidden":true},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Check Pytorch installation\nimport torch, torchvision\nprint(torch.__version__, torch.cuda.is_available())\n\n# Check MMDetection installation\nimport mmdet\nprint(mmdet.__version__)\n\n# Check mmcv installation\nfrom mmcv.ops import get_compiling_cuda_version, get_compiler_version\nprint(get_compiling_cuda_version())\nprint(get_compiler_version())","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:01.752817Z","iopub.execute_input":"2023-05-09T02:38:01.753154Z","iopub.status.idle":"2023-05-09T02:38:04.486072Z","shell.execute_reply.started":"2023-05-09T02:38:01.753123Z","shell.execute_reply":"2023-05-09T02:38:04.485031Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Firstly, i try inference with pretrained detector to see if everything is working fine","metadata":{}},{"cell_type":"code","source":"# We download the pre-trained checkpoints for inference and finetuning.\n!mkdir checkpoints\n!wget -c https://download.openmmlab.com/mmdetection/v2.0/faster_rcnn/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco_20210526_095054-1f77628b.pth \\\n      -O checkpoints/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco_20210526_095054-1f77628b.pth\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:04.487673Z","iopub.execute_input":"2023-05-09T02:38:04.488358Z","iopub.status.idle":"2023-05-09T02:38:17.346502Z","shell.execute_reply.started":"2023-05-09T02:38:04.488305Z","shell.execute_reply":"2023-05-09T02:38:17.345350Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import mmcv\nfrom mmcv.runner import load_checkpoint\n\nfrom mmdet.apis import inference_detector, show_result_pyplot\nfrom mmdet.models import build_detector\n\n# Choose to use a config and initialize the detector\nconfig = 'configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco.py'\n# Setup a checkpoint file to load\ncheckpoint = 'checkpoints/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco_20210526_095054-1f77628b.pth'\n\n# Set the device to be used for evaluation\ndevice='cuda:0'\n\n# Load the config\nconfig = mmcv.Config.fromfile(config)\n# Set pretrained to be None since we do not need pretrained model here\nconfig.model.pretrained = None\n\n# Initialize the detector\nmodel = build_detector(config.model)\n\n# Load checkpoint\ncheckpoint = load_checkpoint(model, checkpoint, map_location=device)\n\n# Set the classes of models for inference\nmodel.CLASSES = checkpoint['meta']['CLASSES']\n\n# We need to set the model's cfg for inference\nmodel.cfg = config\n\n# Convert the model to GPU\nmodel.to(device)\n# Convert the model into evaluation mode\nmodel.eval()","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:17.348795Z","iopub.execute_input":"2023-05-09T02:38:17.349183Z","iopub.status.idle":"2023-05-09T02:38:22.401108Z","shell.execute_reply.started":"2023-05-09T02:38:17.349147Z","shell.execute_reply":"2023-05-09T02:38:22.400119Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Use the detector to do inference\nimg = 'demo/demo.jpg'\nresult = inference_detector(model, img)\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:22.402494Z","iopub.execute_input":"2023-05-09T02:38:22.402834Z","iopub.status.idle":"2023-05-09T02:38:24.926106Z","shell.execute_reply.started":"2023-05-09T02:38:22.402803Z","shell.execute_reply":"2023-05-09T02:38:24.924329Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Let's plot the result\nshow_result_pyplot(model, img, result, score_thr=0.3)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:24.927700Z","iopub.execute_input":"2023-05-09T02:38:24.928144Z","iopub.status.idle":"2023-05-09T02:38:25.570758Z","shell.execute_reply.started":"2023-05-09T02:38:24.928107Z","shell.execute_reply":"2023-05-09T02:38:25.569525Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# !apt-get -q install tree","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:25.571964Z","iopub.execute_input":"2023-05-09T02:38:25.572259Z","iopub.status.idle":"2023-05-09T02:38:25.578617Z","shell.execute_reply.started":"2023-05-09T02:38:25.572234Z","shell.execute_reply":"2023-05-09T02:38:25.577751Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# !tree /kaggle/input/rsna-pneumonia-detection-challenge","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:25.582919Z","iopub.execute_input":"2023-05-09T02:38:25.583657Z","iopub.status.idle":"2023-05-09T02:38:25.587461Z","shell.execute_reply.started":"2023-05-09T02:38:25.583627Z","shell.execute_reply":"2023-05-09T02:38:25.586570Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Now we can try to train detector on pneumonia task","metadata":{}},{"cell_type":"markdown","source":"## i like to copy input in working directory to make it not read-only","metadata":{}},{"cell_type":"code","source":"%cp -r /kaggle/input/rsna-pneumonia-detection-challenge/stage_2_test_images /kaggle/working/","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:25.588972Z","iopub.execute_input":"2023-05-09T02:38:25.589595Z","iopub.status.idle":"2023-05-09T02:38:55.586156Z","shell.execute_reply.started":"2023-05-09T02:38:25.589566Z","shell.execute_reply":"2023-05-09T02:38:55.584748Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%cp -r /kaggle/input/rsna-pneumonia-detection-challenge/stage_2_train_images /kaggle/working/","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:38:55.587919Z","iopub.execute_input":"2023-05-09T02:38:55.588340Z","iopub.status.idle":"2023-05-09T02:43:07.971876Z","shell.execute_reply.started":"2023-05-09T02:38:55.588260Z","shell.execute_reply":"2023-05-09T02:43:07.970307Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Convert images from dcm to jpg format","metadata":{}},{"cell_type":"code","source":"import pydicom as dicom # read dcm images from dataset\nimport cv2 # convert images\nfrom tqdm import tqdm # progress bar","metadata":{"execution":{"iopub.status.busy":"2023-05-09T05:15:10.080424Z","iopub.status.idle":"2023-05-09T05:15:10.081123Z","shell.execute_reply.started":"2023-05-09T05:15:10.080875Z","shell.execute_reply":"2023-05-09T05:15:10.080899Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%cd /kaggle/working/stage_2_train_images","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:43:08.100796Z","iopub.execute_input":"2023-05-09T02:43:08.101149Z","iopub.status.idle":"2023-05-09T02:43:08.108106Z","shell.execute_reply.started":"2023-05-09T02:43:08.101118Z","shell.execute_reply":"2023-05-09T02:43:08.107101Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for filename in tqdm(os.listdir('.')):  # train images\n    ds = dicom.dcmread(filename)\n    img = np.asarray(ds.pixel_array) # shape = (height, width)\n    img = np.expand_dims(img, axis=0) #shape = (1,height, width)\n    img = np.moveaxis(img, -1, 0) # shape = (height, 1, width)\n    img = np.moveaxis(img, -1, 0) # shape = (height, width, 1) -> gray scale image array (1 channel)\n    img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB) # shape = (height, width, 3) -> color image array(3 channel)\n    cv2.imwrite(filename[:-4]+ '.jpg', img) # save image in jpg format ","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:43:08.109672Z","iopub.execute_input":"2023-05-09T02:43:08.110257Z","iopub.status.idle":"2023-05-09T02:54:11.968123Z","shell.execute_reply.started":"2023-05-09T02:43:08.110226Z","shell.execute_reply":"2023-05-09T02:54:11.966997Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%rm *.dcm # removing dcm files","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:54:11.971152Z","iopub.execute_input":"2023-05-09T02:54:11.971832Z","iopub.status.idle":"2023-05-09T02:54:14.315458Z","shell.execute_reply.started":"2023-05-09T02:54:11.971790Z","shell.execute_reply":"2023-05-09T02:54:14.314174Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%cd /kaggle/working/stage_2_test_images","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:54:14.317553Z","iopub.execute_input":"2023-05-09T02:54:14.317973Z","iopub.status.idle":"2023-05-09T02:54:14.335711Z","shell.execute_reply.started":"2023-05-09T02:54:14.317932Z","shell.execute_reply":"2023-05-09T02:54:14.334220Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for filename in tqdm(os.listdir('.')): # test images\n    ds = dicom.dcmread(filename)\n    img = np.asarray(ds.pixel_array) # shape = (height, width)\n    img = np.expand_dims(img, axis=0) #shape = (1,height, width)\n    img = np.moveaxis(img, -1, 0) # shape = (height, 1, width)\n    img = np.moveaxis(img, -1, 0) # shape = (height, width, 1) -> gray scale image array (1 channel)\n    img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB) # shape = (height, width, 3) -> color image array(3 channel)\n    cv2.imwrite(filename[:-4] + '.jpg', img) # save image in jpg format ","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:54:14.337449Z","iopub.execute_input":"2023-05-09T02:54:14.337957Z","iopub.status.idle":"2023-05-09T02:55:27.249983Z","shell.execute_reply.started":"2023-05-09T02:54:14.337924Z","shell.execute_reply":"2023-05-09T02:55:27.248876Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%rm *.dcm","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:55:27.251536Z","iopub.execute_input":"2023-05-09T02:55:27.252365Z","iopub.status.idle":"2023-05-09T02:55:28.427441Z","shell.execute_reply.started":"2023-05-09T02:55:27.252330Z","shell.execute_reply":"2023-05-09T02:55:28.426162Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Quick look on a image from the dataset","metadata":{}},{"cell_type":"code","source":"import mmcv\nimport matplotlib.pyplot as plt\n\nimg = mmcv.imread('/kaggle/working/stage_2_train_images/e65d66fe-2835-4e27-859b-a65065758cab.jpg')\nplt.imshow(img)\nplt.show()\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:08:10.024108Z","iopub.execute_input":"2023-05-09T04:08:10.024504Z","iopub.status.idle":"2023-05-09T04:08:10.448179Z","shell.execute_reply.started":"2023-05-09T04:08:10.024474Z","shell.execute_reply":"2023-05-09T04:08:10.447211Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"img.shape # image shape","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:08:15.808405Z","iopub.execute_input":"2023-05-09T04:08:15.808762Z","iopub.status.idle":"2023-05-09T04:08:15.816863Z","shell.execute_reply.started":"2023-05-09T04:08:15.808735Z","shell.execute_reply":"2023-05-09T04:08:15.815586Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"!cat /kaggle/input/rsna-pneumonia-detection-challenge/stage_2_train_labels.csv # if you want to see labels","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(os.listdir('/kaggle/working/stage_2_train_images'))","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:55:30.120552Z","iopub.execute_input":"2023-05-09T02:55:30.121222Z","iopub.status.idle":"2023-05-09T02:55:30.164444Z","shell.execute_reply.started":"2023-05-09T02:55:30.121186Z","shell.execute_reply":"2023-05-09T02:55:30.163641Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%cd /kaggle/working","metadata":{"execution":{"iopub.status.busy":"2023-05-09T02:55:30.168782Z","iopub.execute_input":"2023-05-09T02:55:30.171992Z","iopub.status.idle":"2023-05-09T02:55:30.182272Z","shell.execute_reply.started":"2023-05-09T02:55:30.171961Z","shell.execute_reply":"2023-05-09T02:55:30.181310Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Convert labels format to coco, it helps to simplify next steps with configuration","metadata":{}},{"cell_type":"markdown","source":"### With help of this paper:   \nhttps://medium.com/analytics-vidhya/how-to-convert-tensorflow-object-detection-csv-data-to-coco-json-format-d0693d5b2f75","metadata":{}},{"cell_type":"code","source":"import numpy as np\nimport json\nimport pandas as pd\n\npath = '/kaggle/input/rsna-pneumonia-detection-challenge/stage_2_train_labels.csv' # the path to the CSV file\nsave_json_path = 'labels.json'\n\n\ndata = pd.read_csv(path)\n\nimages = []\ncategories = []\nannotations = []\n\ncategory = {}\ncategory[\"supercategory\"] = 'none'\ncategory[\"id\"] = 0\ncategory[\"name\"] = 'None'\ncategories.append(category)\n\ndata['fileid'] = data['patientId'].astype('category').cat.codes\ndata['categoryid']= pd.Categorical(data['Target'],ordered= True).codes\ndata['categoryid'] = data['categoryid']+1\ndata['annid'] = data.index\n\ndef image(row):\n    image = {}\n    image[\"height\"] = 1024\n    image[\"width\"] = 1024\n    image[\"id\"] = row.fileid\n    image[\"file_name\"] = row.patientId + '.jpg'\n    return image\n\ndef category(row):\n    category = {}\n    category[\"supercategory\"] = 'None'\n    category[\"id\"] = row.categoryid\n    category[\"name\"] = row[6] # 6 column is the Target\n    return category\n\ndef annotation(row):\n    annotation = {}\n    area = (row.width)*(row.height)\n    annotation[\"segmentation\"] = []\n    annotation[\"iscrowd\"] = 0\n    annotation[\"area\"] = area\n    annotation[\"image_id\"] = row.fileid\n\n    annotation[\"bbox\"] = [row.x, row.y, row.width,row.height]\n\n    annotation[\"category_id\"] = row.categoryid\n    annotation[\"id\"] = row.annid\n    return annotation\n\nfor row in data.itertuples():\n    annotations.append(annotation(row))\n\nimagedf = data.drop_duplicates(subset=['fileid']).sort_values(by='fileid')\nfor row in imagedf.itertuples():\n    images.append(image(row))\n\ncatdf = data.drop_duplicates(subset=['categoryid']).sort_values(by='categoryid')\nfor row in catdf.itertuples():\n    categories.append(category(row))\n\ndata_coco = {}\ndata_coco[\"images\"] = images\ndata_coco[\"categories\"] = categories\ndata_coco[\"annotations\"] = annotations\n\n\njson.dump(data_coco, open(save_json_path, \"w\"), indent=4)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:00.002130Z","iopub.execute_input":"2023-05-09T04:32:00.002916Z","iopub.status.idle":"2023-05-09T04:32:01.220476Z","shell.execute_reply.started":"2023-05-09T04:32:00.002879Z","shell.execute_reply":"2023-05-09T04:32:01.219267Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"row","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:02.140873Z","iopub.execute_input":"2023-05-09T04:32:02.143565Z","iopub.status.idle":"2023-05-09T04:32:02.151018Z","shell.execute_reply.started":"2023-05-09T04:32:02.143510Z","shell.execute_reply":"2023-05-09T04:32:02.149929Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# I decided to split train data json to train, valid and test for a natural use in a config, but you can split only to train/valid, because we have test images in dataset in additional directory and really use only train and val ","metadata":{}},{"cell_type":"code","source":"!pip install echo1-coco-split","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:07.267528Z","iopub.execute_input":"2023-05-09T04:32:07.267898Z","iopub.status.idle":"2023-05-09T04:32:16.935583Z","shell.execute_reply.started":"2023-05-09T04:32:07.267870Z","shell.execute_reply":"2023-05-09T04:32:16.934192Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%cd /kaggle/working/","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:16.939611Z","iopub.execute_input":"2023-05-09T04:32:16.940081Z","iopub.status.idle":"2023-05-09T04:32:16.948735Z","shell.execute_reply.started":"2023-05-09T04:32:16.940032Z","shell.execute_reply":"2023-05-09T04:32:16.947762Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%rm train.json\n%rm test.json\n%rm valid.json","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:50.244257Z","iopub.execute_input":"2023-05-09T04:32:50.244835Z","iopub.status.idle":"2023-05-09T04:32:53.562539Z","shell.execute_reply.started":"2023-05-09T04:32:50.244802Z","shell.execute_reply":"2023-05-09T04:32:53.560998Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"!coco-split \\\n    --has_annotations \\\n    --valid_ratio .2 \\\n    --test_ratio .1 \\\n    --annotations_file /kaggle/working/labels.json","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:32:56.755998Z","iopub.execute_input":"2023-05-09T04:32:56.756432Z","iopub.status.idle":"2023-05-09T04:33:10.908252Z","shell.execute_reply.started":"2023-05-09T04:32:56.756396Z","shell.execute_reply":"2023-05-09T04:33:10.907075Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Load existing config and edit it","metadata":{}},{"cell_type":"code","source":"from mmcv import Config\ncfg = Config.fromfile('/kaggle/working/mmdetection/configs/faster_rcnn/faster_rcnn_r50_caffe_fpn_mstrain_1x_coco.py')","metadata":{"execution":{"iopub.status.busy":"2023-05-09T05:32:56.337447Z","iopub.execute_input":"2023-05-09T05:32:56.337900Z","iopub.status.idle":"2023-05-09T05:32:56.367578Z","shell.execute_reply.started":"2023-05-09T05:32:56.337867Z","shell.execute_reply":"2023-05-09T05:32:56.366691Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cfg.keys()","metadata":{"execution":{"iopub.status.busy":"2023-05-09T04:03:32.525952Z","iopub.execute_input":"2023-05-09T04:03:32.526320Z","iopub.status.idle":"2023-05-09T04:03:32.532474Z","shell.execute_reply.started":"2023-05-09T04:03:32.526272Z","shell.execute_reply":"2023-05-09T04:03:32.531506Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from mmdet.apis import set_random_seed\n# Modify dataset type and path\ncfg.dataset_type = 'CocoDataset'\ncfg.data_root = '/kaggle/working/'\ncfg.device = 'cuda'\nclasses = (0,1,) # \n\ncfg.data.test.type = 'CocoDataset'\ncfg.data.test.classes = classes\ncfg.data.test.data_root = '/kaggle/working/stage_2_train_images'\ncfg.data.test.ann_file = '/kaggle/working/test.json'\ncfg.data.test.img_prefix = ''\n\ncfg.data.train.type = 'CocoDataset'\ncfg.data.train.classes = classes\ncfg.data.train.data_root = '/kaggle/working/stage_2_train_images'\ncfg.data.train.ann_file = '/kaggle/working/train.json'\ncfg.data.train.img_prefix = ''\n\n\ncfg.data.val.type = 'CocoDataset'\ncfg.data.val.classes = classes\ncfg.data.val.data_root = '/kaggle/working/stage_2_train_images'\ncfg.data.val.ann_file = '/kaggle/working/valid.json'\ncfg.data.val.img_prefix = ''\n\n# modify num classes of the model in box head\ncfg.model.roi_head.bbox_head.num_classes = 2\n# If we need to finetune a model based on a pre-trained detector, we need to\n# use load_from to set the path of checkpoints.\ncfg.load_from = '/kaggle/working/mmdetection/checkpoints/faster_rcnn_r50_caffe_fpn_mstrain_3x_coco_20210526_095054-1f77628b.pth'\n\n# Set up working dir to save files and logs.\ncfg.work_dir = '/kaggle/working/logs_exps'\n\n# The original learning rate (LR) is set for 8-GPU training.\n# We divide it by 8 since we only use one GPU.\ncfg.optimizer.lr = 0.02 / 8\ncfg.lr_config.warmup = None\ncfg.log_config.interval = 10\n\n# Change the evaluation metric since we use customized dataset.\ncfg.evaluation.metric = 'bbox'\n# We can set the evaluation interval to reduce the evaluation times\ncfg.evaluation.interval = 1\n# We can set the checkpoint saving interval to reduce the storage cost\ncfg.checkpoint_config.interval = 1\n\n\n# I will train for 2 epochs to get fast result\ncfg.runner.max_epochs = 2\n\n# Set seed thus the results are more reproducible\ncfg.seed = 0\nset_random_seed(0, deterministic=False)\ncfg.gpu_ids = range(1)\n\n# We can also use tensorboard to log the training process\ncfg.log_config.hooks = [\n    dict(type='TextLoggerHook'),\n    dict(type='TensorboardLoggerHook')]\n\n\n# We can initialize the logger for training and have a look\n# at the final config used for training\nprint(f'Config:\\n{cfg.pretty_text}')","metadata":{"execution":{"iopub.status.busy":"2023-05-09T10:09:25.658798Z","iopub.execute_input":"2023-05-09T10:09:25.659262Z","iopub.status.idle":"2023-05-09T10:09:26.602987Z","shell.execute_reply.started":"2023-05-09T10:09:25.659224Z","shell.execute_reply":"2023-05-09T10:09:26.600637Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Now we can build detector to train","metadata":{}},{"cell_type":"code","source":"from mmdet.datasets import build_dataset\nfrom mmdet.models import build_detector\nfrom mmdet.apis import train_detector\nimport os.path as osp\n\n\n\n# Build dataset\ndatasets = [build_dataset(cfg.data.train)]\n\n# Build the detector\nmodel = build_detector(cfg.model)\n# Add an attribute for visualization convenience\nclasses = ('nothing', 'pneumonia')\nmodel.CLASSES = classes\n\n\n\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T10:09:26.622523Z","iopub.execute_input":"2023-05-09T10:09:26.622862Z","iopub.status.idle":"2023-05-09T10:09:27.293232Z","shell.execute_reply.started":"2023-05-09T10:09:26.622835Z","shell.execute_reply":"2023-05-09T10:09:27.292026Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### quick check if dataset load went wrong","metadata":{}},{"cell_type":"code","source":"print(datasets)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T10:09:27.520635Z","iopub.execute_input":"2023-05-09T10:09:27.521622Z","iopub.status.idle":"2023-05-09T10:09:28.144553Z","shell.execute_reply.started":"2023-05-09T10:09:27.521575Z","shell.execute_reply":"2023-05-09T10:09:28.143543Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Create work_dir\nmmcv.mkdir_or_exist(osp.abspath(cfg.work_dir))\n# Train\ntrain_detector(model, datasets, cfg, distributed=False, validate=True)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T10:09:28.209046Z","iopub.execute_input":"2023-05-09T10:09:28.209620Z","iopub.status.idle":"2023-05-09T11:26:59.325365Z","shell.execute_reply.started":"2023-05-09T10:09:28.209583Z","shell.execute_reply":"2023-05-09T11:26:59.324185Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Test (on 5 epochs, from previous session)","metadata":{}},{"cell_type":"markdown","source":"## Pneumonia image ![image.png](attachment:786c27ae-74b7-4fc8-9006-3f3b10f31417.png)","metadata":{},"attachments":{"786c27ae-74b7-4fc8-9006-3f3b10f31417.png":{"image/png":"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"}}},{"cell_type":"code","source":"img = mmcv.imread('/kaggle/working/stage_2_train_images/00436515-870c-4b36-a041-de91049b9ab4.jpg')\nmodel.cfg = cfg\nresult = inference_detector(model, img)\nshow_result_pyplot(model, img, result)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T08:44:08.918828Z","iopub.execute_input":"2023-05-09T08:44:08.919306Z","iopub.status.idle":"2023-05-09T08:44:09.787429Z","shell.execute_reply.started":"2023-05-09T08:44:08.919249Z","shell.execute_reply":"2023-05-09T08:44:09.786587Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## non-pneumonia ![image.png](attachment:1f48835e-5387-4805-99ca-a39edf880e3f.png)","metadata":{},"attachments":{"1f48835e-5387-4805-99ca-a39edf880e3f.png":{"image/png":"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"}}},{"cell_type":"code","source":"img3 = mmcv.imread('/kaggle/working/stage_2_train_images/00569f44-917d-4c86-a842-81832af98c30.jpg')\nresult = inference_detector(model, img3)\nshow_result_pyplot(model, img3, result)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T08:41:51.125222Z","iopub.execute_input":"2023-05-09T08:41:51.126275Z","iopub.status.idle":"2023-05-09T08:41:52.017999Z","shell.execute_reply.started":"2023-05-09T08:41:51.126238Z","shell.execute_reply":"2023-05-09T08:41:52.017091Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Now pneumonia image from test.json: ![image.png](attachment:5d9054dd-6fdd-4f23-8435-dfd8031e6b71.png) ![image.png](attachment:fe4d5a53-f0eb-4139-b4b6-268512b2592f.png)","metadata":{},"attachments":{"5d9054dd-6fdd-4f23-8435-dfd8031e6b71.png":{"image/png":"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"},"fe4d5a53-f0eb-4139-b4b6-268512b2592f.png":{"image/png":"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"}}},{"cell_type":"code","source":"img4 = mmcv.imread('/kaggle/working/stage_2_train_images/0572881e-d1dd-4757-a54e-b240b30da946.jpg') # from test.json\nresult = inference_detector(model, img4)\nshow_result_pyplot(model, img4, result)","metadata":{"execution":{"iopub.status.busy":"2023-05-09T08:47:06.401670Z","iopub.execute_input":"2023-05-09T08:47:06.402806Z","iopub.status.idle":"2023-05-09T08:47:07.343829Z","shell.execute_reply.started":"2023-05-09T08:47:06.402768Z","shell.execute_reply":"2023-05-09T08:47:07.342987Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## non-pneumonia image from test.json ![image.png](attachment:e45ef3e7-085c-4622-9b18-b0157d3c4ddb.png) ![image.png](attachment:de4d1e56-8630-4d7f-afad-9bf2fb554fdc.png)","metadata":{},"attachments":{"e45ef3e7-085c-4622-9b18-b0157d3c4ddb.png":{"image/png":"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"},"de4d1e56-8630-4d7f-afad-9bf2fb554fdc.png":{"image/png":"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"}}},{"cell_type":"code","source":"img5 = mmcv.imread('/kaggle/working/stage_2_train_images/0d121525-812e-4ab2-b29f-ff04b3d97ffa.jpg') # from test.json\nresult = inference_detector(model, img5)\nshow_result_pyplot(model, img5, result)\n","metadata":{"execution":{"iopub.status.busy":"2023-05-09T08:53:33.500654Z","iopub.execute_input":"2023-05-09T08:53:33.501122Z","iopub.status.idle":"2023-05-09T08:53:34.325740Z","shell.execute_reply.started":"2023-05-09T08:53:33.501089Z","shell.execute_reply":"2023-05-09T08:53:34.324498Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}