Functions — github.com/Junjue-Wang/LoveNAS

hub / github.com/Junjue-Wang/LoveNAS / functions

Functions277 in github.com/Junjue-Wang/LoveNAS

⨍ Functions 277◇ Types & classes 68

↓ 10 callersFunction_resnet

(arch, block, layers, pretrained, progress, **kwargs)

module/baseline/base_resnet/_resnets.py:230

↓ 8 callersMethod__init__

(self, config)

module/baseline/unet.py:317

↓ 5 callersMethod__init__

(self, size=None, scale_factor=None)

module/tta.py:119

↓ 5 callersMethodapply

(self, img, **params)

data/transforms.py:23

↓ 5 callersFunctionload_url_dist

In distributed setting, this function only download checkpoint at local rank 0.

module/baseline/swin/checkpoint.py:109

↓ 4 callersMethod__init__

Arguments: in_channels_list (list[int]): number of channels for each feature map that will be fed out

module/baseline/base.py:98

↓ 4 callersMethod__init__

(self, embed_dims, num_heads, window_size,

module/baseline/swin/swin_transformer.py:40

↓ 4 callersMethod_make_layer

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

module/baseline/base_resnet/_resnets.py:181

↓ 4 callersMethodforward

(self, x, y=None)

module/baseline/unet.py:326

↓ 4 callersMethodget_function

(module)

module/baseline/base_resnet/resnet.py:133

↓ 3 callersMethod__init__

(self, in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1,

module/nas/operations.py:19

↓ 3 callersMethod__init__

(self, channels, m)

module/baseline/lrn.py:22

↓ 3 callersMethod__init__

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

module/baseline/base_hrnet/_hrnet.py:189

↓ 3 callersMethod_make_stage

(self, layer_config, num_inchannels, multi_scale_output=True)

module/baseline/base_hrnet/_hrnet.py:534

↓ 3 callersMethod_make_transition_layer

( self, num_channels_pre_layer, num_channels_cur_layer)

module/baseline/base_hrnet/_hrnet.py:471

↓ 3 callersFunctionconv1x1

1x1 convolution

module/baseline/base_resnet/_resnets.py:27

↓ 3 callersFunctionconv3x3

3x3 convolution with padding

module/baseline/base_resnet/_resnets.py:21

↓ 3 callersMethodlayer1

(self)

module/baseline/base_resnet/resnet.py:77

↓ 3 callersFunctionseed_torch

(seed=42)

train_loveda.py:56

↓ 3 callersFunctiontta

(model, image, tta_config)

module/tta.py:12

↓ 2 callersMethod__init__

(self, config)

module/baseline/pspnet.py:46

↓ 2 callersMethod__init__

(self, block, layers, num_classes=1000, zero_init_residual=False, groups=1, width_per_group=6

module/baseline/base_resnet/_resnets.py:117

↓ 2 callersMethod__init__

(self, in_channels=3, embed_dims=768, conv_type='Conv2d',

module/baseline/swin/embed.py:108

↓ 2 callersMethod_freeze_at

(self, at=2)

module/baseline/base_resnet/resnet.py:119

↓ 2 callersFunction_get_mmcv_home

()

module/baseline/swin/checkpoint.py:30

↓ 2 callersFunction_masked_ignore

(y_pred: torch.Tensor, y_true: torch.Tensor, ignore_index: int)

module/loss.py:5

↓ 2 callersMethodbatch_generate

(self, image_dir, mask_dir)

data/loveda.py:65

↓ 2 callersMethodbatch_generate

(self, image_dir, mask_dir)

data/floodnet.py:49

↓ 2 callersFunctionbinary_cross_entropy_with_logits

(output: torch.Tensor, target: torch.Tensor, reduction: str = 'mean', ign

module/loss.py:29

↓ 2 callersFunctionclip_floodnet

(img_dir, save_dir)

tools/prepare_floodnet.py:45

↓ 2 callersFunctionconv3x3

3x3 convolution with padding

module/baseline/base_hrnet/_hrnet.py:180

↓ 2 callersFunctionfast_normalize

(x)

tools/parse_arch.py:15

↓ 2 callersMethodforward

(self, model, image_tensor, **kwargs)

eval_window.py:86

↓ 2 callersMethodgenerate_ops

(self, out_s, init=False)

module/nas/nasdecoder.py:39

↓ 2 callersMethodget_pad_shape

(self, input_shape)

module/baseline/swin/embed.py:57

↓ 2 callersFunctionget_torchvision_models

()

module/baseline/swin/checkpoint.py:172

↓ 2 callersMethodinv_transform

(self, transformed_inputs)

module/tta.py:131

↓ 2 callersMethodlayer2

(self)

module/baseline/base_resnet/resnet.py:86

↓ 2 callersMethodlayer3

(self)

module/baseline/base_resnet/resnet.py:95

↓ 2 callersMethodlayer4

(self)

module/baseline/base_resnet/resnet.py:104

↓ 2 callersFunctionload_state_dict

Load state_dict to a module. This method is modified from :meth:`torch.nn.Module.load_state_dict`. Default value for ``strict`` is set to ``F

module/baseline/swin/checkpoint.py:41

↓ 2 callersFunctionmulti_binary_loss

(y_pred, y_true, reduction='mean', ignore_index=-1)

module/loss.py:70

↓ 2 callersMethodpatch

divide large image into small patches. Returns:

eval_window.py:62

↓ 2 callersMethodstem_forward

(self, x)

module/baseline/base_resnet/_resnets.py:205

↓ 2 callersMethodtransform

(self, inputs)

module/tta.py:125

↓ 2 callersMethodwindow_partition

Args: x: (B, H, W, C) Returns: windows: (num_windows*B, window_size, window_size, C)

module/baseline/swin/swin_transformer.py:268

↓ 1 callersMethod__init__

(self, in_strides:tuple, in_channels, channels=256, stacked_nums=2, normalized_fusion='normal

module/nas/nasdecoder.py:14

↓ 1 callersMethod__init__

(self, config)

module/nas/lovenas.py:80

↓ 1 callersMethod__init__

(self, config)

module/baseline/hrnet.py:34

↓ 1 callersMethod__init__

(self, config)

module/baseline/fcn8s.py:12

↓ 1 callersMethod_check_branches

(self, num_branches, blocks, num_blocks, num_inchannels, num_channels)

module/baseline/base_hrnet/_hrnet.py:277

↓ 1 callersMethod_check_map

(self)

module/nas/nasdecoder.py:140

↓ 1 callersMethod_forward

(self, model, image_tensor, **kwargs)

eval_window.py:92

↓ 1 callersMethod_freeze_stages

(self)

module/baseline/swin/swin_transformer.py:636

↓ 1 callersMethod_frozen_res_bn

(self)

module/baseline/base_resnet/resnet.py:112

↓ 1 callersMethod_frozen_stages

(self)

module/baseline/base_hrnet/_hrnet.py:524

↓ 1 callersFunction_load_checkpoint

Load checkpoint from somewhere (modelzoo, file, url). Args: filename (str): Accept local filepath, URL, ``torchvision://xxx``,

module/baseline/swin/checkpoint.py:225

↓ 1 callersMethod_make_branches

(self, num_branches, block, num_blocks, num_channels)

module/baseline/base_hrnet/_hrnet.py:318

↓ 1 callersMethod_make_fuse_layers

(self)

module/baseline/base_hrnet/_hrnet.py:327

↓ 1 callersMethod_make_layer

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

module/baseline/base_hrnet/_hrnet.py:507

↓ 1 callersMethod_make_one_branch

(self, branch_index, block, num_blocks, num_channels, stride=1)

module/baseline/base_hrnet/_hrnet.py:294

↓ 1 callersMethod_make_stage

(self, features, size)

module/baseline/pspnet.py:17

↓ 1 callersFunction_process_mmcls_checkpoint

(checkpoint)

module/baseline/swin/checkpoint.py:214

↓ 1 callersFunction_save_to_state_dict

Saves module state to `destination` dictionary. This method is modified from :meth:`torch.nn.Module._save_to_state_dict`. Args: modu

module/baseline/swin/checkpoint.py:370

↓ 1 callersFunctioncombine_prior

(appearance_kernel, geometry_kernel)

module/baseline/lrn.py:52

↓ 1 callersFunctionconstant_init

(module, val, bias=0)

module/baseline/base_hrnet/_hrnet.py:157

↓ 1 callersFunctionconv_with_kaiming_uniform

(use_bn=False, use_relu=False)

module/baseline/base.py:78

↓ 1 callersMethoddouble_step_seq

(step1, len1, step2, len2)

module/baseline/swin/swin_transformer.py:119

↓ 1 callersFunctionevaluate

(ckpt_path, config_path='base.hrnetw32', use_tta=False)

eval.py:30

↓ 1 callersFunctionevaluate

(ckpt_path, config_path='base.hrnetw32', use_tta=False)

eval_window.py:116

↓ 1 callersFunctionevaluate

(ckpt_path, config_path='base.hrnetw32', use_tta=False)

submit.py:31

↓ 1 callersFunctionevaluate

(ckpt_path, config_path='base.hrnetw32')

convert.py:32

↓ 1 callersFunctionget_deprecated_model_names

()

module/baseline/swin/checkpoint.py:205

↓ 1 callersFunctionget_external_models

()

module/baseline/swin/checkpoint.py:184

↓ 1 callersFunctionget_mmcls_models

()

module/baseline/swin/checkpoint.py:198

↓ 1 callersMethodget_num_inchannels

(self)

module/baseline/base_hrnet/_hrnet.py:373

↓ 1 callersFunctionget_root_logger

Get the root logger. The logger will be initialized if it has not been initialized. By default a StreamHandler will be added. If `log_file` i

module/baseline/swin/logger.py:6

↓ 1 callersFunctionget_state_dict

Returns a dictionary containing a whole state of the module. Both parameters and persistent buffers (e.g. running averages) are included. Key

module/baseline/swin/checkpoint.py:390

↓ 1 callersMethodinit_weights

(self)

module/baseline/swin/swin_transformer.py:658

↓ 1 callersFunctionload

(module, prefix='')

module/baseline/swin/checkpoint.py:67

↓ 1 callersFunctionload_fileclient_dist

In distributed setting, this function only download checkpoint at local rank 0.

module/baseline/swin/checkpoint.py:151

↓ 1 callersFunctionload_nas_encoder_weights

(encoder:nn.Module, weight_path)

module/nas/lovenas.py:58

↓ 1 callersFunctionload_pavimodel_dist

In distributed setting, this function only download checkpoint at local rank 0.

module/baseline/swin/checkpoint.py:123

↓ 1 callersMethodmerge

(self, out_list)

eval_window.py:72

↓ 1 callersFunctionmulti_binary_label

(batched_mask: torch.Tensor, num_classes, ignore_index)

module/loss.py:60

↓ 1 callersFunctionrun

(args)

tools/benchmark.py:18

↓ 1 callersMethodsame_padding

(kernel_size, dilation)

module/baseline/base.py:66

↓ 1 callersFunctionseed_torch

(seed=2333)

train_floodnet.py:43

↓ 1 callersFunctionsliding_window

(input_size, kernel_size, stride)

eval_window.py:27

↓ 1 callersFunctionsliding_window

(input_size, kernel_size, stride)

tools/prepare_floodnet.py:18

↓ 1 callersFunctionsoftmax_focalloss

Args: y_pred: [N, #class, H, W] y_true: [N, H, W] from 0 to #class gamma: scalar Returns:

module/loss.py:36

↓ 1 callersMethodstage2

(self)

module/baseline/base_hrnet/hrnet_encoder.py:62

↓ 1 callersMethodstage3

(self)

module/baseline/base_hrnet/hrnet_encoder.py:72

↓ 1 callersMethodstage4

(self)

module/baseline/base_hrnet/hrnet_encoder.py:82

↓ 1 callersFunctiontversky_loss_with_logits

(y_pred: torch.Tensor, y_true: torch.Tensor, alpha: float, beta: float, smooth_va

module/loss.py:14

↓ 1 callersFunctionweights_to_cpu

Copy a model state_dict to cpu. Args: state_dict (OrderedDict): Model weights on GPU. Returns: OrderedDict: Model weights on

module/baseline/swin/checkpoint.py:355

↓ 1 callersMethodwindow_reverse

Args: windows: (num_windows*B, window_size, window_size, C) H (int): Height of image W (int): Width of im

module/baseline/swin/swin_transformer.py:252

next →1–100 of 277, ranked by callers