Functions — github.com/NVlabs/RADIO

CodeHub MCPcopy Create free account

hub / github.com/NVlabs/RADIO / functions

Functions572 in github.com/NVlabs/RADIO

⨍ Functions 572◇ Types & classes 107

(self, x: torch.Tensor)

radio/dinov2_arch.py:139

(self, x: torch.Tensor, attn_bias=None)

radio/dinov2_arch.py:166

(self, x: torch.Tensor)

radio/dinov2_arch.py:228

radio/dinov2_arch.py:279

(self, x: torch.Tensor)

radio/dinov2_arch.py:294

(self, x: torch.Tensor)

radio/dinov2_arch.py:368

(self, x_or_x_list)

radio/dinov2_arch.py:437

radio/dinov2_arch.py:554

(self, *args, is_training=False, **kwargs)

radio/dinov2_arch.py:827

Forward pass through C2f layer.

radio/eradio_model.py:54

forward()' applies the YOLOv5 FPN to input data.

radio/eradio_model.py:77

radio/eradio_model.py:97

radio/eradio_model.py:139

radio/eradio_model.py:184

(self, input_tensor)

radio/eradio_model.py:323

radio/eradio_model.py:451

radio/eradio_model.py:574

radio/eradio_model.py:611

radio/eradio_model.py:652

radio/eradio_model.py:689

radio/eradio_model.py:722

(self, x, attn_mask = None)

radio/eradio_model.py:768

radio/eradio_model.py:911

radio/eradio_model.py:962

(self, x, il_level=-1, full_features=None)

radio/eradio_model.py:1012

radio/eradio_model.py:1190

(self, x: torch.Tensor)

radio/cls_token.py:44

Forward pass through C2f layer.

radio/block.py:32

forward()' applies the YOLOv5 FPN to input data.

radio/block.py:56

(self, weight: torch.Tensor, *args, **kwargs)

radio/enable_spectral_reparam.py:76

(self, weight: torch.Tensor, *args, **kwargs)

radio/enable_spectral_reparam.py:115

(self, x: torch.Tensor)

radio/adaptor_mlp.py:43

(self, x: torch.Tensor, images: Optional[torch.Tensor] = None, patch_size: Optional[int] = None)

radio/adaptor_mlp.py:92

(self, input: AdaptorInput)

radio/adaptor_generic.py:65

(self, x: torch.Tensor)

radio/vit_patch_generator.py:94

(self, x: torch.Tensor)

radio/vit_patch_generator.py:281

(self, x: torch.Tensor)

radio/hf_model.py:207

Forward process for model. Args: x: Input tensor. Unless `make_preprocessor_external` has been called, then the dynamic r

radio/radio_model.py:181

(self, src: torch.Tensor, tgt: torch.Tensor)

radio/dual_hybrid_vit.py:30

(self, x: torch.Tensor, twod_shape: Tuple[int, int])

radio/dual_hybrid_vit.py:76

(self, x: torch.Tensor)

radio/dual_hybrid_vit.py:150

(self, x: torch.Tensor)

radio/extra_models.py:41

(self, *args, **kwargs)

radio/extra_models.py:150

Forward pass. Args: x: Input features with shape (B, H, W, C). Returns: Output features with shape (B, H//2,

radio/radio1d.py:261

(ctx: torch.autograd.function.FunctionCtx, x: torch.Tensor, lambda_: Union[float, torch.Tensor] = 1.0)

radio/radio1d.py:328

Forward pass. Args: x: Input features with shape (B, H, W, C). Returns: Output features with shape (B, 2*H,

radio/radio1d.py:434

Forward pass through decoder. Args: global_tokens: Global tokens from encoder (B, num_tokens, C_in), possibly padded

radio/radio1d.py:613

Forward pass through encoder only. Args: x: Input image tensor of shape (B, C, H, W) attn_mask: Optional attention ma

radio/radio1d.py:1668

(self, q: torch.Tensor, k: torch.Tensor, token_shape: Tuple[int, int])

radio/vision_transformer_xpos.py:107

(self, x: torch.Tensor, num_prefix_tokens: int, patch_shape: Tuple[int, int])

radio/vision_transformer_xpos.py:134

(self, x: torch.Tensor, num_prefix_tokens: int, patch_shape: Tuple[int, int])

radio/vision_transformer_xpos.py:205

radio/conv.py:39

(self, x: torch.Tensor, **kwargs)

radio/adaptor_attn.py:45

(self, x: torch.Tensor)

radio/input_conditioner.py:31

(self, input: AdaptorInput)

radio/adaptor_base.py:53

(self, x: torch.Tensor)

examples/position_predictor.py:359

examples/count_params.py:49

(self, x, *args, feature_fmt: str = 'NLC', **kwargs)

examples/common/model_loader.py:54

(self, *args, **kwargs)

examples/common/model_loader.py:91

(self, *args, **kwargs)

examples/common/model_loader.py:138

(self, x: torch.Tensor, *args, **kwargs)

examples/common/model_loader.py:156

(self, x: torch.Tensor, **kwargs)

examples/common/model_loader.py:215

(self, x: torch.Tensor, **kwargs)

examples/common/model_loader.py:253

(self, x: torch.Tensor)

examples/common/model_loader.py:305

Methodforward_features

(self, x: torch.Tensor)

radio/extra_models.py:54

Methodforward_features

(self, x: torch.Tensor)

radio/vision_transformer_xpos.py:290

Methodforward_intermediates

Forward features that returns intermediates. Args: x: Input image tensor indices: Take last n blocks if int, select m

radio/radio_model.py:277

Methodforward_intermediates

(self, x: torch.Tensor, norm: bool = False, **kwargs, )

radio/extra_models.py:168

Methodforward_intermediates

(self, x: torch.Tensor, norm: bool = False, **kwargs)

radio/vision_transformer_xpos.py:301

Methodforward_split

Forward pass using split() instead of chunk().

radio/eradio_model.py:60

Methodforward_split

Forward pass using split() instead of chunk().

radio/block.py:38

Methodget_expected_tokens

(self, outside_max: int)

radio/radio1d.py:728

Methodget_expected_tokens

(self, outside_max: int)

radio/radio1d.py:757

Methodget_expected_tokens

(self, outside_max: int)

radio/radio1d.py:777

Methodget_expected_tokens

(self, outside_max: int)

radio/radio1d.py:798

Methodget_first_downscale_block_idx

Return the index of the first downscaling block, or None if no downscaling.

radio/radio1d.py:1662

Methodget_intermediate_layers

( self, x: torch.Tensor, n: Union[int, Sequence] = 1, # Layers or n last layers to ta

radio/dinov2_arch.py:801

Methodget_max_tokens

(self, outside_max: int)

radio/radio1d.py:711

Methodget_max_tokens

(self, outside_max: int)

radio/radio1d.py:795

Functionget_scale_map

scalar_map: (1, h, w, C) is the feature map of a single image.

examples/visualize_features.py:631

Methodhead_reshape

(t: torch.Tensor)

radio/vision_transformer_xpos.py:152

Functionhybrid_base

(pretrained=False, concatenate: bool = False, weight_init: str = 'skip', **kwargs)

radio/dual_hybrid_vit.py:190

Functionhybrid_huge

(pretrained=False, concatenate: bool = False, weight_init: str = 'skip', **kwargs)

radio/dual_hybrid_vit.py:208

Functionhybrid_large

(pretrained=False, concatenate: bool = False, weight_init: str = 'skip', **kwargs)

radio/dual_hybrid_vit.py:199

Methodinit_weights

mmseg/linear_head.py:28

Methodmake_params

(self, flat_inputs: List[Any])

examples/common/resize_transform.py:35

Methodmake_params

(self, flat_inputs: List[Any])

examples/common/resize_transform.py:93

Methodmake_preprocessor_external

radio/radio_model.py:147

Methodmax_resolution

radio/hf_model.py:183

Methodmax_resolution

radio/radio_model.py:111

Methodmin_resolution_step

radio/hf_model.py:195

Methodmin_resolution_step

radio/radio_model.py:123

Functionnamed_apply

(fn: Callable, module: nn.Module, name="", depth_first=True, include_root=False)

radio/dinov2_arch.py:542

Methodno_fsdp_wrap_types

radio/dual_hybrid_vit.py:147

Methodno_weight_decay

radio/cls_token.py:56

Methodno_weight_decay

radio/vit_patch_generator.py:123

Methodno_weight_decay_keywords

radio/eradio_model.py:1167

Methodnum_cls_patches

radio/vit_patch_generator.py:112

Methodnum_cls_tokens

radio/vit_patch_generator.py:108

Methodnum_cls_tokens

radio/radio_model.py:96

← previous next →401–500 of 572, ranked by callers