Functions — github.com/bcmi/FOPA-Fast-Object-Placement-Assessment

hub / github.com/bcmi/FOPA-Fast-Object-Placement-Assessment / functions

Functions61 in github.com/bcmi/FOPA-Fast-Object-Placement-Assessment

⨍ Functions 61 ◇ Types & classes 13

↓ 4 callersMethod_make_layer

(self, block, planes, blocks, stride=1)

backbone/ResNet.py:128

↓ 4 callersFunctioncreate_loader

(table_path, bg_dir, fg_dir, mask_dir, in_size, datatype, batch_size, num_workers, shuffle)

data/OBdataset.py:199

↓ 3 callersFunction_collect_info

load json file and return required information Args: json_file(str): json file with train/test information bg_dir(str): fold

data/OBdataset.py:158

↓ 3 callersMethodchange_lr

(self, curr)

train.py:162

↓ 3 callersFunctionconv1x1

1x1 convolution

backbone/ResNet.py:23

↓ 3 callersFunctionconv3x3

3x3 convolution with padding

backbone/ResNet.py:18

↓ 3 callersFunctionmake_log

(path, context)

utils/misc.py:33

↓ 3 callersMethodupdate

(self, val, n=1)

utils/misc.py:14

↓ 2 callersMethod__init__

(self, block, layers, zero_init_residual=False)

backbone/ResNet.py:99

↓ 2 callersFunctioncheck_mkdir

(dir_name)

utils/misc.py:28

↓ 2 callersFunctionmake_composite

(fg_img, mask_img, bg_img, pos, isflip=False)

data/OBdataset.py:206

↓ 2 callersFunctionmake_composite_PIL

(fg_img, mask_img, bg_img, pos, return_mask=False)

data/OBdataset.py:262

↓ 2 callersFunctionpretrained_resnet18_4ch

(pretrained=True, **kwargs)

backbone/ResNet.py:214

↓ 2 callersMethodresume_checkpoint

Args: load_path (str): path of pretrained model mode (str): 'all'：resume all information；'onlynet':only resume model

train.py:233

↓ 1 callersFunctionBackbone_ResNet18_in3

(pretrained=True)

backbone/ResNet.py:176

↓ 1 callersFunctionBackbone_ResNet18_in3_1

(pretrained=True)

backbone/ResNet.py:189

↓ 1 callersFunction_obtain_target

put 0, 1 labels on a 256x256 score map Args: original_width(int): width of original background original_height(int): height o

data/OBdataset.py:122

↓ 1 callersFunctioncollect_info

(json_file, bg_dir, fg_dir)

prepare_multi_fg_scales.py:14

↓ 1 callersFunctionconstruct_path_dict

(proj_root, exp_name)

utils/misc.py:45

↓ 1 callersMethodevalutate_model

calculate F1 and bAcc metrics

test.py:32

↓ 1 callersFunctionfeature_mimicking

(composites, feature_pos, feature_map, num, device)

train.py:256

↓ 1 callersMethodgenerate_composite

generate composite images for each pair of scaled foreground and background

test.py:118

↓ 1 callersMethodgenerate_composite_multi_scales

generate composite images for each pair of scaled foreground and background

test_multi_fg_scales.py:71

↓ 1 callersMethodget_heatmap

generate heatmap for each pair of scaled foreground and background

test.py:82

↓ 1 callersMethodget_heatmap_multi_scales

generate heatmap for each pair of scaled foreground and background

test_multi_fg_scales.py:32

↓ 1 callersMethodmake_optim

(self)

train.py:180

↓ 1 callersFunctionpre_mkdir

(path_config)

utils/misc.py:21

↓ 1 callersMethodreset

(self)

utils/misc.py:8

↓ 1 callersFunctionresnet18

Constructs a ResNet-18 model. Args: pretrained (bool): If True, returns a model pre-trained on ImageNet

backbone/ResNet.py:156

↓ 1 callersMethodsave_checkpoint

(self, current_epoch, full_net_path, state_net_path)

train.py:224

↓ 1 callersFunctionsetup_seed

(seed)

train.py:34

↓ 1 callersMethodtrain

data/all_transforms.py:12

Method__call__

(self, bg, fg, mask)

data/all_transforms.py:22

Method__getitem__

load each item return: i: the image index, bg_t:(1 * 3 * in_size * in_size) background image, mas

(self, args, checkpoint_path)

test.py:17

Method__init__

(self, args, checkpoint_path)

test_multi_fg_scales.py:17

Method__init__

(self, inplanes, planes, stride=1, downsample=None)

backbone/ResNet.py:31

Method__init__

(self, inplanes, planes, stride=1, downsample=None)

backbone/ResNet.py:63

( self, in_planes, out_planes, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=False,

network/BaseBlocks.py:5

Method__init__

(self, pretrained=True, ks=3, scale=3)

network/ObPlaNet_simple.py:13

Method__init__

use nn.Unfold to realize dynamic convolution Args: in_xC (int): channel number of first input in_yC (int): channel nu

network/DynamicModules.py:6

Method__init__

(self, size)

data/all_transforms.py:4

Method__init__

(self, transforms)

data/all_transforms.py:19

Method__init__

initialize dataset Args: file(str): file with training/test data information bg_dir(str): folder with backgr

conver bbox to center pixel

data/OBdataset.py:193

Functioncheck_dir_path_valid

network/tensor_ops.py:4

Methodforward

(self, x)

backbone/ResNet.py:41

Methodforward

(self, x)

backbone/ResNet.py:75

Methodforward

(self, x)

backbone/ResNet.py:143

Methodforward

(self, x)

network/BaseBlocks.py:25

Methodforward

Args: bg_in_data: (batch_size * 3 * H * W) background image fg_in_data: (batch_size * 3 * H * W) scaled foreground im

network/ObPlaNet_simple.py:61

Methodforward

(self, x, y)

network/DynamicModules.py:29

Functionshuffle_channels

(x, groups)

network/tensor_ops.py:34

Functionupsample_add

(*xs)

network/tensor_ops.py:9

Functionupsample_cat

(*xs)

network/tensor_ops.py:16

Functionupsample_reduce

(b, a)

network/tensor_ops.py:24