{"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":"markdown","source":"## Acknowledgment\nMain code and models in this notebook are from [Egor Trushin](https://www.kaggle.com/code/egortrushin/gr-icrgw-pytorch-lightning-baseline-unet-resnest) and [Shashwat Raman](https://www.kaggle.com/code/shashwatraman/simple-unet-baseline-infer-lb-0-580). Thanks to their amazing notebooks, I can try out ensemble here.\n\nThanks to [@imeintanis](https://www.kaggle.com/imeintanis)'s patient guidance, I've corrected a serious mistake.\n## Method Description\nI tried to apply **and operation** and **or operation** to combine the masks of different models, but the resulting was poor. In this notebook, I calculate the **weighted sum of probabilities** and then apply a threshold to get masks. \nNotice that there are K hyperparameters instead of K+1, where K is the number of models. This is because that I don't force the sum of weights equals to 1. In addition, I **scale their probabilities** based on the thresholds in their original notebooks.\n\n\n","metadata":{}},{"cell_type":"markdown","source":"## Experiments\n\n| | model_1 | model_2 |\n|--------|---------:|---------:|\n| **lb** | 0.580 | 0.618 |\n\n\n| ratio_1 | ratio_2 | lb |\n|----------:|----------:|-------:|\n| 0.3 | 0.7 | 0.626 |\n| 0.4 | 0.6 | 0.626 |\n| 0.45 | 0.55 | 0.626 |\n| 0.5 | 0.5 | 0.626 |\n| 0.55 | 0.45 | 0.587 |\n| 0.6 | 0.4 | 0.585 |\n\n\n**Lb remains 0.626 as long as the weight of the better model is greater than or equals to the weight of another model.\nThis pattern exists when I try to ensemble my model.**\n\n\n| | model_1 | model_2 |\n|--------|---------:|---------:|\n| **lb** | 0.552 | 0.580 |\n\n| ratio_1 | ratio_2 | lb |\n|---------:|---------:|-------:|\n| 0.5 | 0.5 | 0.592 |\n| 0.6 | 0.5 | 0.589 |\n| 0.5 | 0.6 | 0.592 |\n\n","metadata":{}},{"cell_type":"markdown","source":"## Hyperparameters","metadata":{}},{"cell_type":"code","source":"ratio_1 = 0.55\nratio_2 = 0.45\nthr_1 = 0.02\nthr_2 = 0.5","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:44:58.319978Z","iopub.execute_input":"2023-06-17T09:44:58.320422Z","iopub.status.idle":"2023-06-17T09:44:58.325936Z","shell.execute_reply.started":"2023-06-17T09:44:58.320388Z","shell.execute_reply":"2023-06-17T09:44:58.324722Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Import","metadata":{}},{"cell_type":"code","source":"import warnings\nwarnings.filterwarnings(\"ignore\")\n\nimport os\nimport math\nimport random\nfrom PIL import Image\nfrom collections import defaultdict\n\nimport numpy as np\nimport pandas as pd\nfrom pathlib import Path\nfrom tqdm.notebook import tqdm\nimport matplotlib.pyplot as plt\n\nimport torch\nfrom torch import nn\nimport pytorch_lightning as pl\nfrom torchvision import transforms\nimport torchvision.transforms as T\nfrom torch.utils.data import Dataset, DataLoader\nfrom transformers import get_cosine_schedule_with_warmup\n\nimport sys\nsys.path.append(\"../input/pretrained-models-pytorch\")\nsys.path.append(\"../input/efficientnet-pytorch\")\nsys.path.append(\"/kaggle/input/smp-github/segmentation_models.pytorch-master\")\nsys.path.append(\"/kaggle/input/timm-pretrained-resnest/resnest/\")\nimport segmentation_models_pytorch as smp\n\nprint(f\"Segmentation Models version: {smp.__version__}\")\n\n!mkdir -p /root/.cache/torch/hub/checkpoints/\n!cp /kaggle/input/timm-pretrained-resnest/resnest/gluon_resnest26-50eb607c.pth /root/.cache/torch/hub/checkpoints/gluon_resnest26-50eb607c.pth","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:44:58.334227Z","iopub.execute_input":"2023-06-17T09:44:58.334622Z","iopub.status.idle":"2023-06-17T09:45:00.833939Z","shell.execute_reply.started":"2023-06-17T09:44:58.334591Z","shell.execute_reply":"2023-06-17T09:45:00.832657Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## U_Net#1 [LB 0.580]","metadata":{}},{"cell_type":"code","source":"class Config:\n batch_size = 32\n seed = 42\n thr = 0.02\n \n encoder = 'efficientnet-b0'\n pretrained = False\n weights = None\n classes = ['contrail']\n activation = None\n in_chans = 3\n \n device = 'cuda' if torch.cuda.is_available() else 'cpu'\n \n image_size = 256\n \n model_ckpt = '/kaggle/input/unet-model/epoch-9.pth'\n \nclass Paths:\n data = '/kaggle/input/google-research-identify-contrails-reduce-global-warming'\n data_root = '/kaggle/input/google-research-identify-contrails-reduce-global-warming/test/'","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:00.836315Z","iopub.execute_input":"2023-06-17T09:45:00.836955Z","iopub.status.idle":"2023-06-17T09:45:00.844983Z","shell.execute_reply.started":"2023-06-17T09:45:00.836917Z","shell.execute_reply":"2023-06-17T09:45:00.843724Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def set_seed(seed=1234):\n random.seed(seed)\n np.random.seed(seed)\n os.environ[\"PYTHONHASHSEED\"] = str(seed)\n \n torch.manual_seed(seed)\n torch.cuda.manual_seed(seed)\n torch.backends.cudnn.deterministic = False\n torch.backends.cudnn.benchmark = True\n \nset_seed(Config.seed)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:00.846817Z","iopub.execute_input":"2023-06-17T09:45:00.84756Z","iopub.status.idle":"2023-06-17T09:45:01.056931Z","shell.execute_reply.started":"2023-06-17T09:45:00.847517Z","shell.execute_reply":"2023-06-17T09:45:01.055595Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"filenames = os.listdir(Paths.data_root)\ntest_df = pd.DataFrame(filenames, columns=['record_id'])\n\ntest_df['path'] = Paths.data_root + test_df['record_id'].astype(str)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.059802Z","iopub.execute_input":"2023-06-17T09:45:01.060152Z","iopub.status.idle":"2023-06-17T09:45:01.072827Z","shell.execute_reply.started":"2023-06-17T09:45:01.060125Z","shell.execute_reply":"2023-06-17T09:45:01.071756Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class ContrailsDataset(torch.utils.data.Dataset):\n def __init__(self, df, train=True):\n \n self.df = df\n self.trn = train\n \n def read_record(self, directory):\n record_data = {}\n for x in [\n \"band_11\", \n \"band_14\", \n \"band_15\"\n ]:\n\n record_data[x] = np.load(os.path.join(directory, x + \".npy\"))\n\n return record_data\n\n def normalize_range(self, data, bounds):\n \"\"\"Maps data to the range [0, 1].\"\"\"\n return (data - bounds[0]) / (bounds[1] - bounds[0])\n \n def get_false_color(self, record_data):\n _T11_BOUNDS = (243, 303)\n _CLOUD_TOP_TDIFF_BOUNDS = (-4, 5)\n _TDIFF_BOUNDS = (-4, 2)\n \n N_TIMES_BEFORE = 4\n\n r = self.normalize_range(record_data[\"band_15\"] - record_data[\"band_14\"], _TDIFF_BOUNDS)\n g = self.normalize_range(record_data[\"band_14\"] - record_data[\"band_11\"], _CLOUD_TOP_TDIFF_BOUNDS)\n b = self.normalize_range(record_data[\"band_14\"], _T11_BOUNDS)\n false_color = np.clip(np.stack([r, g, b], axis=2), 0, 1)\n img = false_color[..., N_TIMES_BEFORE]\n\n return img\n \n def __getitem__(self, index):\n row = self.df.iloc[index]\n con_path = row.path\n data = self.read_record(con_path) \n \n img = self.get_false_color(data)\n \n img = torch.tensor(img)\n img = img.permute(2, 0, 1)\n \n return img.float()\n \n def __len__(self):\n return len(self.df)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.074173Z","iopub.execute_input":"2023-06-17T09:45:01.075049Z","iopub.status.idle":"2023-06-17T09:45:01.089242Z","shell.execute_reply.started":"2023-06-17T09:45:01.075018Z","shell.execute_reply":"2023-06-17T09:45:01.088357Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_ds = ContrailsDataset(\n test_df,\n train = False\n )\n \ntest_dl = DataLoader(test_ds, batch_size=Config.batch_size, num_workers = 2)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.090754Z","iopub.execute_input":"2023-06-17T09:45:01.091875Z","iopub.status.idle":"2023-06-17T09:45:01.108146Z","shell.execute_reply.started":"2023-06-17T09:45:01.09184Z","shell.execute_reply":"2023-06-17T09:45:01.106928Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class UNet(nn.Module):\n def __init__(self, cfg):\n super(UNet, self).__init__()\n \n self.cfg = cfg\n self.training = True\n \n self.model = smp.Unet(\n encoder_name=cfg.encoder, \n encoder_weights=cfg.weights, \n decoder_use_batchnorm=True,\n classes=len(cfg.classes), \n activation=cfg.activation,\n )\n \n self.loss_fn = smp.losses.DiceLoss(mode='binary')\n \n def forward(self, imgs):\n \n x = imgs\n logits = self.model(x)\n \n return {\"logits\": logits.sigmoid()}","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.110591Z","iopub.execute_input":"2023-06-17T09:45:01.111339Z","iopub.status.idle":"2023-06-17T09:45:01.120517Z","shell.execute_reply.started":"2023-06-17T09:45:01.111255Z","shell.execute_reply":"2023-06-17T09:45:01.119438Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model = UNet(Config).to(Config.device)\nmodel.load_state_dict(torch.load(Config.model_ckpt, map_location=torch.device(Config.device)))","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.122733Z","iopub.execute_input":"2023-06-17T09:45:01.123064Z","iopub.status.idle":"2023-06-17T09:45:01.376832Z","shell.execute_reply.started":"2023-06-17T09:45:01.123036Z","shell.execute_reply":"2023-06-17T09:45:01.375562Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.eval()\ntorch.set_grad_enabled(False)\n\nval_data = defaultdict(list)\npbar = tqdm(enumerate(test_dl), total=len(test_dl), desc='Test')\nfor step, X in pbar: \n X = X.to(Config.device)\n\n output = model(X)\n for key, val in output.items():\n val_data[key] += [output[key]]\n\nfor key, val in output.items():\n value = val_data[key]\n if len(value[0].shape) == 0:\n val_data[key] = torch.stack(value)\n else:\n val_data[key] = torch.cat(value, dim=0).cpu().detach().numpy()","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.378087Z","iopub.execute_input":"2023-06-17T09:45:01.37849Z","iopub.status.idle":"2023-06-17T09:45:01.89474Z","shell.execute_reply.started":"2023-06-17T09:45:01.378456Z","shell.execute_reply":"2023-06-17T09:45:01.893505Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def rle_encode(x, fg_val=1):\n \"\"\"\n Args:\n x: numpy array of shape (height, width), 1 - mask, 0 - background\n Returns: run length encoding as list\n \"\"\"\n\n dots = np.where(\n x.T.flatten() == fg_val)[0] # .T sets Fortran order down-then-right\n run_lengths = []\n prev = -2\n for b in dots:\n if b > prev + 1:\n run_lengths.extend((b + 1, 0))\n run_lengths[-1] += 1\n prev = b\n return run_lengths\n\ndef list_to_string(x):\n \"\"\"\n Converts list to a string representation\n Empty list returns '-'\n \"\"\"\n if x: # non-empty list\n s = str(x).replace(\"[\", \"\").replace(\"]\", \"\").replace(\",\", \"\")\n else:\n s = '-'\n return s","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.900327Z","iopub.execute_input":"2023-06-17T09:45:01.900714Z","iopub.status.idle":"2023-06-17T09:45:01.910071Z","shell.execute_reply.started":"2023-06-17T09:45:01.90068Z","shell.execute_reply":"2023-06-17T09:45:01.908941Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Ids = []\nProbs_1 = []","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.911808Z","iopub.execute_input":"2023-06-17T09:45:01.912156Z","iopub.status.idle":"2023-06-17T09:45:01.925178Z","shell.execute_reply.started":"2023-06-17T09:45:01.91213Z","shell.execute_reply":"2023-06-17T09:45:01.92405Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for i, pred in enumerate(val_data['logits']):\n rec = test_df['record_id'][i]\n Ids.append(rec)\n Probs_1.append(pred[0])","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.926607Z","iopub.execute_input":"2023-06-17T09:45:01.927366Z","iopub.status.idle":"2023-06-17T09:45:01.938929Z","shell.execute_reply.started":"2023-06-17T09:45:01.927331Z","shell.execute_reply":"2023-06-17T09:45:01.937684Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import gc\ndel filenames, test_df, test_ds, test_dl, model, val_data, pbar\ngc.collect()","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:01.940463Z","iopub.execute_input":"2023-06-17T09:45:01.941274Z","iopub.status.idle":"2023-06-17T09:45:02.371771Z","shell.execute_reply.started":"2023-06-17T09:45:01.941216Z","shell.execute_reply":"2023-06-17T09:45:02.37093Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## U_Net#2 [LB 0.618]","metadata":{}},{"cell_type":"code","source":"batch_size = 16\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\ndata = '/kaggle/input/google-research-identify-contrails-reduce-global-warming'\ndata_root = '/kaggle/input/google-research-identify-contrails-reduce-global-warming/test/'","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.372957Z","iopub.execute_input":"2023-06-17T09:45:02.373905Z","iopub.status.idle":"2023-06-17T09:45:02.378916Z","shell.execute_reply.started":"2023-06-17T09:45:02.373871Z","shell.execute_reply":"2023-06-17T09:45:02.377824Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"filenames = os.listdir(data_root)\ntest_df = pd.DataFrame(filenames, columns=['record_id'])\ntest_df['path'] = data_root + test_df['record_id'].astype(str)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.380609Z","iopub.execute_input":"2023-06-17T09:45:02.38098Z","iopub.status.idle":"2023-06-17T09:45:02.398252Z","shell.execute_reply.started":"2023-06-17T09:45:02.38095Z","shell.execute_reply":"2023-06-17T09:45:02.396727Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class ContrailsDataset(torch.utils.data.Dataset):\n def __init__(self, df, train=True):\n \n self.df = df\n self.trn = train\n self.df_idx: pd.DataFrame = pd.DataFrame({'idx': os.listdir(f'/kaggle/input/google-research-identify-contrails-reduce-global-warming/test')})\n self.normalize_image = T.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))\n \n def read_record(self, directory):\n record_data = {}\n for x in [\n \"band_11\", \n \"band_14\", \n \"band_15\"\n ]:\n\n record_data[x] = np.load(os.path.join(directory, x + \".npy\"))\n\n return record_data\n\n def normalize_range(self, data, bounds):\n \"\"\"Maps data to the range [0, 1].\"\"\"\n return (data - bounds[0]) / (bounds[1] - bounds[0])\n \n def get_false_color(self, record_data):\n _T11_BOUNDS = (243, 303)\n _CLOUD_TOP_TDIFF_BOUNDS = (-4, 5)\n _TDIFF_BOUNDS = (-4, 2)\n \n N_TIMES_BEFORE = 4\n\n r = self.normalize_range(record_data[\"band_15\"] - record_data[\"band_14\"], _TDIFF_BOUNDS)\n g = self.normalize_range(record_data[\"band_14\"] - record_data[\"band_11\"], _CLOUD_TOP_TDIFF_BOUNDS)\n b = self.normalize_range(record_data[\"band_14\"], _T11_BOUNDS)\n false_color = np.clip(np.stack([r, g, b], axis=2), 0, 1)\n img = false_color[..., N_TIMES_BEFORE]\n\n return img\n \n def __getitem__(self, index):\n row = self.df.iloc[index]\n con_path = row.path\n data = self.read_record(con_path) \n \n img = self.get_false_color(data)\n \n img = torch.tensor(np.reshape(img, (256, 256, 3))).to(torch.float32).permute(2, 0, 1)\n \n img = self.normalize_image(img)\n \n image_id = int(self.df_idx.iloc[index]['idx'])\n \n return img.float(), torch.tensor(image_id)\n \n def __len__(self):\n return len(self.df)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.400571Z","iopub.execute_input":"2023-06-17T09:45:02.400974Z","iopub.status.idle":"2023-06-17T09:45:02.417728Z","shell.execute_reply.started":"2023-06-17T09:45:02.400938Z","shell.execute_reply":"2023-06-17T09:45:02.416556Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_ds = ContrailsDataset(\n test_df,\n train = False\n )\n \ntest_dl = DataLoader(test_ds, batch_size=batch_size, num_workers = 1)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.419398Z","iopub.execute_input":"2023-06-17T09:45:02.420692Z","iopub.status.idle":"2023-06-17T09:45:02.438218Z","shell.execute_reply.started":"2023-06-17T09:45:02.420654Z","shell.execute_reply":"2023-06-17T09:45:02.437057Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class LightningModule(pl.LightningModule):\n\n def __init__(self):\n super().__init__()\n self.model = smp.Unet(encoder_name=\"timm-resnest26d\",\n encoder_weights=None,\n in_channels=3,\n classes=1,\n activation=None,\n )\n\n def forward(self, batch):\n return self.model(batch)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.440062Z","iopub.execute_input":"2023-06-17T09:45:02.440445Z","iopub.status.idle":"2023-06-17T09:45:02.453983Z","shell.execute_reply.started":"2023-06-17T09:45:02.440414Z","shell.execute_reply":"2023-06-17T09:45:02.452774Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"if torch.cuda.is_available():\n model = LightningModule().load_from_checkpoint(\"/kaggle/input/models/model.ckpt\")\nelse:\n model = LightningModule().load_from_checkpoint(\"/kaggle/input/models/model.ckpt\", map_location=torch.device('cpu'))\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel.to(device)\nmodel.eval()\nmodel.zero_grad()","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:02.455343Z","iopub.execute_input":"2023-06-17T09:45:02.455728Z","iopub.status.idle":"2023-06-17T09:45:03.877627Z","shell.execute_reply.started":"2023-06-17T09:45:02.455684Z","shell.execute_reply":"2023-06-17T09:45:03.876697Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Probs_2 = []","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:03.879227Z","iopub.execute_input":"2023-06-17T09:45:03.879875Z","iopub.status.idle":"2023-06-17T09:45:03.883805Z","shell.execute_reply.started":"2023-06-17T09:45:03.879843Z","shell.execute_reply":"2023-06-17T09:45:03.882915Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for i, data in enumerate(test_dl):\n images, image_id = data\n \n # Predict mask for this instance\n images = images.to(device)\n predicated_mask = model.forward(images[:, :, :, :])\n predicated_mask = torch.sigmoid(predicated_mask).cpu().detach().numpy()\n \n # Apply threshold\n# predicated_mask_with_threshold = np.zeros((images.shape[0], 256, 256))\n# predicated_mask_with_threshold[predicated_mask[:, 0, :, :] < 0.5] = 0\n# predicated_mask_with_threshold[predicated_mask[:, 0, :, :] > 0.5] = 1\n \n for img_num in range(0, images.shape[0]):\n current_mask = predicated_mask[img_num, :, :]\n current_image_id = image_id[img_num].item()\n \n Probs_2.append(current_mask)","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:03.885313Z","iopub.execute_input":"2023-06-17T09:45:03.885952Z","iopub.status.idle":"2023-06-17T09:45:04.697563Z","shell.execute_reply.started":"2023-06-17T09:45:03.88592Z","shell.execute_reply":"2023-06-17T09:45:04.695907Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"del filenames, test_df, test_ds, model\ngc.collect()","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:04.6998Z","iopub.execute_input":"2023-06-17T09:45:04.700554Z","iopub.status.idle":"2023-06-17T09:45:05.140176Z","shell.execute_reply.started":"2023-06-17T09:45:04.700513Z","shell.execute_reply":"2023-06-17T09:45:05.139032Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Ensemble & Submission","metadata":{}},{"cell_type":"code","source":"submission = pd.read_csv('/kaggle/input/google-research-identify-contrails-reduce-global-warming/sample_submission.csv', index_col='record_id')","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:05.141824Z","iopub.execute_input":"2023-06-17T09:45:05.142275Z","iopub.status.idle":"2023-06-17T09:45:05.151922Z","shell.execute_reply.started":"2023-06-17T09:45:05.142232Z","shell.execute_reply":"2023-06-17T09:45:05.150865Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def ensemble(p1, p2):\n w1 = ratio_1 * (1 / thr_1)\n w2 = ratio_2 * (1 / thr_2)\n thr = 1 # fixed \n p = p1*w1 + p2*w2\n p = (p > thr).astype(np.int32)\n return p","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:05.153267Z","iopub.execute_input":"2023-06-17T09:45:05.15361Z","iopub.status.idle":"2023-06-17T09:45:05.166879Z","shell.execute_reply.started":"2023-06-17T09:45:05.153582Z","shell.execute_reply":"2023-06-17T09:45:05.165702Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for img_id, p1, p2 in zip(Ids, Probs_1, Probs_2):\n submission.loc[int(img_id), 'encoded_pixels'] = list_to_string(rle_encode(ensemble(p1, p2)))","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:05.16862Z","iopub.execute_input":"2023-06-17T09:45:05.168953Z","iopub.status.idle":"2023-06-17T09:45:05.185473Z","shell.execute_reply.started":"2023-06-17T09:45:05.168926Z","shell.execute_reply":"2023-06-17T09:45:05.184219Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission.head()","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:05.186603Z","iopub.execute_input":"2023-06-17T09:45:05.186918Z","iopub.status.idle":"2023-06-17T09:45:05.201204Z","shell.execute_reply.started":"2023-06-17T09:45:05.186892Z","shell.execute_reply":"2023-06-17T09:45:05.200022Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission.to_csv('submission.csv')","metadata":{"execution":{"iopub.status.busy":"2023-06-17T09:45:05.20299Z","iopub.execute_input":"2023-06-17T09:45:05.203466Z","iopub.status.idle":"2023-06-17T09:45:05.216398Z","shell.execute_reply.started":"2023-06-17T09:45:05.203426Z","shell.execute_reply":"2023-06-17T09:45:05.214943Z"},"trusted":true},"execution_count":null,"outputs":[]}]}