SYMBOL INDEX (1119 symbols across 103 files)

FILE: MaksimovKA_solution/augmentations/transforms.py
  function augmentations (line 13) | def augmentations(prob=0.5):

FILE: MaksimovKA_solution/dataset/base_dataset.py
  class BaseMaskDatasetIterator (line 7) | class BaseMaskDatasetIterator(Iterator):
    method __init__ (line 8) | def __init__(self,
    method pad_mask_image (line 32) | def pad_mask_image(self, mask, image, img_id, crop_shape):
    method transform_batch_y (line 36) | def transform_batch_y(self, batch_y):
    method _get_batches_of_transformed_samples (line 39) | def _get_batches_of_transformed_samples(self, index_array):
    method preprocess_input (line 81) | def preprocess_input(batch_x):
    method transform_batch_x (line 88) | def transform_batch_x(self, batch_x):
    method next (line 91) | def next(self):

FILE: MaksimovKA_solution/dataset/spacenet_binary_dataset.py
  class SpacenetBinaryDataset (line 6) | class SpacenetBinaryDataset:
    method __init__ (line 7) | def __init__(self,
    method get_generator (line 26) | def get_generator(self, image_ids, crop_shape, preprocessing_function=1,
    method train_generator (line 42) | def train_generator(self, crop_shape, preprocessing_function=1, random...
    method val_generator (line 46) | def val_generator(self, preprocessing_function=1, batch_size=1):
    method generate_ids (line 49) | def generate_ids(self):
  class SpacenetDatasetIterator (line 61) | class SpacenetDatasetIterator(BaseMaskDatasetIterator):
    method __init__ (line 63) | def __init__(self,
    method pad_mask_image (line 87) | def pad_mask_image(self, mask, image, img_id, crop_shape):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/backbones.py
  function get_backbone (line 31) | def get_backbone(name, *args, **kwargs):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnet/blocks.py
  function handle_block_names (line 11) | def handle_block_names(stage, block):
  function basic_identity_block (line 20) | def basic_identity_block(filters, stage, block):
  function basic_conv_block (line 53) | def basic_conv_block(filters, stage, block, strides=(2, 2)):
  function conv_block (line 89) | def conv_block(filters, stage, block, strides=(2, 2)):
  function identity_block (line 128) | def identity_block(filters, stage, block):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnet/builder.py
  function build_resnet (line 28) | def build_resnet(

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnet/models.py
  function ResNet18 (line 6) | def ResNet18(input_shape, input_tensor=None, weights=None, classes=1000,...
  function ResNet34 (line 20) | def ResNet34(input_shape, input_tensor=None, weights=None, classes=1000,...
  function ResNet50 (line 34) | def ResNet50(input_shape, input_tensor=None, weights=None, classes=1000,...
  function ResNet101 (line 47) | def ResNet101(input_shape, input_tensor=None, weights=None, classes=1000...
  function ResNet152 (line 60) | def ResNet152(input_shape, input_tensor=None, weights=None, classes=1000...

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnet/params.py
  function get_conv_params (line 5) | def get_conv_params(**params):
  function get_bn_params (line 15) | def get_bn_params(**params):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnet/preprocessing.py
  function preprocess_input (line 4) | def preprocess_input(x, size=None, BGRTranspose=True):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnext/blocks.py
  function handle_block_names (line 13) | def handle_block_names(stage, block):
  function GroupConv2D (line 22) | def GroupConv2D(filters, kernel_size, conv_params, conv_name, strides=(1...
  function conv_block (line 41) | def conv_block(filters, stage, block, strides=(2, 2)):
  function identity_block (line 81) | def identity_block(filters, stage, block):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnext/builder.py
  function build_resnext (line 29) | def build_resnext(

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnext/models.py
  function ResNeXt50 (line 6) | def ResNeXt50(input_shape, input_tensor=None, weights=None, classes=1000...
  function ResNeXt101 (line 20) | def ResNeXt101(input_shape, input_tensor=None, weights=None, classes=100...

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnext/params.py
  function get_conv_params (line 5) | def get_conv_params(**params):
  function get_bn_params (line 15) | def get_bn_params(**params):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/resnext/preprocessing.py
  function preprocess_input (line 4) | def preprocess_input(x, size=None):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/classification_models/utils.py
  function find_weights (line 4) | def find_weights(weights_collection, model_name, dataset, include_top):
  function load_model_weights (line 11) | def load_model_weights(weights_collection, model, dataset, classes, incl...

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/classification_models/tests/test_imagenet.py
  function get_top (line 101) | def get_top(y, top=5):
  function is_equal (line 109) | def is_equal(gt, pr, eps=10e-5):
  function test_model (line 123) | def test_model(model, preprocessing_func, sample, ground_truth):
  function main (line 141) | def main():

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/inception_resnet_v2.py
  function preprocess_input (line 49) | def preprocess_input(x):
  function conv2d_bn (line 59) | def conv2d_bn(x,
  function inception_resnet_block (line 97) | def inception_resnet_block(x, scale, block_type, block_idx, activation='...
  function InceptionResNetV2 (line 173) | def InceptionResNetV2(include_top=True,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/backbones/inception_v3.py
  function conv2d_bn (line 45) | def conv2d_bn(x,
  function InceptionV3 (line 87) | def InceptionV3(include_top=True,
  function preprocess_input (line 395) | def preprocess_input(x):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/common/blocks.py
  function Conv2DBlock (line 6) | def Conv2DBlock(n_filters, kernel_size,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/common/functions.py
  function transpose_shape (line 5) | def transpose_shape(shape, target_format, spatial_axes):
  function permute_dimensions (line 40) | def permute_dimensions(x, pattern):
  function int_shape (line 52) | def int_shape(x):
  function resize_images (line 67) | def resize_images(x,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/common/layers.py
  class ResizeImage (line 10) | class ResizeImage(Layer):
    method __init__ (line 45) | def __init__(self, factor=(2, 2), data_format='channels_last', interpo...
    method compute_output_shape (line 55) | def compute_output_shape(self, input_shape):
    method call (line 71) | def call(self, inputs):
    method get_config (line 75) | def get_config(self):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/fpn/blocks.py
  function Conv (line 14) | def Conv(n_filters, kernel_size, activation='relu', use_batchnorm=False,...
  function pyramid_block (line 27) | def pyramid_block(pyramid_filters=256, segmentation_filters=128, upsampl...

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/fpn/builder.py
  function build_fpn (line 21) | def build_fpn(backbone,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/fpn/layers.py
  function transpose_shape (line 10) | def transpose_shape(shape, target_format, spatial_axes):
  function permute_dimensions (line 45) | def permute_dimensions(x, pattern):
  function int_shape (line 57) | def int_shape(x):
  function resize_images (line 72) | def resize_images(x,
  class UpSampling2D (line 125) | class UpSampling2D(Layer):
    method __init__ (line 160) | def __init__(self, size=(2, 2), data_format='channels_last', interpola...
    method compute_output_shape (line 170) | def compute_output_shape(self, input_shape):
    method call (line 186) | def call(self, inputs):
    method get_config (line 190) | def get_config(self):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/fpn/model.py
  function FPN (line 24) | def FPN(backbone_name='vgg16',

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/linknet/blocks.py
  function handle_block_names (line 10) | def handle_block_names(stage):
  function ConvRelu (line 18) | def ConvRelu(filters,
  function Conv2DUpsample (line 42) | def Conv2DUpsample(filters,
  function Conv2DTranspose (line 60) | def Conv2DTranspose(filters,
  function UpsampleBlock (line 80) | def UpsampleBlock(filters,
  function DecoderBlock (line 121) | def DecoderBlock(stage,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/linknet/builder.py
  function build_linknet (line 9) | def build_linknet(backbone,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/linknet/model.py
  function Linknet (line 24) | def Linknet(backbone_name='vgg16',

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/pspnet/blocks.py
  function InterpBlock (line 13) | def InterpBlock(level, feature_map_shape,
  function DUC (line 53) | def DUC(factor=(8, 8)):
  function PyramidPoolingModule (line 78) | def PyramidPoolingModule(**params):

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/pspnet/builder.py
  function build_psp (line 23) | def build_psp(backbone,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/pspnet/model.py
  function _get_layer_by_factor (line 23) | def _get_layer_by_factor(backbone_name, factor):
  function _shape_guard (line 35) | def _shape_guard(factor, shape):
  function PSPNet (line 47) | def PSPNet(backbone_name='vgg16',

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/unet/blocks.py
  function handle_block_names (line 9) | def handle_block_names(stage):
  function ConvRelu (line 17) | def ConvRelu(filters, kernel_size, use_batchnorm=False, conv_name='conv'...
  function Upsample2D_block (line 27) | def Upsample2D_block(filters, stage, kernel_size=(3,3), upsample_rate=(2...
  function Transpose2D_block (line 49) | def Transpose2D_block(filters, stage, kernel_size=(3,3), upsample_rate=(...

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/unet/builder.py
  function build_unet (line 10) | def build_unet(backbone, classes, skip_connection_layers,

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/unet/model.py
  function Unet (line 24) | def Unet(backbone_name='vgg16',

FILE: MaksimovKA_solution/models/qubvel_segmentation_models/utils.py
  function get_layer_number (line 5) | def get_layer_number(model, layer_name):
  function extract_outputs (line 24) | def extract_outputs(model, layers, include_top=False):
  function reverse (line 45) | def reverse(l):
  function add_docstring (line 51) | def add_docstring(doc_string=None):
  function recompile (line 65) | def recompile(model):
  function freeze_model (line 69) | def freeze_model(model):
  function set_trainable (line 75) | def set_trainable(model):
  function to_tuple (line 81) | def to_tuple(x):

FILE: MaksimovKA_solution/predict/predict_segmentation.py
  function _remove_interiors (line 23) | def _remove_interiors(line):
  function my_watershed (line 35) | def my_watershed(what, mask1, mask2):
  function wsh (line 41) | def wsh(mask_img, threshold, border_img, seeds, shift):
  function preprocess_input (line 56) | def preprocess_input(batch_x):

FILE: MaksimovKA_solution/train/model_factory.py
  function make_model (line 5) | def make_model(network, freeze_encoder=1, predict_flag=0):

FILE: MaksimovKA_solution/train/train_segmentation.py
  class ModelCheckpointMGPU (line 14) | class ModelCheckpointMGPU(ModelCheckpoint):
    method __init__ (line 15) | def __init__(self, original_model, filepath,
    method on_epoch_end (line 21) | def on_epoch_end(self, epoch, logs=None):
  function main (line 26) | def main():

FILE: MaksimovKA_solution/utils/losses.py
  function binary_crossentropy (line 4) | def binary_crossentropy(y, p):
  function dice_coef (line 8) | def dice_coef(y_true, y_pred, smooth=1):
  function dice_coef_loss (line 15) | def dice_coef_loss(y_true, y_pred):
  function dice_coef_loss_bce (line 19) | def dice_coef_loss_bce(y_true, y_pred, dice=0.5, bce=0.5):
  function jacard_coef (line 23) | def jacard_coef(y_true, y_pred, smooth=1e-3):
  function jacard_coef_loss (line 30) | def jacard_coef_loss(y_true, y_pred):
  function double_head (line 34) | def double_head(y_true, y_pred, instance=1.0, border=1.0):
  function double_head_changed (line 39) | def double_head_changed(y_true, y_pred, instance=1.0, border=1.0):
  function make_loss (line 45) | def make_loss(loss_name):

FILE: MaksimovKA_solution/utils/metrics.py
  function hard_dice_coef_mask (line 5) | def hard_dice_coef_mask(y_true, y_pred, smooth=1e-3):
  function hard_jacard_coef_mask (line 12) | def hard_jacard_coef_mask(y_true, y_pred, smooth=1e-3):
  function hard_dice_coef_border (line 20) | def hard_dice_coef_border(y_true, y_pred, smooth=1e-3):
  function hard_jacard_coef_border (line 27) | def hard_jacard_coef_border(y_true, y_pred, smooth=1e-3):
  function calc_iou (line 35) | def calc_iou(gt_masks, predicted_masks, height=768, width=768):
  function precision_at (line 70) | def precision_at(threshold, iou):

FILE: XD_XD/main.py
  class conv_relu (line 57) | class conv_relu(nn.Module):
    method __init__ (line 58) | def __init__(self, in_, out):
    method forward (line 63) | def forward(self, x):
  class decoder_block (line 69) | class decoder_block(nn.Module):
    method __init__ (line 70) | def __init__(self, in_channels, middle_channels, out_channels):
    method forward (line 79) | def forward(self, x):
  class unet_vgg16 (line 83) | class unet_vgg16(nn.Module):
    method __init__ (line 84) | def __init__(self, num_filters=32, pretrained=False):
    method forward (line 116) | def forward(self, x):
  function get_image (line 132) | def get_image(imageid, basepath='/wdata/dataset', rgbdir='train_rgb'):
  class AtlantaDataset (line 139) | class AtlantaDataset(Dataset):
    method __init__ (line 140) | def __init__(self, image_ids, aug=None, basepath='/wdata/dataset'):
    method __len__ (line 145) | def __len__(self):
    method __getitem__ (line 148) | def __getitem__(self, idx):
  class AtlantaTestDataset (line 170) | class AtlantaTestDataset(Dataset):
    method __init__ (line 171) | def __init__(self, image_ids, aug=None, basepath='/wdata/dataset'):
    method __len__ (line 176) | def __len__(self):
    method __getitem__ (line 179) | def __getitem__(self, idx):
  function cli (line 189) | def cli():
  function check (line 196) | def check(inputs):
  function preproctrain (line 206) | def preproctrain(inputs, working_dir):
  function masks_from_geojson (line 234) | def masks_from_geojson(mask_dir, inputs, ref_name, geojson_fn):
  function read_cv_splits (line 245) | def read_cv_splits(inputs):
  function train (line 273) | def train(inputs, working_dir, fold_id):
  function validation (line 447) | def validation(model, criterion, val_loader,
  class Metrics (line 503) | class Metrics(object):
  class binary_loss (line 509) | class binary_loss(object):
    method __init__ (line 510) | def __init__(self, jaccard_weight=0):
    method __call__ (line 516) | def __call__(self, outputs, targets):
  function save (line 536) | def save(model, epoch, step, model_name):
  function copy_best (line 547) | def copy_best(model, epoch, model_name, step):
  function write_event (line 553) | def write_event(log, **data):
  function open_log (line 560) | def open_log(model_name):
  function make_train_val_loader (line 568) | def make_train_val_loader(train_transformer,
  function inference (line 605) | def inference(inputs, working_dir, output):
  function make_sub (line 637) | def make_sub(model_names, test_collection, output_fn):  # noqa: C901
  function __createCSVSummaryFile (line 719) | def __createCSVSummaryFile(chipSummaryList, outputFileName, pixPrecision...
  function inference_by_model (line 753) | def inference_by_model(model_name, filenames,
  function preproctest (line 849) | def preproctest(inputs, working_dir):
  function pan_to_bgr (line 872) | def pan_to_bgr(src, dst, thresh=3000):
  function filecheck (line 889) | def filecheck(inputs, working_dir):
  function filecheck_inference_models (line 900) | def filecheck_inference_models(working_dir):
  function filecheck_inference_images (line 923) | def filecheck_inference_images(working_dir):
  function __filecheck (line 935) | def __filecheck(path, max_length=80):
  function systemcheck_inference (line 951) | def systemcheck_inference():
  function systemcheck_train (line 956) | def systemcheck_train():
  function helper_assertion_check (line 963) | def helper_assertion_check(msg, res, max_length=80):

FILE: cannab/adamw.py
  class AdamW (line 6) | class AdamW(torch.optim.Optimizer):
    method __init__ (line 25) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8,
    method step (line 31) | def step(self, closure=None):

FILE: cannab/convert_test.py
  function process_image (line 34) | def process_image(img_id):

FILE: cannab/create_masks.py
  function mask_for_polygon (line 35) | def mask_for_polygon(poly, im_size):
  function process_image (line 50) | def process_image(img_id):

FILE: cannab/create_submission_lgbm.py
  function mask_to_polygons (line 23) | def mask_to_polygons(mask, min_area=8.):

FILE: cannab/losses.py
  function dice_round (line 14) | def dice_round(preds, trues):
  function soft_dice_loss (line 18) | def soft_dice_loss(outputs, targets, per_image=False):
  function jaccard (line 30) | def jaccard(outputs, targets, per_image=True, non_empty=False, min_pixel...
  class DiceLoss (line 54) | class DiceLoss(nn.Module):
    method __init__ (line 55) | def __init__(self, weight=None, size_average=True, per_image=False):
    method forward (line 61) | def forward(self, input, target):
  class JaccardLoss (line 64) | class JaccardLoss(nn.Module):
    method __init__ (line 65) | def __init__(self, weight=None, size_average=True, per_image=False, no...
    method forward (line 75) | def forward(self, input, target):
  class StableBCELoss (line 80) | class StableBCELoss(nn.Module):
    method __init__ (line 81) | def __init__(self):
    method forward (line 84) | def forward(self, input, target):
  class ComboLoss (line 92) | class ComboLoss(nn.Module):
    method __init__ (line 93) | def __init__(self, weights, per_image=False):
    method forward (line 111) | def forward(self, outputs, targets):
  function lovasz_grad (line 123) | def lovasz_grad(gt_sorted):
  function lovasz_hinge (line 137) | def lovasz_hinge(logits, labels, per_image=True, ignore=None):
  function lovasz_hinge_flat (line 152) | def lovasz_hinge_flat(logits, labels):
  function flatten_binary_scores (line 171) | def flatten_binary_scores(scores, labels, ignore=None):
  function lovasz_sigmoid (line 185) | def lovasz_sigmoid(probas, labels, per_image=False, ignore=None):
  function lovasz_sigmoid_flat (line 201) | def lovasz_sigmoid_flat(probas, labels):
  function mean (line 216) | def mean(l, ignore_nan=False, empty=0):
  class LovaszLoss (line 236) | class LovaszLoss(nn.Module):
    method __init__ (line 237) | def __init__(self, ignore_index=255, per_image=True):
    method forward (line 242) | def forward(self, outputs, targets):
  class LovaszLossSigmoid (line 247) | class LovaszLossSigmoid(nn.Module):
    method __init__ (line 248) | def __init__(self, ignore_index=255, per_image=True):
    method forward (line 253) | def forward(self, outputs, targets):
  class FocalLoss2d (line 259) | class FocalLoss2d(nn.Module):
    method __init__ (line 260) | def __init__(self, gamma=2, ignore_index=255):
    method forward (line 265) | def forward(self, outputs, targets):

FILE: cannab/merge_oof.py
  function process_image (line 31) | def process_image(fid):

FILE: cannab/train101_9ch_fold.py
  function shift_image (line 48) | def shift_image(img, shift_pnt):
  function rotate_image (line 53) | def rotate_image(image, angle, scale, rot_pnt):
  function gauss_noise (line 58) | def gauss_noise(img, var=30):
  function clahe (line 67) | def clahe(img, clipLimit=2.0, tileGridSize=(5,5)):
  function _blend (line 74) | def _blend(img1, img2, alpha):
  function _grayscale (line 78) | def _grayscale(img):
  function saturation (line 81) | def saturation(img, alpha):
  function brightness (line 85) | def brightness(img, alpha):
  function contrast (line 89) | def contrast(img, alpha):
  class TrainData (line 94) | class TrainData(Dataset):
    method __init__ (line 95) | def __init__(self, image_ids, epoch_size):
    method __len__ (line 101) | def __len__(self):
    method __getitem__ (line 104) | def __getitem__(self, idx):
  class ValData (line 208) | class ValData(Dataset):
    method __init__ (line 209) | def __init__(self, image_ids):
    method __len__ (line 213) | def __len__(self):
    method __getitem__ (line 216) | def __getitem__(self, idx):
  class AverageMeter (line 245) | class AverageMeter(object):
    method __init__ (line 247) | def __init__(self):
    method reset (line 249) | def reset(self):
    method update (line 254) | def update(self, val, n=1):
  function validate (line 261) | def validate(net, data_loader):
  function evaluate_val (line 280) | def evaluate_val(data_val, best_score, model, snapshot_name, current_epo...
  function train_epoch (line 296) | def train_epoch(current_epoch, loss_function, l1_loss, model, optimizer,...

FILE: cannab/train154_9ch_fold.py
  function shift_image (line 48) | def shift_image(img, shift_pnt):
  function rotate_image (line 53) | def rotate_image(image, angle, scale, rot_pnt):
  function gauss_noise (line 58) | def gauss_noise(img, var=30):
  function clahe (line 67) | def clahe(img, clipLimit=2.0, tileGridSize=(5,5)):
  function _blend (line 74) | def _blend(img1, img2, alpha):
  function _grayscale (line 78) | def _grayscale(img):
  function saturation (line 81) | def saturation(img, alpha):
  function brightness (line 85) | def brightness(img, alpha):
  function contrast (line 89) | def contrast(img, alpha):
  class TrainData (line 94) | class TrainData(Dataset):
    method __init__ (line 95) | def __init__(self, image_ids, epoch_size):
    method __len__ (line 102) | def __len__(self):
    method __getitem__ (line 105) | def __getitem__(self, idx):
  class ValData (line 213) | class ValData(Dataset):
    method __init__ (line 214) | def __init__(self, image_ids):
    method __len__ (line 218) | def __len__(self):
    method __getitem__ (line 221) | def __getitem__(self, idx):
  class AverageMeter (line 250) | class AverageMeter(object):
    method __init__ (line 252) | def __init__(self):
    method reset (line 254) | def reset(self):
    method update (line 259) | def update(self, val, n=1):
  function validate (line 266) | def validate(net, data_loader):
  function evaluate_val (line 285) | def evaluate_val(data_val, best_score, model, snapshot_name, current_epo...
  function train_epoch (line 301) | def train_epoch(current_epoch, loss_function, l1_loss, model, optimizer,...

FILE: cannab/train50_9ch_fold.py
  function shift_image (line 48) | def shift_image(img, shift_pnt):
  function rotate_image (line 53) | def rotate_image(image, angle, scale, rot_pnt):
  function gauss_noise (line 58) | def gauss_noise(img, var=30):
  function clahe (line 67) | def clahe(img, clipLimit=2.0, tileGridSize=(5,5)):
  function _blend (line 74) | def _blend(img1, img2, alpha):
  function _grayscale (line 78) | def _grayscale(img):
  function saturation (line 81) | def saturation(img, alpha):
  function brightness (line 85) | def brightness(img, alpha):
  function contrast (line 89) | def contrast(img, alpha):
  class TrainData (line 94) | class TrainData(Dataset):
    method __init__ (line 95) | def __init__(self, image_ids, epoch_size):
    method __len__ (line 101) | def __len__(self):
    method __getitem__ (line 104) | def __getitem__(self, idx):
  class ValData (line 209) | class ValData(Dataset):
    method __init__ (line 210) | def __init__(self, image_ids):
    method __len__ (line 214) | def __len__(self):
    method __getitem__ (line 217) | def __getitem__(self, idx):
  class AverageMeter (line 246) | class AverageMeter(object):
    method __init__ (line 248) | def __init__(self):
    method reset (line 250) | def reset(self):
    method update (line 255) | def update(self, val, n=1):
  function validate (line 262) | def validate(net, data_loader):
  function evaluate_val (line 281) | def evaluate_val(data_val, best_score, model, snapshot_name, current_epo...
  function train_epoch (line 297) | def train_epoch(current_epoch, loss_function, l1_loss, model, optimizer,...

FILE: cannab/train92_9ch_fold.py
  function shift_image (line 48) | def shift_image(img, shift_pnt):
  function rotate_image (line 53) | def rotate_image(image, angle, scale, rot_pnt):
  function gauss_noise (line 58) | def gauss_noise(img, var=30):
  function clahe (line 67) | def clahe(img, clipLimit=2.0, tileGridSize=(5,5)):
  function _blend (line 74) | def _blend(img1, img2, alpha):
  function _grayscale (line 78) | def _grayscale(img):
  function saturation (line 81) | def saturation(img, alpha):
  function brightness (line 85) | def brightness(img, alpha):
  function contrast (line 89) | def contrast(img, alpha):
  class TrainData (line 94) | class TrainData(Dataset):
    method __init__ (line 95) | def __init__(self, image_ids, epoch_size):
    method __len__ (line 101) | def __len__(self):
    method __getitem__ (line 104) | def __getitem__(self, idx):
  class ValData (line 209) | class ValData(Dataset):
    method __init__ (line 210) | def __init__(self, image_ids):
    method __len__ (line 214) | def __len__(self):
    method __getitem__ (line 217) | def __getitem__(self, idx):
  class AverageMeter (line 245) | class AverageMeter(object):
    method __init__ (line 247) | def __init__(self):
    method reset (line 249) | def reset(self):
    method update (line 254) | def update(self, val, n=1):
  function validate (line 261) | def validate(net, data_loader):
  function evaluate_val (line 280) | def evaluate_val(data_val, best_score, model, snapshot_name, current_epo...
  function train_epoch (line 296) | def train_epoch(current_epoch, loss_function, l1_loss, model, optimizer,...

FILE: cannab/train_classifier.py
  function get_inputs (line 47) | def get_inputs(filename, pred_folder, add_features=[], return_labels=Fal...

FILE: cannab/utils.py
  function parse_img_id (line 9) | def parse_img_id(img_id):
  function preprocess_inputs (line 54) | def preprocess_inputs(x):
  function dice (line 60) | def dice(im1, im2, empty_score=1.0):

FILE: cannab/zoo/dpn.py
  function dpn68 (line 97) | def dpn68(num_classes=1000, pretrained='imagenet'):
  function dpn68b (line 115) | def dpn68b(num_classes=1000, pretrained='imagenet+5k'):
  function dpn92 (line 133) | def dpn92(num_classes=1000, pretrained='imagenet+5k'):
  function dpn98 (line 151) | def dpn98(num_classes=1000, pretrained='imagenet'):
  function dpn131 (line 169) | def dpn131(num_classes=1000, pretrained='imagenet'):
  function dpn107 (line 187) | def dpn107(num_classes=1000, pretrained='imagenet+5k'):
  class CatBnAct (line 206) | class CatBnAct(nn.Module):
    method __init__ (line 207) | def __init__(self, in_chs, activation_fn=nn.ReLU(inplace=True)):
    method forward (line 212) | def forward(self, x):
  class BnActConv2d (line 217) | class BnActConv2d(nn.Module):
    method __init__ (line 218) | def __init__(self, in_chs, out_chs, kernel_size, stride,
    method forward (line 225) | def forward(self, x):
  class InputBlock (line 229) | class InputBlock(nn.Module):
    method __init__ (line 230) | def __init__(self, num_init_features, kernel_size=7,
    method forward (line 239) | def forward(self, x):
  class DualPathBlock (line 247) | class DualPathBlock(nn.Module):
    method __init__ (line 248) | def __init__(
    method forward (line 284) | def forward(self, x):
  class DPN (line 311) | class DPN(nn.Module):
    method __init__ (line 312) | def __init__(self, small=False, num_init_features=64, k_r=96, groups=32,
    method logits (line 381) | def logits(self, features):
    method forward (line 392) | def forward(self, input):
  function pooling_factor (line 409) | def pooling_factor(pool_type='avg'):
  function adaptive_avgmax_pool2d (line 413) | def adaptive_avgmax_pool2d(x, pool_type='avg', padding=0, count_include_...
  class AdaptiveAvgMaxPool2d (line 437) | class AdaptiveAvgMaxPool2d(torch.nn.Module):
    method __init__ (line 440) | def __init__(self, output_size=1, pool_type='avg'):
    method forward (line 453) | def forward(self, x):
    method factor (line 462) | def factor(self):
    method __repr__ (line 465) | def __repr__(self):

FILE: cannab/zoo/models.py
  class ConvReluBN (line 8) | class ConvReluBN(nn.Module):
    method __init__ (line 9) | def __init__(self, in_channels, out_channels, kernel_size=3):
    method forward (line 16) | def forward(self, x):
  class ConvRelu (line 19) | class ConvRelu(nn.Module):
    method __init__ (line 20) | def __init__(self, in_channels, out_channels, kernel_size=3):
    method forward (line 26) | def forward(self, x):
  class SCSEModule (line 29) | class SCSEModule(nn.Module):
    method __init__ (line 31) | def __init__(self, channels, reduction=16, concat=False):
    method forward (line 46) | def forward(self, x):
  class SeResNext50_9ch_Unet (line 64) | class SeResNext50_9ch_Unet(nn.Module):
    method __init__ (line 65) | def __init__(self, pretrained='imagenet', **kwargs):
    method forward (line 100) | def forward(self, x, y, cat_inp, coord_inp):
    method _initialize_weights (line 161) | def _initialize_weights(self):
  class Dpn92_9ch_Unet (line 172) | class Dpn92_9ch_Unet(nn.Module):
    method __init__ (line 173) | def __init__(self, pretrained='imagenet+5k', **kwargs):
    method forward (line 215) | def forward(self, x, y, cat_inp, coord_inp):
    method _initialize_weights (line 282) | def _initialize_weights(self):
  class ScSeSenet154_9ch_Unet (line 293) | class ScSeSenet154_9ch_Unet(nn.Module):
    method __init__ (line 294) | def __init__(self, pretrained='imagenet', **kwargs):
    method forward (line 328) | def forward(self, x, y, cat_inp, coord_inp):
    method _initialize_weights (line 389) | def _initialize_weights(self):
  class ScSeResNext101_9ch_Unet (line 400) | class ScSeResNext101_9ch_Unet(nn.Module):
    method __init__ (line 401) | def __init__(self, pretrained='imagenet', **kwargs):
    method forward (line 436) | def forward(self, x, y, cat_inp, coord_inp):
    method _initialize_weights (line 497) | def _initialize_weights(self):

FILE: cannab/zoo/senet.py
  class SEModule (line 86) | class SEModule(nn.Module):
    method __init__ (line 88) | def __init__(self, channels, reduction, concat=False):
    method forward (line 98) | def forward(self, x):
  class SCSEModule (line 107) | class SCSEModule(nn.Module):
    method __init__ (line 109) | def __init__(self, channels, reduction=16, concat=False):
    method forward (line 124) | def forward(self, x):
  class Bottleneck (line 141) | class Bottleneck(nn.Module):
    method forward (line 145) | def forward(self, x):
  class SEBottleneck (line 168) | class SEBottleneck(Bottleneck):
    method __init__ (line 174) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SCSEBottleneck (line 192) | class SCSEBottleneck(Bottleneck):
    method __init__ (line 198) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SEResNetBottleneck (line 216) | class SEResNetBottleneck(Bottleneck):
    method __init__ (line 224) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SEResNeXtBottleneck (line 241) | class SEResNeXtBottleneck(Bottleneck):
    method __init__ (line 247) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SCSEResNeXtBottleneck (line 266) | class SCSEResNeXtBottleneck(Bottleneck):
    method __init__ (line 272) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SENet (line 290) | class SENet(nn.Module):
    method __init__ (line 292) | def __init__(self, block, layers, groups, reduction, dropout_p=0.2,
    method _make_layer (line 410) | def _make_layer(self, block, planes, blocks, groups, reduction, stride=1,
    method _initialize_weights (line 430) | def _initialize_weights(self):
    method features (line 440) | def features(self, x):
    method logits (line 449) | def logits(self, x):
    method forward (line 457) | def forward(self, x):
  function initialize_pretrained_model (line 463) | def initialize_pretrained_model(model, num_classes, settings):
  function senet154 (line 475) | def senet154(num_classes=1000, pretrained='imagenet'):
  function scsenet154 (line 483) | def scsenet154(num_classes=1000, pretrained='imagenet'):
  function se_resnet50 (line 493) | def se_resnet50(num_classes=1000, pretrained='imagenet'):
  function se_resnet101 (line 504) | def se_resnet101(num_classes=1000, pretrained='imagenet'):
  function se_resnet152 (line 515) | def se_resnet152(num_classes=1000, pretrained='imagenet'):
  function se_resnext50_32x4d (line 526) | def se_resnext50_32x4d(num_classes=1000, pretrained='imagenet'):
  function scse_resnext50_32x4d (line 537) | def scse_resnext50_32x4d(num_classes=1000, pretrained='imagenet'):
  function se_resnext101_32x4d (line 548) | def se_resnext101_32x4d(num_classes=1000, pretrained='imagenet'):

FILE: number13/number13/src/Mask_RCNN/mrcnn/config.py
  class Config (line 18) | class Config(object):
    method __init__ (line 193) | def __init__(self):
    method display (line 208) | def display(self):

FILE: number13/number13/src/Mask_RCNN/mrcnn/model.py
  function log (line 39) | def log(text, array=None):
  class BatchNorm (line 53) | class BatchNorm(KL.BatchNormalization):
    method call (line 61) | def call(self, inputs, training=None):
  function compute_backbone_shapes (line 71) | def compute_backbone_shapes(config, image_shape):
  function identity_block (line 92) | def identity_block(input_tensor, kernel_size, filters, stage, block,
  function conv_block (line 127) | def conv_block(input_tensor, kernel_size, filters, stage, block,
  function resnet_graph (line 168) | def resnet_graph(input_image, architecture, stage5=False, train_bn=True):
  function apply_box_deltas_graph (line 210) | def apply_box_deltas_graph(boxes, deltas):
  function clip_boxes_graph (line 234) | def clip_boxes_graph(boxes, window):
  class ProposalLayer (line 252) | class ProposalLayer(KE.Layer):
    method __init__ (line 267) | def __init__(self, proposal_count, nms_threshold, config=None, **kwargs):
    method call (line 273) | def call(self, inputs):
    method compute_output_shape (line 328) | def compute_output_shape(self, input_shape):
  function log2_graph (line 336) | def log2_graph(x):
  class PyramidROIAlign (line 341) | class PyramidROIAlign(KE.Layer):
    method __init__ (line 361) | def __init__(self, pool_shape, **kwargs):
    method call (line 365) | def call(self, inputs):
    method compute_output_shape (line 445) | def compute_output_shape(self, input_shape):
  function overlaps_graph (line 453) | def overlaps_graph(boxes1, boxes2):
  function detection_targets_graph (line 482) | def detection_targets_graph(proposals, gt_class_ids, gt_boxes, gt_masks,...
  class DetectionTargetLayer (line 616) | class DetectionTargetLayer(KE.Layer):
    method __init__ (line 643) | def __init__(self, config, **kwargs):
    method call (line 647) | def call(self, inputs):
    method compute_output_shape (line 663) | def compute_output_shape(self, input_shape):
    method compute_mask (line 672) | def compute_mask(self, inputs, mask=None):
  function refine_detections_graph (line 680) | def refine_detections_graph(rois, probs, deltas, window, config):
  class DetectionLayer (line 778) | class DetectionLayer(KE.Layer):
    method __init__ (line 787) | def __init__(self, config=None, **kwargs):
    method call (line 791) | def call(self, inputs):
    method compute_output_shape (line 818) | def compute_output_shape(self, input_shape):
  function rpn_graph (line 826) | def rpn_graph(feature_map, anchors_per_location, anchor_stride):
  function build_rpn_model (line 870) | def build_rpn_model(anchor_stride, anchors_per_location, depth):
  function fpn_classifier_graph (line 896) | def fpn_classifier_graph(rois, feature_maps, image_meta,
  function build_fpn_mask_graph (line 950) | def build_fpn_mask_graph(rois, feature_maps, image_meta,
  function smooth_l1_loss (line 1006) | def smooth_l1_loss(y_true, y_pred):
  function rpn_class_loss_graph (line 1016) | def rpn_class_loss_graph(rpn_match, rpn_class_logits):
  function rpn_bbox_loss_graph (line 1041) | def rpn_bbox_loss_graph(config, target_bbox, rpn_match, rpn_bbox):
  function mrcnn_class_loss_graph (line 1074) | def mrcnn_class_loss_graph(target_class_ids, pred_class_logits,
  function mrcnn_bbox_loss_graph (line 1110) | def mrcnn_bbox_loss_graph(target_bbox, target_class_ids, pred_bbox):
  function mrcnn_mask_loss_graph (line 1141) | def mrcnn_mask_loss_graph(target_masks, target_class_ids, pred_masks):
  function load_image_gt (line 1184) | def load_image_gt(dataset, config, image_id, augment=False, augmentation...
  function build_detection_targets (line 1287) | def build_detection_targets(rpn_rois, gt_class_ids, gt_boxes, gt_masks, ...
  function build_rpn_targets (line 1444) | def build_rpn_targets(image_shape, anchors, gt_class_ids, gt_boxes, conf...
  function generate_random_rois (line 1555) | def generate_random_rois(image_shape, count, gt_class_ids, gt_boxes):
  function data_generator (line 1629) | def data_generator(dataset, config, shuffle=True, augment=False, augment...
  class MaskRCNN (line 1806) | class MaskRCNN():
    method __init__ (line 1812) | def __init__(self, mode, config, model_dir):
    method build (line 1825) | def build(self, mode, config):
    method find_last (line 2046) | def find_last(self):
    method load_weights (line 2071) | def load_weights(self, filepath, by_name=False, exclude=None):
    method get_imagenet_weights (line 2109) | def get_imagenet_weights(self):
    method compile (line 2123) | def compile(self, learning_rate, momentum):
    method set_trainable (line 2171) | def set_trainable(self, layer_regex, keras_model=None, indent=0, verbo...
    method set_log_dir (line 2208) | def set_log_dir(self, model_path=None):
    method train (line 2244) | def train(self, train_dataset, val_dataset, learning_rate, epochs, lay...
    method mold_inputs (line 2332) | def mold_inputs(self, images):
    method unmold_detections (line 2371) | def unmold_detections(self, detections, mrcnn_mask, original_image_shape,
    method detect (line 2436) | def detect(self, images, verbose=0):
    method detect_molded (line 2494) | def detect_molded(self, molded_images, image_metas, verbose=0):
    method get_anchors (line 2552) | def get_anchors(self, image_shape):
    method ancestor (line 2574) | def ancestor(self, tensor, name, checked=None):
    method find_trainable_layer (line 2602) | def find_trainable_layer(self, layer):
    method get_trainable_layers (line 2611) | def get_trainable_layers(self):
    method run_graph (line 2623) | def run_graph(self, images, outputs, image_metas=None):
  function compose_image_meta (line 2679) | def compose_image_meta(image_id, original_image_shape, image_shape,
  function parse_image_meta (line 2704) | def parse_image_meta(meta):
  function parse_image_meta_graph (line 2728) | def parse_image_meta_graph(meta):
  function mold_image (line 2752) | def mold_image(images, config):
  function unmold_image (line 2760) | def unmold_image(normalized_images, config):
  function trim_zeros_graph (line 2769) | def trim_zeros_graph(boxes, name=None):
  function batch_pack_graph (line 2781) | def batch_pack_graph(x, counts, num_rows):
  function norm_boxes_graph (line 2791) | def norm_boxes_graph(boxes, shape):
  function denorm_boxes_graph (line 2808) | def denorm_boxes_graph(boxes, shape):

FILE: number13/number13/src/Mask_RCNN/mrcnn/model_mod_mpan.py
  function log (line 40) | def log(text, array=None):
  class WeightsSaver (line 54) | class WeightsSaver(Callback):
    method __init__ (line 55) | def __init__(self,snapshot_path, model, N):
    method on_batch_end (line 61) | def on_batch_end(self, batch, logs={}):
  class BatchNorm (line 71) | class BatchNorm(KL.BatchNormalization):
    method call (line 79) | def call(self, inputs, training=None):
  function compute_backbone_shapes (line 89) | def compute_backbone_shapes(config, image_shape):
  function bottom_up_agg (line 102) | def bottom_up_agg(Ps):
  function identity_block (line 119) | def identity_block(input_tensor, kernel_size, filters, stage, block,
  function conv_block (line 154) | def conv_block(input_tensor, kernel_size, filters, stage, block,
  function det_conv_block (line 195) | def det_conv_block(input_tensor, kernel_size, filters, stage, block,
  function resnet_graph (line 236) | def resnet_graph(input_image, architecture, stage5=False, train_bn=True,...
  function apply_box_deltas_graph (line 297) | def apply_box_deltas_graph(boxes, deltas):
  function clip_boxes_graph (line 321) | def clip_boxes_graph(boxes, window):
  class ProposalLayer (line 339) | class ProposalLayer(KE.Layer):
    method __init__ (line 354) | def __init__(self, proposal_count, nms_threshold, config=None, **kwargs):
    method call (line 360) | def call(self, inputs):
    method compute_output_shape (line 417) | def compute_output_shape(self, input_shape):
  function log2_graph (line 425) | def log2_graph(x):
  class PyramidROIAlign (line 430) | class PyramidROIAlign(KE.Layer):
    method __init__ (line 450) | def __init__(self, pool_shape, **kwargs):
    method call (line 454) | def call(self, inputs):
    method compute_output_shape (line 534) | def compute_output_shape(self, input_shape):
  function overlaps_graph (line 542) | def overlaps_graph(boxes1, boxes2):
  function detection_targets_graph (line 571) | def detection_targets_graph(proposals, gt_class_ids, gt_boxes, gt_masks,...
  class DetectionTargetLayer (line 748) | class DetectionTargetLayer(KE.Layer):
    method __init__ (line 775) | def __init__(self, config, **kwargs):
    method call (line 779) | def call(self, inputs):
    method compute_output_shape (line 795) | def compute_output_shape(self, input_shape):
    method compute_mask (line 804) | def compute_mask(self, inputs, mask=None):
  function refine_detections_graph (line 812) | def refine_detections_graph(rois, probs, deltas, window, config):
  class DetectionLayer (line 910) | class DetectionLayer(KE.Layer):
    method __init__ (line 919) | def __init__(self, config=None, **kwargs):
    method call (line 923) | def call(self, inputs):
    method compute_output_shape (line 950) | def compute_output_shape(self, input_shape):
  function rpn_graph (line 958) | def rpn_graph(feature_map, anchors_per_location, anchor_stride):
  function build_rpn_model (line 1002) | def build_rpn_model(anchor_stride, anchors_per_location, depth):
  function fpn_classifier_graph (line 1028) | def fpn_classifier_graph(rois, feature_maps, image_meta,
  function build_fpn_mask_graph (line 1092) | def build_fpn_mask_graph(rois, feature_maps, image_meta,
  function smooth_l1_loss (line 1148) | def smooth_l1_loss(y_true, y_pred):
  function rpn_class_loss_graph (line 1158) | def rpn_class_loss_graph(rpn_match, rpn_class_logits):
  function rpn_bbox_loss_graph (line 1187) | def rpn_bbox_loss_graph(config, target_bbox, rpn_match, rpn_bbox):
  function mrcnn_class_loss_graph (line 1220) | def mrcnn_class_loss_graph(target_class_ids, pred_class_logits,
  function mrcnn_bbox_loss_graph (line 1256) | def mrcnn_bbox_loss_graph(target_bbox, target_class_ids, pred_bbox):
  function mrcnn_mask_loss_graph (line 1287) | def mrcnn_mask_loss_graph(target_masks, target_class_ids, pred_masks):
  function load_image_gt (line 1330) | def load_image_gt(dataset, config, image_id, augment=False, augmentation...
  function build_detection_targets (line 1433) | def build_detection_targets(rpn_rois, gt_class_ids, gt_boxes, gt_masks, ...
  function build_rpn_targets (line 1590) | def build_rpn_targets(image_shape, anchors, gt_class_ids, gt_boxes, conf...
  function generate_random_rois (line 1701) | def generate_random_rois(image_shape, count, gt_class_ids, gt_boxes):
  function data_generator (line 1775) | def data_generator(dataset, config, shuffle=True, augment=False, augment...
  class MaskRCNN (line 1952) | class MaskRCNN():
    method __init__ (line 1958) | def __init__(self, mode, config, model_dir):
    method build (line 1971) | def build(self, mode, config):
    method find_last (line 2198) | def find_last(self):
    method load_weights (line 2223) | def load_weights(self, filepath, by_name=False, exclude=None):
    method get_imagenet_weights (line 2261) | def get_imagenet_weights(self):
    method compile (line 2275) | def compile(self, learning_rate, momentum):
    method set_trainable (line 2323) | def set_trainable(self, layer_regex, keras_model=None, indent=0, verbo...
    method set_log_dir (line 2360) | def set_log_dir(self, model_path=None):
    method train (line 2396) | def train(self, train_dataset, val_dataset, learning_rate, epochs, lay...
    method mold_inputs (line 2486) | def mold_inputs(self, images):
    method unmold_detections (line 2525) | def unmold_detections(self, detections, mrcnn_mask, original_image_shape,
    method detect (line 2590) | def detect(self, images, verbose=0):
    method detect_molded (line 2648) | def detect_molded(self, molded_images, image_metas, verbose=0):
    method get_anchors (line 2706) | def get_anchors(self, image_shape):
    method ancestor (line 2728) | def ancestor(self, tensor, name, checked=None):
    method find_trainable_layer (line 2756) | def find_trainable_layer(self, layer):
    method get_trainable_layers (line 2765) | def get_trainable_layers(self):
    method run_graph (line 2777) | def run_graph(self, images, outputs, image_metas=None):
  function compose_image_meta (line 2833) | def compose_image_meta(image_id, original_image_shape, image_shape,
  function parse_image_meta (line 2858) | def parse_image_meta(meta):
  function parse_image_meta_graph (line 2882) | def parse_image_meta_graph(meta):
  function mold_image (line 2906) | def mold_image(images, config):
  function unmold_image (line 2914) | def unmold_image(normalized_images, config):
  function trim_zeros_graph (line 2923) | def trim_zeros_graph(boxes, name=None):
  function batch_pack_graph (line 2935) | def batch_pack_graph(x, counts, num_rows):
  function norm_boxes_graph (line 2945) | def norm_boxes_graph(boxes, shape):
  function boost_boxes_graph (line 2962) | def boost_boxes_graph(boxes):
  function denorm_boxes_graph (line 2971) | def denorm_boxes_graph(boxes, shape):

FILE: number13/number13/src/Mask_RCNN/mrcnn/model_mod_rgb.py
  function log (line 40) | def log(text, array=None):
  class WeightsSaver (line 54) | class WeightsSaver(Callback):
    method __init__ (line 55) | def __init__(self,snapshot_path, model, N):
    method on_batch_end (line 61) | def on_batch_end(self, batch, logs={}):
  class BatchNorm (line 71) | class BatchNorm(KL.BatchNormalization):
    method call (line 79) | def call(self, inputs, training=None):
  function compute_backbone_shapes (line 89) | def compute_backbone_shapes(config, image_shape):
  function bottom_up_agg (line 102) | def bottom_up_agg(Ps):
  function identity_block (line 119) | def identity_block(input_tensor, kernel_size, filters, stage, block,
  function conv_block (line 154) | def conv_block(input_tensor, kernel_size, filters, stage, block,
  function det_conv_block (line 195) | def det_conv_block(input_tensor, kernel_size, filters, stage, block,
  function resnet_graph (line 236) | def resnet_graph(input_image, architecture, stage5=False, train_bn=True,...
  function apply_box_deltas_graph (line 297) | def apply_box_deltas_graph(boxes, deltas):
  function clip_boxes_graph (line 321) | def clip_boxes_graph(boxes, window):
  class ProposalLayer (line 339) | class ProposalLayer(KE.Layer):
    method __init__ (line 354) | def __init__(self, proposal_count, nms_threshold, config=None, **kwargs):
    method call (line 360) | def call(self, inputs):
    method compute_output_shape (line 417) | def compute_output_shape(self, input_shape):
  function log2_graph (line 425) | def log2_graph(x):
  class PyramidROIAlign (line 430) | class PyramidROIAlign(KE.Layer):
    method __init__ (line 450) | def __init__(self, pool_shape, **kwargs):
    method call (line 454) | def call(self, inputs):
    method compute_output_shape (line 534) | def compute_output_shape(self, input_shape):
  function overlaps_graph (line 542) | def overlaps_graph(boxes1, boxes2):
  function detection_targets_graph (line 571) | def detection_targets_graph(proposals, gt_class_ids, gt_boxes, gt_masks,...
  class DetectionTargetLayer (line 748) | class DetectionTargetLayer(KE.Layer):
    method __init__ (line 775) | def __init__(self, config, **kwargs):
    method call (line 779) | def call(self, inputs):
    method compute_output_shape (line 795) | def compute_output_shape(self, input_shape):
    method compute_mask (line 804) | def compute_mask(self, inputs, mask=None):
  function refine_detections_graph (line 812) | def refine_detections_graph(rois, probs, deltas, window, config):
  class DetectionLayer (line 910) | class DetectionLayer(KE.Layer):
    method __init__ (line 919) | def __init__(self, config=None, **kwargs):
    method call (line 923) | def call(self, inputs):
    method compute_output_shape (line 950) | def compute_output_shape(self, input_shape):
  function rpn_graph (line 958) | def rpn_graph(feature_map, anchors_per_location, anchor_stride):
  function build_rpn_model (line 1002) | def build_rpn_model(anchor_stride, anchors_per_location, depth):
  function fpn_classifier_graph (line 1028) | def fpn_classifier_graph(rois, feature_maps, image_meta,
  function build_fpn_mask_graph (line 1092) | def build_fpn_mask_graph(rois, feature_maps, image_meta,
  function smooth_l1_loss (line 1148) | def smooth_l1_loss(y_true, y_pred):
  function rpn_class_loss_graph (line 1158) | def rpn_class_loss_graph(rpn_match, rpn_class_logits):
  function rpn_bbox_loss_graph (line 1187) | def rpn_bbox_loss_graph(config, target_bbox, rpn_match, rpn_bbox):
  function mrcnn_class_loss_graph (line 1220) | def mrcnn_class_loss_graph(target_class_ids, pred_class_logits,
  function mrcnn_bbox_loss_graph (line 1256) | def mrcnn_bbox_loss_graph(target_bbox, target_class_ids, pred_bbox):
  function mrcnn_mask_loss_graph (line 1287) | def mrcnn_mask_loss_graph(target_masks, target_class_ids, pred_masks):
  function load_image_gt (line 1330) | def load_image_gt(dataset, config, image_id, augment=False, augmentation...
  function build_detection_targets (line 1433) | def build_detection_targets(rpn_rois, gt_class_ids, gt_boxes, gt_masks, ...
  function build_rpn_targets (line 1590) | def build_rpn_targets(image_shape, anchors, gt_class_ids, gt_boxes, conf...
  function generate_random_rois (line 1701) | def generate_random_rois(image_shape, count, gt_class_ids, gt_boxes):
  function data_generator (line 1775) | def data_generator(dataset, config, shuffle=True, augment=False, augment...
  class MaskRCNN (line 1952) | class MaskRCNN():
    method __init__ (line 1958) | def __init__(self, mode, config, model_dir):
    method build (line 1971) | def build(self, mode, config):
    method find_last (line 2198) | def find_last(self):
    method load_weights (line 2223) | def load_weights(self, filepath, by_name=False, exclude=None):
    method get_imagenet_weights (line 2261) | def get_imagenet_weights(self):
    method compile (line 2275) | def compile(self, learning_rate, momentum):
    method set_trainable (line 2323) | def set_trainable(self, layer_regex, keras_model=None, indent=0, verbo...
    method set_log_dir (line 2360) | def set_log_dir(self, model_path=None):
    method train (line 2396) | def train(self, train_dataset, val_dataset, learning_rate, epochs, lay...
    method mold_inputs (line 2486) | def mold_inputs(self, images):
    method unmold_detections (line 2525) | def unmold_detections(self, detections, mrcnn_mask, original_image_shape,
    method detect (line 2590) | def detect(self, images, verbose=0):
    method detect_molded (line 2648) | def detect_molded(self, molded_images, image_metas, verbose=0):
    method get_anchors (line 2706) | def get_anchors(self, image_shape):
    method ancestor (line 2728) | def ancestor(self, tensor, name, checked=None):
    method find_trainable_layer (line 2756) | def find_trainable_layer(self, layer):
    method get_trainable_layers (line 2765) | def get_trainable_layers(self):
    method run_graph (line 2777) | def run_graph(self, images, outputs, image_metas=None):
  function compose_image_meta (line 2833) | def compose_image_meta(image_id, original_image_shape, image_shape,
  function parse_image_meta (line 2858) | def parse_image_meta(meta):
  function parse_image_meta_graph (line 2882) | def parse_image_meta_graph(meta):
  function mold_image (line 2906) | def mold_image(images, config):
  function unmold_image (line 2914) | def unmold_image(normalized_images, config):
  function trim_zeros_graph (line 2923) | def trim_zeros_graph(boxes, name=None):
  function batch_pack_graph (line 2935) | def batch_pack_graph(x, counts, num_rows):
  function norm_boxes_graph (line 2945) | def norm_boxes_graph(boxes, shape):
  function boost_boxes_graph (line 2962) | def boost_boxes_graph(boxes):
  function denorm_boxes_graph (line 2971) | def denorm_boxes_graph(boxes, shape):

FILE: number13/number13/src/Mask_RCNN/mrcnn/parallel_model.py
  class ParallelModel (line 22) | class ParallelModel(KM.Model):
    method __init__ (line 30) | def __init__(self, keras_model, gpu_count):
    method __getattribute__ (line 41) | def __getattribute__(self, attrname):
    method summary (line 48) | def summary(self, *args, **kwargs):
    method make_parallel (line 54) | def make_parallel(self):
  function build_model (line 128) | def build_model(x_train, num_classes):

FILE: number13/number13/src/Mask_RCNN/mrcnn/utils.py
  function extract_bboxes (line 32) | def extract_bboxes(mask):
  function compute_iou (line 58) | def compute_iou(box, boxes, box_area, boxes_area):
  function compute_overlaps (line 79) | def compute_overlaps(boxes1, boxes2):
  function compute_overlaps_masks (line 98) | def compute_overlaps_masks(masks1, masks2):
  function non_max_suppression (line 116) | def non_max_suppression(boxes, scores, threshold):
  function apply_box_deltas (line 153) | def apply_box_deltas(boxes, deltas):
  function box_refinement_graph (line 177) | def box_refinement_graph(box, gt_box):
  function box_refinement (line 203) | def box_refinement(box, gt_box):
  class Dataset (line 233) | class Dataset(object):
    method __init__ (line 249) | def __init__(self, class_map=None):
    method add_class (line 256) | def add_class(self, source, class_id, class_name):
    method add_image (line 270) | def add_image(self, source, image_id, path, **kwargs):
    method image_reference (line 279) | def image_reference(self, image_id):
    method prepare (line 288) | def prepare(self, class_map=None):
    method map_source_class_id (line 324) | def map_source_class_id(self, source_class_id):
    method get_source_class_id (line 332) | def get_source_class_id(self, class_id, source):
    method append_data (line 338) | def append_data(self, class_info, image_info):
    method image_ids (line 350) | def image_ids(self):
    method source_image_link (line 353) | def source_image_link(self, image_id):
    method load_image (line 360) | def load_image(self, image_id):
    method load_mask (line 373) | def load_mask(self, image_id):
  function resize_image (line 392) | def resize_image(image, min_dim=None, max_dim=None, min_scale=None, mode...
  function resize_mask (line 500) | def resize_mask(mask, scale, padding, crop=None):
  function minimize_mask (line 522) | def minimize_mask(bbox, mask, mini_shape):
  function expand_mask (line 542) | def expand_mask(bbox, mini_mask, image_shape):
  function mold_mask (line 561) | def mold_mask(mask, config):
  function unmold_mask (line 565) | def unmold_mask(mask, bbox, image_shape):
  function generate_anchors (line 588) | def generate_anchors(scales, ratios, shape, feature_stride, anchor_stride):
  function generate_pyramid_anchors (line 627) | def generate_pyramid_anchors(scales, ratios, feature_shapes, feature_str...
  function trim_zeros (line 651) | def trim_zeros(x):
  function compute_matches (line 661) | def compute_matches(gt_boxes, gt_class_ids, gt_masks,
  function compute_ap (line 720) | def compute_ap(gt_boxes, gt_class_ids, gt_masks,
  function compute_ap_range (line 759) | def compute_ap_range(gt_box, gt_class_id, gt_mask,
  function compute_recall (line 783) | def compute_recall(pred_boxes, gt_boxes, iou):
  function batch_slice (line 808) | def batch_slice(inputs, graph_fn, batch_size, names=None):
  function download_trained_weights (line 845) | def download_trained_weights(coco_model_path, verbose=1):
  function norm_boxes (line 858) | def norm_boxes(boxes, shape):
  function denorm_boxes (line 875) | def denorm_boxes(boxes, shape):

FILE: number13/number13/src/Mask_RCNN/mrcnn/visualize.py
  function display_images (line 36) | def display_images(images, titles=None, cols=4, cmap=None, norm=None,
  function random_colors (line 60) | def random_colors(N, bright=True):
  function apply_mask (line 73) | def apply_mask(image, mask, color, alpha=0.5):
  function display_instances (line 84) | def display_instances(image, boxes, masks, class_ids, class_names,
  function display_differences (line 172) | def display_differences(image,
  function draw_rois (line 210) | def draw_rois(image, rois, refined_rois, mask, class_ids, class_names, l...
  function draw_box (line 272) | def draw_box(image, box, color):
  function display_top_masks (line 284) | def display_top_masks(image, mask, class_ids, class_names, limit=4):
  function plot_precision_recall (line 307) | def plot_precision_recall(AP, precisions, recalls):
  function plot_overlaps (line 322) | def plot_overlaps(gt_class_ids, pred_class_ids, pred_scores,
  function draw_boxes (line 361) | def draw_boxes(image, boxes=None, refined_boxes=None,
  function display_table (line 464) | def display_table(table):
  function display_weight_stats (line 478) | def display_weight_stats(model):

FILE: number13/number13/src/Mask_RCNN/setup.py
  function _parse_requirements (line 16) | def _parse_requirements(file_path):

FILE: number13/number13/src/cocoeval.py
  class COCOeval (line 49) | class COCOeval:
    method __init__ (line 99) | def __init__(self, cocoGt=None, cocoDt=None, iouType='segm'):
    method _prepare (line 124) | def _prepare(self):
    method evaluate (line 161) | def evaluate(self):
    method computeIoU (line 203) | def computeIoU(self, imgId, catId):
    method computeOks (line 232) | def computeOks(self, imgId, catId):
    method evaluateImg (line 275) | def evaluateImg(self, imgId, catId, aRng, maxDet):
    method accumulate (line 355) | def accumulate(self, p = None):
    method _summarize (line 456) | def _summarize(self, ap=1, iouThr=None, areaRng='all', maxDets=100 ):
    method summarize (line 488) | def summarize(self):
    method __str__ (line 530) | def __str__(self):
  class Params (line 533) | class Params:
    method setDetParams (line 537) | def setDetParams(self):
    method setKpParams (line 548) | def setKpParams(self):
    method __init__ (line 559) | def __init__(self, iouType='segm'):

FILE: number13/number13/src/create_patches_all.py
  function close_contour (line 22) | def close_contour(contour):
  function binary_mask_to_polygon (line 28) | def binary_mask_to_polygon(binary_mask, tolerance=0):
  function get_pan_sharpend (line 54) | def get_pan_sharpend(Dir, imageid):
  function create_coco_anns (line 63) | def create_coco_anns(file_name, counter, mask_patch):
  function patches_and_cocoann (line 85) | def patches_and_cocoann(gt, outdir_rgb, outdir_mpan, count=1,create_anns...
  function make_patch_data (line 127) | def make_patch_data(gt, outdir_rgb, outdir_mpan, count=1, create_anns=Fa...
  function main_single (line 138) | def main_single(gt_file, outdir_rgb, outdir_mpan):
  function main (line 155) | def main(angles, outdir_rgb, outdir_mpan):

FILE: number13/number13/src/crowdai_train.py
  class CrowdAiConfig (line 25) | class CrowdAiConfig(Config):
  class CrowdAiDataset (line 68) | class CrowdAiDataset(utils.Dataset):
    method load_data (line 69) | def load_data(self, coco,imgIds,dataDir ):
    method load_image (line 77) | def load_image(self, image_id):
    method image_reference (line 89) | def image_reference(self, image_id):
    method load_mask (line 98) | def load_mask(self, image_id):
  function get_dataset (line 113) | def get_dataset(Dir, fltarea=None):
  function train (line 148) | def train(init_with="coco", weights=None, fine=0, last_epoch=0, epochs=1...
  function copy_final_model (line 204) | def copy_final_model(): # Just so that spacenet finds the final required...

FILE: number13/number13/src/eval_val.py
  function fix_gt (line 24) | def fix_gt(gt,angle):
  function eval_spacenet (line 53) | def eval_spacenet(annFile, weight_path, angles, is_uint16=False, group='...

FILE: number13/number13/src/inference.py
  function get_weights_angle_specific (line 21) | def get_weights_angle_specific(use_epoch=2, group='rgb'):
  function infer_nadir_angle (line 41) | def infer_nadir_angle(input_dir, group='rgb', infer_angles='all', nms_th...
  function ensure_model_available (line 137) | def ensure_model_available():
  function infer_all (line 156) | def infer_all(input_dir):
  function save_results (line 170) | def save_results(result,save_as):

FILE: number13/number13/src/models.py
  class SpacenetConfig (line 11) | class SpacenetConfig(Config):
  class SpacenetConfigIRGB_u16 (line 50) | class SpacenetConfigIRGB_u16(SpacenetConfig):
  class SpacenetConfigRGB_u8 (line 63) | class SpacenetConfigRGB_u8(SpacenetConfig):
  class SpacenetConfigMPAN_u16 (line 76) | class SpacenetConfigMPAN_u16(SpacenetConfig):
  class SpacenetConfigMPAN_u8 (line 88) | class SpacenetConfigMPAN_u8(SpacenetConfig):

FILE: number13/number13/src/patchify.py
  class PatchGenerator (line 10) | class PatchGenerator:
    method __init__ (line 11) | def __init__(self,stepsize, imsize, winsize=None):
    method generate_coords (line 23) | def generate_coords(self):
    method create (line 37) | def create(self, img,mask=None,nonzero=False,standarize=False,coords=N...
    method reconstruct (line 78) | def reconstruct(self,patches,resize=None):
    method get_coords (line 90) | def get_coords(self):
  function tests (line 95) | def tests():

FILE: number13/number13/src/prediction.py
  function get_overlap_poly (line 24) | def get_overlap_poly():
  function get_mpan_image_patches (line 34) | def get_mpan_image_patches(ms,pan,patch_creator):
  function get_rgb_image_patches (line 46) | def get_rgb_image_patches(rgb, patch_creator):
  function fix_multipolygon (line 59) | def fix_multipolygon(mpoly):
  function fix_poly (line 72) | def fix_poly(polys):
  function polygonize_and_shift (line 84) | def polygonize_and_shift(poly,shifts):
  function close_contour (line 119) | def close_contour(contour):
  function binary_mask_to_polygon (line 125) | def binary_mask_to_polygon(binary_mask, tolerance=0):
  function get_predictions_spacenet (line 151) | def get_predictions_spacenet(image_id, img_patches, model, shifts, conf_...
  function get_final_annotations (line 185) | def get_final_annotations(image_id,anns,mx,my):
  function predict_mpan512 (line 235) | def predict_mpan512(subdir, model,conf_thres=0.95):
  function predict_rgb512 (line 276) | def predict_rgb512(subdir, model,conf_thres=0.95):
  function get_predictions_coco (line 317) | def get_predictions_coco(dataset, model, subname='tmp.json',conf_thres=0...

FILE: number13/number13/src/train.py
  class SpacenetDataset (line 38) | class SpacenetDataset(utils.Dataset):
    method load_data (line 39) | def load_data(self, coco,imgIds, dataDir, group='rgb',is_uint16 = False):
    method load_image (line 51) | def load_image(self, image_id):
    method image_reference (line 65) | def image_reference(self, image_id):
    method load_mask (line 74) | def load_mask(self, image_id):
  function get_train_val_split (line 93) | def get_train_val_split(dir, angles, is_uint16=False):
  function get_datasets (line 124) | def get_datasets(annFile, angles, is_uint16=False, group = 'rgb'):
  function limit_mem (line 143) | def limit_mem():
  function train (line 150) | def train(annFile, angle, previous_weights, group='rgb',is_uint16=False,...
  function train_cascade (line 218) | def train_cascade(annFile, initial_weight, angles, train_type, group='rg...
  function train_extra (line 243) | def train_extra(annfile,group,model_dir):
  function train_all (line 270) | def train_all(pretrained_weights,gpus=1):

FILE: number13/number13/src/util.py
  function mask_for_polygons (line 13) | def mask_for_polygons(polygons, im_size=(900, 900)):
  function contours_hierarchy (line 36) | def contours_hierarchy(mask):
  function mask_to_polygons (line 45) | def mask_to_polygons(mask, epsilon=0.0, min_area=0):
  function mask_to_multipolygons (line 78) | def mask_to_multipolygons(mask, epsilon=0.0, min_area=0, shift=(0, 0)):
  function stretch_8bit (line 124) | def stretch_8bit(bands, lower_percent=2, higher_percent=98, chan=3):
  function pansharpen (line 138) | def pansharpen(m, pan):
  function strip_tail (line 174) | def strip_tail(annotations):
  function comput_mean_jpg (line 178) | def comput_mean_jpg(dir):

FILE: selim_sef/create_folds.py
  function get_id (line 10) | def get_id(f):
  function get_nadir (line 13) | def get_nadir(f):

FILE: selim_sef/dataset/dense_data.py
  function stretch_8bit (line 73) | def stretch_8bit(bands, lower_percent=0, higher_percent=100):
  class DenseData (line 91) | class DenseData(Dataset):
    method __init__ (line 92) | def __init__(self, data_path, nadir, mode="train", csv_path="folds.csv...
    method __len__ (line 111) | def __len__(self):
    method __getitem__ (line 114) | def __getitem__(self, idx):
  class TestDenseData (line 140) | class TestDenseData(Dataset):
    method __init__ (line 141) | def __init__(self, data_path, transform=None):
    method __len__ (line 149) | def __len__(self):
    method __getitem__ (line 152) | def __getitem__(self, idx):

FILE: selim_sef/dataset/dense_transform.py
  class Normalize (line 18) | class Normalize(object):
    method __init__ (line 19) | def __init__(self, mean, std):
    method __call__ (line 23) | def __call__(self, sample):
    method normalize (line 29) | def normalize(self, tensor, mean, std):
  class HFlip (line 37) | class HFlip(object):
    method __call__ (line 38) | def __call__(self, sample):
  class VFlip (line 51) | class VFlip(object):
    method __call__ (line 52) | def __call__(self, sample):
  function rot90 (line 60) | def rot90(img, factor):
  class Rotate90 (line 65) | class Rotate90(object):
    method __call__ (line 66) | def __call__(self, sample):
  class Pad (line 74) | class Pad(object):
    method __init__ (line 75) | def __init__(self, block=32, mode='reflect'):
    method __call__ (line 80) | def __call__(self, sample):
  function pad (line 87) | def pad(image, block, type='reflect', **kwargs):
  class ToTensor (line 104) | class ToTensor(object):
    method __call__ (line 105) | def __call__(self, sample):
  class ColorJitterImage (line 113) | class ColorJitterImage(object):
    method __init__ (line 114) | def __init__(self):
    method __call__ (line 117) | def __call__(self, sample):
  class LightingImage (line 123) | class LightingImage(object):
    method __init__ (line 124) | def __init__(self):
    method __call__ (line 127) | def __call__(self, sample):
  class RandomCropAndScale (line 132) | class RandomCropAndScale(object):
    method __init__ (line 133) | def __init__(self, height, width, scale_range=(0.5, 2.0), rescale_prob...
    method __call__ (line 140) | def __call__(self, sample):
  function random_crop (line 154) | def random_crop(img, height, width, scale, random_state, mode=None):
  function shift_scale_rotate (line 183) | def shift_scale_rotate(img, angle, scale, dx, dy, borderMode=cv2.BORDER_...
  class RandomRotate (line 204) | class RandomRotate(object):
    method __init__ (line 205) | def __init__(self, angle=15, prob=0.3):
    method __call__ (line 210) | def __call__(self, sample):
  function _grayscale (line 222) | def _grayscale(img):
  function _blend (line 227) | def _blend(img1, img2, alpha):
  class Lighting (line 231) | class Lighting(object):
    method __init__ (line 232) | def __init__(self, alphastd=_DEFAULT_ALPHASTD, eigval=_DEFAULT_EIGVAL,...
    method __call__ (line 237) | def __call__(self, img):
  class Saturation (line 246) | class Saturation(object):
    method __init__ (line 247) | def __init__(self, var):
    method __call__ (line 250) | def __call__(self, img):
  class Brightness (line 256) | class Brightness(object):
    method __init__ (line 257) | def __init__(self, var):
    method __call__ (line 260) | def __call__(self, img):
  class Contrast (line 266) | class Contrast(object):
    method __init__ (line 267) | def __init__(self, var):
    method __call__ (line 270) | def __call__(self, img):
  class ColorJitter (line 277) | class ColorJitter(object):
    method __init__ (line 278) | def __init__(self, saturation=_DEFAULT_BCS[0], brightness=_DEFAULT_BCS...
    method __call__ (line 287) | def __call__(self, img):

FILE: selim_sef/ensemble.py
  function average_strategy (line 15) | def average_strategy(images):
  function hard_voting (line 19) | def hard_voting(images):
  function ensemble_image (line 24) | def ensemble_image(params):
  function ensemble (line 41) | def ensemble(dirs, strategy, ensembling_dir, n_threads):

FILE: selim_sef/evaluate_labels.py
  function calc (line 19) | def calc(f):

FILE: selim_sef/generate_polygons.py
  function label_mask (line 27) | def label_mask(pred, main_threshold=0.3, seed_threshold=0.7, w_pixel_t=2...
  function _internal_test (line 55) | def _internal_test(mask_dir, out_file):
  function mask_to_poly (line 85) | def mask_to_poly(mask, min_polygon_area_th=MIN_AREA):
  function _remove_interiors (line 114) | def _remove_interiors(line):

FILE: selim_sef/inference/predict.py
  function predict_tta (line 20) | def predict_tta(model, batch, apply_sigmoid, transforms):
  function get_nadir (line 28) | def get_nadir(f):

FILE: selim_sef/inference/predict_oof.py
  function predict_tta (line 21) | def predict_tta(model, batch, apply_sigmoid, transforms):

FILE: selim_sef/inference/tta.py
  class TTAOp (line 7) | class TTAOp:
    method __init__ (line 8) | def __init__(self, sigmoid=True):
    method __call__ (line 11) | def __call__(self, model, batch):
    method forward (line 17) | def forward(self, img):
    method backward (line 20) | def backward(self, img):
    method to_numpy (line 23) | def to_numpy(self, batch):
  class BasicTTAOp (line 32) | class BasicTTAOp(TTAOp):
    method op (line 34) | def op(img):
    method forward (line 37) | def forward(self, img):
    method backward (line 40) | def backward(self, img):
  class Nothing (line 49) | class Nothing(BasicTTAOp):
    method op (line 51) | def op(img):
  class ScaleUp2 (line 55) | class ScaleUp2(BasicTTAOp):
    method op (line 57) | def op(img):
    method backward (line 67) | def backward(self, img):
  class ScaleUp1 (line 77) | class ScaleUp1(BasicTTAOp):
    method op (line 79) | def op(img):
    method backward (line 89) | def backward(self, img):
  class ScaleDown (line 99) | class ScaleDown(BasicTTAOp):
    method op (line 101) | def op(img):
    method backward (line 111) | def backward(self, img):
  class HFlip (line 122) | class HFlip(BasicTTAOp):
    method op (line 124) | def op(img):
  class VFlip (line 128) | class VFlip(BasicTTAOp):
    method op (line 130) | def op(img):
  class Transpose (line 134) | class Transpose(BasicTTAOp):
    method op (line 136) | def op(img):
  function chain_op (line 140) | def chain_op(data, operations):
  class ChainedTTA (line 146) | class ChainedTTA(TTAOp):
    method operations (line 148) | def operations(self):
    method forward (line 151) | def forward(self, img):
    method backward (line 154) | def backward(self, img):
  class HVFlip (line 163) | class HVFlip(ChainedTTA):
    method operations (line 165) | def operations(self):
  class TransposeHFlip (line 169) | class TransposeHFlip(ChainedTTA):
    method operations (line 171) | def operations(self):
  class TransposeVFlip (line 175) | class TransposeVFlip(ChainedTTA):
    method operations (line 177) | def operations(self):
  class TransposeHVFlip (line 181) | class TransposeHVFlip(ChainedTTA):
    method operations (line 183) | def operations(self):

FILE: selim_sef/predict_trees.py
  function process_images (line 40) | def process_images(step):

FILE: selim_sef/tools/adamw.py
  class AdamW (line 6) | class AdamW(torch.optim.Optimizer):
    method __init__ (line 25) | def __init__(self, params, lr=1e-3, betas=(0.9, 0.999), eps=1e-8,
    method step (line 31) | def step(self, closure=None):

FILE: selim_sef/tools/clr.py
  class CyclicLR (line 4) | class CyclicLR(object):
    method __init__ (line 68) | def __init__(self, optimizer, base_lr=5e-5, max_lr=2e-4,
    method step (line 119) | def step(self, batch_iteration=None):
    method _triangular_scale_fn (line 126) | def _triangular_scale_fn(self, x):
    method _triangular2_scale_fn (line 129) | def _triangular2_scale_fn(self, x):
    method _exp_range_scale_fn (line 132) | def _exp_range_scale_fn(self, x):
    method get_lr (line 135) | def get_lr(self):

FILE: selim_sef/tools/config.py
  function _merge (line 48) | def _merge(src, dst):
  function load_config (line 57) | def load_config(config_file, defaults=DEFAULTS):

FILE: selim_sef/tools/lr_policy.py
  class PolyLR (line 4) | class PolyLR(_LRScheduler):
    method __init__ (line 7) | def __init__(self, optimizer, max_iter=90000, power=0.9, last_epoch=-1):
    method get_lr (line 12) | def get_lr(self):

FILE: selim_sef/tools/mask_from_geo.py
  function masks_from_geojsons (line 6) | def masks_from_geojsons(geojson_dir, im_src_dir, mask_dest_dir,
  function main (line 47) | def main(train_dir):

FILE: selim_sef/tools/mask_utils.py
  function create_separation (line 28) | def create_separation(labels):
  function create_mask (line 57) | def create_mask(img_id, data_dir):
  function save_mask_and_label (line 90) | def save_mask_and_label(image_name):
  function main (line 100) | def main():

FILE: selim_sef/tools/rle.py
  function multi_rle_encode (line 4) | def multi_rle_encode(labels):
  function rle_encode (line 8) | def rle_encode(img):
  function rle_decode (line 20) | def rle_decode(mask_rle, shape=(768, 768)):
  function masks_as_image (line 36) | def masks_as_image(in_mask_list, all_masks=None, shape=(768, 768)):
  function masks_as_label (line 45) | def masks_as_label(in_mask_list, all_masks=None, shape=(768, 768)):

FILE: selim_sef/train.py
  function get_model_name (line 28) | def get_model_name(model, num_classes, snapshot_prefix, dataset_name):
  function main (line 57) | def main():
  function evaluate_val (line 137) | def evaluate_val(args, data_val, miou_best, model, snapshot_name, conf, ...
  function train_epoch (line 164) | def train_epoch(args, conf, current_epoch, loss_function, model, optimiz...

FILE: selim_sef/train_classifier.py
  function get_nadir (line 44) | def get_nadir(f):
  function get_inputs (line 48) | def get_inputs(filename, pred_folder, truth_folder=None):
  function check_id (line 366) | def check_id(id: str):
  function get_id (line 374) | def get_id(f):

FILE: selim_sef/training/eval.py
  function postprocess (line 15) | def postprocess(mask, min_size=10):
  function validate (line 25) | def validate(net, data_loader):

FILE: selim_sef/training/losses.py
  function dice_round (line 15) | def dice_round(preds, trues):
  function soft_dice_loss (line 20) | def soft_dice_loss(outputs, targets, per_image=False):
  function jaccard (line 34) | def jaccard(outputs, targets, per_image=False, non_empty=False, min_pixe...
  class DiceLoss (line 60) | class DiceLoss(nn.Module):
    method __init__ (line 61) | def __init__(self, weight=None, size_average=True, per_image=False):
    method forward (line 67) | def forward(self, input, target):
  class JaccardLoss (line 71) | class JaccardLoss(nn.Module):
    method __init__ (line 72) | def __init__(self, weight=None, size_average=True, per_image=False, no...
    method forward (line 82) | def forward(self, input, target):
  class StableBCELoss (line 88) | class StableBCELoss(nn.Module):
    method __init__ (line 89) | def __init__(self):
    method forward (line 92) | def forward(self, input, target):
  class ComboLoss (line 101) | class ComboLoss(nn.Module):
    method __init__ (line 102) | def __init__(self, weights, per_image=False, channel_weights=[1, 0.5, ...
    method forward (line 123) | def forward(self, outputs, targets):
  function lovasz_grad (line 146) | def lovasz_grad(gt_sorted):
  function lovasz_hinge (line 161) | def lovasz_hinge(logits, labels, per_image=True, ignore=None):
  function lovasz_hinge_flat (line 177) | def lovasz_hinge_flat(logits, labels):
  function flatten_binary_scores (line 197) | def flatten_binary_scores(scores, labels, ignore=None):
  function lovasz_sigmoid (line 212) | def lovasz_sigmoid(probas, labels, per_image=False, ignore=None):
  function lovasz_sigmoid_flat (line 229) | def lovasz_sigmoid_flat(probas, labels):
  function symmetric_lovasz (line 244) | def symmetric_lovasz(outputs, targets, ):
  function mean (line 247) | def mean(l, ignore_nan=False, empty=0):
  class LovaszLoss (line 268) | class LovaszLoss(nn.Module):
    method __init__ (line 269) | def __init__(self, ignore_index=255, per_image=True):
    method forward (line 274) | def forward(self, outputs, targets):
  class LovaszLossSigmoid (line 279) | class LovaszLossSigmoid(nn.Module):
    method __init__ (line 280) | def __init__(self, ignore_index=255, per_image=True):
    method forward (line 285) | def forward(self, outputs, targets):
  class FocalLoss2d (line 291) | class FocalLoss2d(nn.Module):
    method __init__ (line 292) | def __init__(self, gamma=2, ignore_index=255):
    method forward (line 297) | def forward(self, outputs, targets):

FILE: selim_sef/training/meters.py
  class AverageMeter (line 2) | class AverageMeter(object):
    method __init__ (line 5) | def __init__(self):
    method reset (line 8) | def reset(self):
    method update (line 14) | def update(self, val, n=1):

FILE: selim_sef/training/metric.py
  function miou_score (line 4) | def miou_score(y, p):
  function calc_score (line 7) | def calc_score(labels, y_pred):
  function precision_at (line 38) | def precision_at(threshold, iou):
  function dice (line 46) | def dice(im1, im2, empty_score=1.0):

FILE: selim_sef/training/utils.py
  function get_model_params (line 12) | def get_model_params(network_config):
  function create_optimizer (line 31) | def create_optimizer(optimizer_config, model, master_params=None):
  function create_transforms (line 111) | def create_transforms(input_config):

FILE: selim_sef/zoo/densenet.py
  function densenet121 (line 19) | def densenet121(pretrained=True, **kwargs):
  function densenet169 (line 49) | def densenet169(pretrained=True, **kwargs):
  function densenet201 (line 79) | def densenet201(pretrained=True, **kwargs):
  function densenet161 (line 106) | def densenet161(pretrained=True, **kwargs):
  class _DenseLayer (line 137) | class _DenseLayer(nn.Sequential):
    method __init__ (line 138) | def __init__(self, num_input_features, growth_rate, bn_size, drop_rate):
    method forward (line 150) | def forward(self, x):
  class _DenseBlock (line 157) | class _DenseBlock(nn.Sequential):
    method __init__ (line 158) | def __init__(self, num_layers, num_input_features, bn_size, growth_rat...
  class _Transition (line 165) | class _Transition(nn.Sequential):
    method __init__ (line 166) | def __init__(self, num_input_features, num_output_features):
  class DenseNet (line 175) | class DenseNet(nn.Module):
    method __init__ (line 188) | def __init__(self, growth_rate=32, block_config=(6, 12, 24, 16),
    method forward (line 229) | def forward(self, x):

FILE: selim_sef/zoo/dpn.py
  function dpn68 (line 97) | def dpn68(num_classes=1000, pretrained='imagenet'):
  function dpn68b (line 115) | def dpn68b(num_classes=1000, pretrained='imagenet+5k'):
  function dpn92 (line 133) | def dpn92(num_classes=1000, pretrained='imagenet+5k'):
  function dpn98 (line 151) | def dpn98(num_classes=1000, pretrained='imagenet'):
  function dpn131 (line 169) | def dpn131(num_classes=1000, pretrained='imagenet'):
  function dpn107 (line 187) | def dpn107(num_classes=1000, pretrained='imagenet+5k'):
  class CatBnAct (line 206) | class CatBnAct(nn.Module):
    method __init__ (line 207) | def __init__(self, in_chs, activation_fn=nn.ReLU(inplace=True)):
    method forward (line 212) | def forward(self, x):
  class BnActConv2d (line 217) | class BnActConv2d(nn.Module):
    method __init__ (line 218) | def __init__(self, in_chs, out_chs, kernel_size, stride,
    method forward (line 225) | def forward(self, x):
  class InputBlock (line 229) | class InputBlock(nn.Module):
    method __init__ (line 230) | def __init__(self, num_init_features, kernel_size=7,
    method forward (line 239) | def forward(self, x):
  class DualPathBlock (line 247) | class DualPathBlock(nn.Module):
    method __init__ (line 248) | def __init__(
    method forward (line 284) | def forward(self, x):
  class DPN (line 311) | class DPN(nn.Module):
    method __init__ (line 312) | def __init__(self, small=False, num_init_features=64, k_r=96, groups=32,
    method logits (line 381) | def logits(self, features):
    method forward (line 392) | def forward(self, input):
  function pooling_factor (line 409) | def pooling_factor(pool_type='avg'):
  function adaptive_avgmax_pool2d (line 413) | def adaptive_avgmax_pool2d(x, pool_type='avg', padding=0, count_include_...
  class AdaptiveAvgMaxPool2d (line 437) | class AdaptiveAvgMaxPool2d(torch.nn.Module):
    method __init__ (line 440) | def __init__(self, output_size=1, pool_type='avg'):
    method forward (line 453) | def forward(self, x):
    method factor (line 462) | def factor(self):
    method __repr__ (line 465) | def __repr__(self):

FILE: selim_sef/zoo/resnet.py
  function conv3x3 (line 19) | def conv3x3(in_planes, out_planes, stride=1):
  class BasicBlock (line 25) | class BasicBlock(nn.Module):
    method __init__ (line 28) | def __init__(self, inplanes, planes, stride=1, downsample=None):
    method forward (line 38) | def forward(self, x):
  class Bottleneck (line 57) | class Bottleneck(nn.Module):
    method __init__ (line 60) | def __init__(self, inplanes, planes, stride=1, downsample=None):
    method forward (line 73) | def forward(self, x):
  class ResNet (line 96) | class ResNet(nn.Module):
    method __init__ (line 97) | def __init__(self, block, layers, in_channels=3):
    method _make_layer (line 118) | def _make_layer(self, block, planes, blocks, stride=1):
    method forward (line 135) | def forward(self, x):
  function resnet18 (line 149) | def resnet18(**kwargs):
  function resnet34 (line 159) | def resnet34(**kwargs):
  function resnet50 (line 169) | def resnet50(**kwargs):
  function resnet101 (line 179) | def resnet101(**kwargs):
  function resnet152 (line 189) | def resnet152(**kwargs):

FILE: selim_sef/zoo/senet.py
  class SEModule (line 89) | class SEModule(nn.Module):
    method __init__ (line 91) | def __init__(self, channels, reduction, concat=False):
    method forward (line 101) | def forward(self, x):
  class SCSEModule (line 110) | class SCSEModule(nn.Module):
    method __init__ (line 112) | def __init__(self, channels, reduction=16, mode='concat'):
    method forward (line 127) | def forward(self, x):
  class Bottleneck (line 146) | class Bottleneck(nn.Module):
    method forward (line 150) | def forward(self, x):
  class SEBottleneck (line 173) | class SEBottleneck(Bottleneck):
    method __init__ (line 179) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SCSEBottleneck (line 197) | class SCSEBottleneck(Bottleneck):
    method __init__ (line 203) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SEResNetBottleneck (line 221) | class SEResNetBottleneck(Bottleneck):
    method __init__ (line 229) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SEResNeXtBottleneck (line 246) | class SEResNeXtBottleneck(Bottleneck):
    method __init__ (line 252) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SCSEResNeXtBottleneck (line 271) | class SCSEResNeXtBottleneck(Bottleneck):
    method __init__ (line 277) | def __init__(self, inplanes, planes, groups, reduction, stride=1,
  class SENet (line 295) | class SENet(nn.Module):
    method __init__ (line 297) | def __init__(self, block, layers, groups, reduction, dropout_p=0.2,
    method _make_layer (line 415) | def _make_layer(self, block, planes, blocks, groups, reduction, stride=1,
    method _initialize_weights (line 435) | def _initialize_weights(self):
    method features (line 445) | def features(self, x):
    method logits (line 454) | def logits(self, x):
    method forward (line 462) | def forward(self, x):
  function initialize_pretrained_model (line 468) | def initialize_pretrained_model(model, num_classes, settings):
  function senet154 (line 480) | def senet154(num_classes=1000, pretrained='imagenet'):
  function scsenet154 (line 488) | def scsenet154(num_classes=1000, pretrained='imagenet'):
  function se_resnet50 (line 498) | def se_resnet50(num_classes=1000, pretrained='imagenet'):
  function se_resnet101 (line 509) | def se_resnet101(num_classes=1000, pretrained='imagenet'):
  function se_resnet152 (line 520) | def se_resnet152(num_classes=1000, pretrained='imagenet'):
  function se_resnext50_32x4d (line 531) | def se_resnext50_32x4d(num_classes=1000, pretrained='imagenet'):
  function scse_resnext50_32x4d (line 542) | def scse_resnext50_32x4d(num_classes=1000, pretrained='imagenet'):
  function se_resnext101_32x4d (line 553) | def se_resnext101_32x4d(num_classes=1000, pretrained='imagenet'):

FILE: selim_sef/zoo/unet.py
  class BasicConvAct (line 81) | class BasicConvAct(nn.Module):
    method __init__ (line 82) | def __init__(self, in_channels, out_channels, kernel_size=1, dilation=...
    method forward (line 90) | def forward(self, x):
  class Conv1x1 (line 96) | class Conv1x1(BasicConvAct):
    method __init__ (line 97) | def __init__(self, in_channels, out_channels, dilation=1, bias=True):
  class Conv3x3 (line 101) | class Conv3x3(BasicConvAct):
    method __init__ (line 102) | def __init__(self, in_channels, out_channels, dilation=1):
  class ConvReLu1x1 (line 106) | class ConvReLu1x1(BasicConvAct):
    method __init__ (line 107) | def __init__(self, in_channels, out_channels, dilation=1):
  class ConvReLu3x3 (line 111) | class ConvReLu3x3(BasicConvAct):
    method __init__ (line 112) | def __init__(self, in_channels, out_channels, dilation=1):
  class BasicUpBlock (line 116) | class BasicUpBlock(nn.Module):
    method __init__ (line 117) | def __init__(self, in_channels, out_channels, kernel_size=3, activatio...
    method forward (line 126) | def forward(self, x):
  class AbstractModel (line 132) | class AbstractModel(nn.Module):
    method _initialize_weights (line 133) | def _initialize_weights(self):
    method initialize_encoder (line 143) | def initialize_encoder(self, model, model_url, num_channels_changed=Fa...
    method first_layer_params_name (line 161) | def first_layer_params_name(self):
  class EncoderDecoder (line 165) | class EncoderDecoder(AbstractModel):
    method __init__ (line 166) | def __init__(self, num_classes, num_channels=3, encoder_name='resnet34'):
    method forward (line 203) | def forward(self, x):
    method get_decoder (line 225) | def get_decoder(self, layer):
    method make_final_classifier (line 230) | def make_final_classifier(self, in_filters, num_classes):
    method get_encoder (line 235) | def get_encoder(self, encoder, layer):
    method first_layer_params (line 239) | def first_layer_params(self):
    method layers_except_first_params (line 243) | def layers_except_first_params(self):
  function _get_layers_params (line 248) | def _get_layers_params(layers):
  function get_slice (line 252) | def get_slice(features, start, end):
  class ConvBottleneck (line 258) | class ConvBottleneck(nn.Module):
    method __init__ (line 259) | def __init__(self, in_channels, out_channels):
    method forward (line 266) | def forward(self, dec, enc):
  class UnetDecoderBlock (line 271) | class UnetDecoderBlock(nn.Module):
    method __init__ (line 272) | def __init__(self, in_channels, middle_channels, out_channels):
    method forward (line 280) | def forward(self, x):
  class Resnet (line 284) | class Resnet(EncoderDecoder):
    method __init__ (line 285) | def __init__(self, seg_classes, backbone_arch):
    method get_encoder (line 289) | def get_encoder(self, encoder, layer):
  class DPNUnet (line 307) | class DPNUnet(EncoderDecoder):
    method __init__ (line 308) | def __init__(self, seg_classes, backbone_arch='dpn92'):
    method get_encoder (line 312) | def get_encoder(self, encoder, layer):
    method first_layer_params_name (line 332) | def first_layer_params_name(self):
  class DensenetUnet (line 336) | class DensenetUnet(EncoderDecoder):
    method __init__ (line 337) | def __init__(self, seg_classes, backbone_arch='densenet121'):
    method get_encoder (line 342) | def get_encoder(self, encoder, layer):
  class SEUnet (line 358) | class SEUnet(EncoderDecoder):
    method __init__ (line 359) | def __init__(self, seg_classes, backbone_arch='senet154'):
    method get_encoder (line 364) | def get_encoder(self, encoder, layer):
    method first_layer_params_name (line 379) | def first_layer_params_name(self):
  class ConvSCSEBottleneckNoBn (line 384) | class ConvSCSEBottleneckNoBn(nn.Module):
    method __init__ (line 385) | def __init__(self, in_channels, out_channels, reduction=2):
    method forward (line 394) | def forward(self, dec, enc):
  class SCSEUnet (line 398) | class SCSEUnet(SEUnet):
    method __init__ (line 399) | def __init__(self, seg_classes, backbone_arch='seresnext50'):