Functions — github.com/LayneH/GreenMIM

hub / github.com/LayneH/GreenMIM / functions

Functions148 in github.com/LayneH/GreenMIM

⨍ Functions 148◇ Types & classes 34

↓ 35 callersFunctionprint

(*args, **kwargs)

util/misc.py:176

↓ 10 callersMethod__init__

(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0., dr

modeling/green_twins_models.py:97

↓ 5 callersMethod__init__

(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.)

modeling/green_swin_models.py:21

↓ 5 callersFunctionget_coordinates

(h, w, device='cpu')

modeling/group_window_attention.py:13

↓ 3 callersMethod__init__

(self, mode='train', max_class=1000, num_classes=1000, transform=None, percent=1., shu

util/base_dataset.py:112

↓ 3 callersFunctionis_dist_avail_and_initialized

()

util/misc.py:187

↓ 3 callersMethodload_state_dict

(self, state_dict)

util/misc.py:293

↓ 3 callersMethodprepare

(self, coords, num_tokens)

modeling/group_window_attention.py:275

↓ 3 callersMethodstate_dict

(self)

util/misc.py:290

↓ 3 callersFunctionto_sparse_tensor

(x: torch.Tensor, indexes: torch.Tensor, H: int, W: int)

modeling/sparse_conv_spconv.py:12

↓ 3 callersFunctionto_sparse_tensor

(x: torch.Tensor, indexes: torch.Tensor)

modeling/sparse_conv_me.py:12

↓ 3 callersMethodupdate

(self, **kwargs)

util/misc.py:91

↓ 2 callersMethod_get_attn_mask

(self, group_id)

modeling/group_window_attention.py:203

↓ 2 callersMethod_get_group_id

(self, coords)

modeling/group_window_attention.py:196

↓ 2 callersMethod_get_rel_pos_idx

(self, coords)

modeling/group_window_attention.py:212

↓ 2 callersMethodapply_mask

(self, x, mask, patches_resolution)

modeling/base_green_models.py:19

↓ 2 callersMethodforward

(self, x: torch.Tensor, indexes: torch.Tensor, H: int, W: int)

modeling/sparse_conv_spconv.py:35

↓ 2 callersFunctionget_1d_sincos_pos_embed_from_grid

embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)

util/pos_embed.py:49

↓ 2 callersFunctionget_world_size

()

util/misc.py:195

↓ 2 callersMethodgroup

(self, x)

modeling/group_window_attention.py:283

↓ 2 callersMethodload_image

(self, filename)

util/base_dataset.py:30

↓ 2 callersMethodmerge

(self, x)

modeling/group_window_attention.py:290

↓ 2 callersFunctionsetup_for_distributed

This function disables printing when not in master process

util/misc.py:170

↓ 2 callersMethodupdate

(self, value, n=1)

util/misc.py:37

↓ 1 callersMethod__init__

(self, window_size, shift_size, group_size=None)

modeling/group_window_attention.py:187

↓ 1 callersMethod__init__

(self, *args, **kwargs)

modeling/sparse_conv_me.py:24

↓ 1 callersMethod__len__

(self)

util/base_dataset.py:121

↓ 1 callersMethod_init_memcached

(self)

util/base_dataset.py:23

↓ 1 callersMethod_prepare_grouping

(self, coords)

modeling/group_window_attention.py:232

↓ 1 callersMethod_prepare_masking

(self, coords)

modeling/group_window_attention.py:220

↓ 1 callersMethodadd_meter

(self, name, meter)

util/misc.py:120

↓ 1 callersFunctionbuild_transform

(is_train, args)

util/datasets.py:34

↓ 1 callersMethodforward

(self, x, H, W)

modeling/green_twins_models.py:110

↓ 1 callersMethodforward

(self, x: torch.Tensor, indexes: torch.Tensor, H: int = None, W: int = None)

modeling/sparse_conv_me.py:27

↓ 1 callersMethodforward_decoder

(self, x, ids_restore)

modeling/base_green_models.py:192

↓ 1 callersMethodforward_encoder

(self, x, mask_ratio)

modeling/base_green_models.py:183

↓ 1 callersMethodforward_features

(self, x, mask)

modeling/base_green_models.py:49

↓ 1 callersMethodforward_loss

imgs: [N, 3, H, W] pred: [N, L, p*p*3] mask: [N, L], 0 is keep, 1 is remove,

modeling/base_green_models.py:214

↓ 1 callersFunctionget_2d_sincos_pos_embed

grid_size: int of the grid height and width return: pos_embed: [grid_size*grid_size, embed_dim] or [1+grid_size*grid_size, embed_dim] (w/

util/pos_embed.py:20

↓ 1 callersFunctionget_2d_sincos_pos_embed_from_grid

(embed_dim, grid)

util/pos_embed.py:38

↓ 1 callersFunctionget_args_parser

()

main_pretrain.py:35

↓ 1 callersFunctionget_grad_norm_

(parameters, norm_type: float = 2.0)

util/misc.py:297

↓ 1 callersFunctionget_keep_index

(samples, percent, num_classes, shuffle=False)

util/base_dataset.py:97

↓ 1 callersFunctionget_layer_id_for_vit

Assign a parameter with its layer id Following BEiT: https://github.com/microsoft/unilm/blob/master/beit/optim_factory.py#L33

util/lr_decay.py:64

↓ 1 callersFunctionget_pretrain_model

(model_str, norm_pix_loss=True)

modeling/model_factory.py:15

↓ 1 callersFunctionget_rank

()

util/misc.py:201

↓ 1 callersFunctiongroup_windows

Greedily apply the DP algorithm to group the elements. Args: group_size (int): maximal size of the group num_ele_win (list[int]):

modeling/group_window_attention.py:157

↓ 1 callersMethodinitialize_weights

(self)

modeling/base_green_models.py:94

↓ 1 callersFunctionis_main_process

()

util/misc.py:207

↓ 1 callersFunctionknapsack

Args: W (int): capacity wt (tuple[int]): the numbers of elements within each window

modeling/group_window_attention.py:111

↓ 1 callersFunctionload_pretrained

(ckpt_path, save_path, model_type='swin')

util/extract_backbone.py:8

↓ 1 callersMethodlog_every

(self, iterable, print_freq, header=None)

util/misc.py:123

↓ 1 callersFunctionmain

↓ 1 callersFunctionparse_host_addr

(s)

util/misc.py:238

↓ 1 callersMethodpatchify

(self, x)

modeling/green_swin_models.py:387

↓ 1 callersMethodpatchify

imgs: (N, 3, H, W) x: (N, L, patch_size**2 *3)

modeling/base_green_models.py:126

↓ 1 callersMethodpatchify

(self, x)

modeling/green_twins_models.py:335

↓ 1 callersMethodrandom_masking

NOTE: Perform PER-BATCH random masking by per-sample shuffling. Per-batch shuffling is done by argsort random noise. x: [N, L

modeling/base_green_models.py:154

↓ 1 callersFunctionremap_pretrained_keys_swin

(checkpoint_model)

util/extract_backbone.py:37

↓ 1 callersFunctionsave_on_master

(*args, **kwargs)

util/misc.py:211

↓ 1 callersMethodsynchronize_between_processes

(self)

util/misc.py:116

↓ 1 callersFunctiontrain_one_epoch

(model: torch.nn.Module, data_loader: Iterable, optimizer: torch.optim.Optimizer,

engine_pretrain.py:22

Method__call__

(self, loss, optimizer, clip_grad=None, parameters=None, create_graph=False, update_grad=True)

util/base_dataset.py:124

Method__getitem__

(self, index)

util/base_dataset.py:146

Method__init__

(self, in_channels: int, out_channels: int, kernel_size: int = 1, stride: int = 1, padding: i

modeling/sparse_conv_spconv.py:45

Method__init__

(self, dim, input_resolution, num_heads, window_size=7, shift_size=0, mlp_ratio=4., qkv_bias=

modeling/green_swin_models.py:57

Method__init__

(self, input_resolution, dim, norm_layer=nn.LayerNorm)

modeling/green_swin_models.py:110

Method__init__

(self, dim, input_resolution, depth, num_heads, window_size, mlp_ratio=4., qkv_bias=True, qk_

modeling/green_swin_models.py:174

Method__init__

(self, img_size=224, patch_size=4, in_chans=3, embed_dim=96, norm_layer=None)

modeling/green_swin_models.py:257

Method__init__

(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000, embed_dim=96, depths=[2, 2,

modeling/green_swin_models.py:313

Method__init__

(self, dim, window_size, num_heads, qkv_bias=True, qk_scale=None, attn_drop=0., proj_drop=0.)

modeling/group_window_attention.py:34

Method__init__

(self, encoder, embed_dim, patch_size, in_chans=3, decoder_num_patches=196, decoder_embed_dim

modeling/base_green_models.py:60

Method__init__

(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0., sr_ratio=1, norm_layer=nn.

modeling/green_twins_models.py:52

Method__init__

(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0., dr

modeling/green_twins_models.py:118

Method__init__

(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0., dr

modeling/green_twins_models.py:128

Method__init__

(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768, norm_layer=nn.LayerNorm)

modeling/green_twins_models.py:151

Method__init__

(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768, norm_layer=nn.LayerNorm)

modeling/green_twins_models.py:179

Method__init__

(self, img_size=224, patch_size=16, in_chans=3, num_classes=1000, embed_dims=[64, 128, 256, 512],

modeling/green_twins_models.py:211

Method__init__

(self, in_chans, embed_dim=768, s=1)

modeling/green_twins_models.py:280

Method__init__

(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000, embed_dims=[64, 128, 256, 512],

modeling/green_twins_models.py:300

Method__init__

(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000, embed_dims=[64, 128, 256], n

modeling/green_twins_models.py:373

Method__init__

(self, img_size=224, patch_size=4, in_chans=3, num_classes=1000, embed_dims=[64, 128, 256], n

modeling/green_twins_models.py:386

Method__init__

(self, *args, **kwargs)

modeling/sparse_conv_me.py:50

Method__init__

(self, window_size=20, fmt=None)

util/misc.py:29

Method__init__

(self, delimiter="\t")

util/base_dataset.py:18

Method__init__

(self, mode='train', max_class=1000, aug=None, prefix='/mnt/cache/share/images/meta',

util/base_dataset.py:48

Method__init__

(self, mode='train', max_class=1000, num_classes=1000, transform=None, idx=None, shuff

util/base_dataset.py:132

Method__len__

(self)

util/base_dataset.py:143

modeling/green_swin_models.py:370

Method_init_weights

(self, m)

modeling/base_green_models.py:116

Method_init_weights

(self, m)

modeling/green_twins_models.py:265

Method_init_weights

(self, m)

modeling/green_twins_models.py:313

Functionadjust_learning_rate

Decay the learning rate with half-cycle cosine after warmup

util/lr_sched.py:9

next →1–100 of 148, ranked by callers