Functions — github.com/cvg/NoPoSplat

CodeHub MCPcopy Create free account

hub / github.com/cvg/NoPoSplat / functions

Functions553 in github.com/cvg/NoPoSplat

⨍ Functions 553◇ Types & classes 135↳ Endpoints 1

Methoddata_stage

src/dataset/dataset_scannet_pose.py:250

Methoddata_stage

src/dataset/dataset_re10k.py:247

Functiondecompose_extrinsic_RT

Decompose the standard extrinsic matrix into RT. Batched I/O.

src/misc/cam_utils.py:8

Functiondepthmap_to_absolute_camera_coordinates

Args: - depthmap (HxW array): - camera_intrinsics: a 3x3 matrix - camera_pose: a 4x3 or 4x4 cam2world matrix Returns:

src/geometry/ptc_geometry.py:201

src/loss/loss_point.py:62

Methodestimate_pose

(self, context, normalize=False)

src/model/distiller/dust3d_backbone.py:210

Methodextra_repr

src/model/encoder/backbone/croco/blocks.py:55

Functionfig_to_image

( fig: Figure, dpi: int = 100, device: torch.device = torch.device("cpu"), )

src/misc/image_io.py:22

Functionfind_reciprocal_matches

returns 3 values: 1 - reciprocal_in_P2: a boolean array of size P2.shape[0], a "True" value indicates a match 2 - nn2_in_P1: a int array

src/geometry/ptc_geometry.py:364

flip so that tensor[0::2] <=> tensor[1::2]

src/model/encoder/backbone/croco/misc.py:43

( self, samples: Float[Tensor, "*batch dim"], )

src/model/encodings/positional_encoding.py:28

( self, context: dict, global_step: int = 0, visualization_dump: Optional[dict

src/model/encoder/encoder_noposplat.py:139

( self, context: dict, global_step: int = 0, visualization_dump: Optional[dict

src/model/encoder/encoder_noposplat_multi.py:126

( self, context: BatchedViews, )

src/model/encoder/backbone/backbone_resnet.py:66

(self, context: dict, symmetrize_batch=False, return_views=Fal

src/model/encoder/backbone/backbone_croco.py:222

( self, context: BatchedViews, )

src/model/encoder/backbone/backbone_dino.py:41

( self, context: BatchedViews, )

src/model/encoder/backbone/backbone.py:20

(self, context: dict, symmetrize_batch=False, return_views=Fal

src/model/encoder/backbone/backbone_croco_multiview.py:190

src/model/encoder/backbone/croco/blocks.py:52

(self, x, xpos)

src/model/encoder/backbone/croco/blocks.py:94

(self, x, xpos)

src/model/encoder/backbone/croco/blocks.py:127

(self, query, key, value, qpos, kpos)

src/model/encoder/backbone/croco/blocks.py:149

(self, x, y, xpos, ypos)

src/model/encoder/backbone/croco/blocks.py:186

src/model/encoder/backbone/croco/blocks.py:227

input: * tokens: batch_size x nheads x ntokens x dim * positions: batch_size x ntokens x 2 (y and x posit

src/model/encoder/backbone/croco/pos_embed.py:142

img1: tensor of size B x 3 x img_size x img_size img2: tensor of size B x 3 x img_size x img_size out will be B x

src/model/encoder/backbone/croco/croco.py:231

(self, x, **kw)

src/model/encoder/backbone/croco/patch_embed.py:20

(self, img, true_shape)

src/model/encoder/backbone/croco/patch_embed.py:42

(ctx, tokens, positions, base, F0=1)

src/model/encoder/backbone/croco/curope/curope2d.py:15

(self, tokens, positions)

src/model/encoder/backbone/croco/curope/curope2d.py:38

( self, extrinsics: Float[Tensor, "*#batch 4 4"], intrinsics: Float[Tensor, "*#batch 3

src/model/encoder/common/gaussian_adapter.py:50

( self, means: Float[Tensor, "*#batch 3"], depths: Float[Tensor, "*#batch"], o

src/model/encoder/common/gaussian_adapter.py:123

(self, encoder_tokens: List[torch.Tensor], image_size=None, ray_embedding=None)

src/model/encoder/heads/dpt_head.py:35

(self, x, img_info, ray_embedding=None)

src/model/encoder/heads/dpt_head.py:94

(self, decout, img_shape)

src/model/encoder/heads/linear_head.py:30

(self, decout, img_shape)

src/model/encoder/heads/linear_head.py:62

(self, encoder_tokens: List[torch.Tensor], depths, imgs, image_size=None, conf=None)

src/model/encoder/heads/dpt_gs_head.py:120

(self, x, depths, imgs, img_info, conf=None)

src/model/encoder/heads/dpt_gs_head.py:181

src/model/encoder/heads/head_modules.py:47

src/model/encoder/heads/head_modules.py:84

src/model/encoder/heads/head_modules.py:112

Forward pass. Args: x (tensor): input Returns: tensor: output

src/model/encoder/heads/dpt_block.py:121

Forward pass. Returns: tensor: output

src/model/encoder/heads/dpt_block.py:189

Forward pass. Args: x (tensor): input Returns: tensor: interpolated data

src/model/encoder/heads/dpt_block.py:247

(self, encoder_tokens: List[torch.Tensor], image_size)

src/model/encoder/heads/dpt_block.py:428

( self, gaussians: Gaussians, extrinsics: Float[Tensor, "batch view 4 4"], int

src/model/decoder/decoder.py:35

( self, gaussians: Gaussians, extrinsics: Float[Tensor, "batch view 4 4"], int

src/model/decoder/decoder_splatting_cuda.py:37

(self, x, z=None)

src/model/transformer/attention.py:54

(self, x, z=None, **kwargs)

src/model/transformer/transformer.py:67

(self, x, **kwargs)

src/model/transformer/pre_norm.py:34

src/model/transformer/feed_forward.py:39

Computes the spherical harmonics.

src/misc/sht.py:1439

src/loss/loss_point.py:43

(self, *args, **kwargs)

src/loss/loss_point.py:108

(self, gt_pts1, gt_pts2, pr_pts1, pr_pts2, conf1=None, conf2=None, dist_clip=None, disable_view1=False)

src/loss/loss_point.py:208

( self, prediction: DecoderOutput, batch: BatchedExample, gaussians: Gaussians

src/loss/loss_mse.py:23

( self, prediction: DecoderOutput, batch: BatchedExample, gaussians: Gaussians

src/loss/loss_lpips.py:36

(self, X: Tensor, Y: Tensor)

src/loss/loss_ssim.py:305

(self, X: Tensor, Y: Tensor)

src/loss/loss_ssim.py:348

( self, prediction: DecoderOutput, batch: BatchedExample, gaussians: Gaussians

src/loss/loss_depth.py:27

Functiongather_discrete_topk

( pdf: Float[Tensor, "*batch bucket"], num_samples: int, eps: float = torch.finfo(torch.float32).e

src/misc/discrete_probability_distribution.py:23

Methodgenerate_ctx_views

src/model/encoder/backbone/backbone_croco_multiview.py:159

Functiongenerate_heterogeneous_index_transpose

Generate an index that can be used to "transpose" the heterogeneous index. Applying the index a second time inverts the "transpose."

src/misc/heterogeneous_pairings.py:27

Functiongenerate_spin

( num_frames: int, device: torch.device, elevation: float, radius: float, )

src/visualization/camera_trajectory/spin.py:9

Methodget_conf_log

src/loss/loss_point.py:205

Methodget_data_shim

src/model/encoder/encoder_noposplat.py:242

Methodget_data_shim

src/model/encoder/encoder_noposplat_multi.py:229

Functionget_depth

Calculate the depths that correspond to the specified 2D points on the epipolar lines defined by the origins and directions. The extrinsics and in

src/geometry/epipolar_lines.py:280

Functionget_joint_pointcloud_center_scale

(pts1, pts2, valid_mask1=None, valid_mask2=None, z_only=False, center=True)

src/geometry/ptc_geometry.py:347

Functionget_joint_pointcloud_depth

(z1, z2, valid_mask1, valid_mask2=None, quantile=0.5)

src/geometry/ptc_geometry.py:332

Functionget_med_dist_between_poses

src/geometry/ptc_geometry.py:383

src/loss/loss_point.py:132

src/loss/loss_point.py:161

Functionget_pred_pts3d

(gt, pred, use_pose=False)

src/loss/loss_point.py:10

Functionget_seed

src/global_cfg.py:18

Methodglobal_step

src/dataset/view_sampler/view_sampler.py:60

(module, input, output)

src/model/encoder/visualization/encoder_visualizer_epipolar.py:47

Functioninterpolate_pos_embed

(model, checkpoint_model)

src/model/encoder/backbone/croco/pos_embed.py:78

Invert a torch or numpy matrix

src/geometry/ptc_geometry.py:105

Functioninverse_normalize_image

(tensor, mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5))

src/dataset/shims/normalize_shim.py:9

Methodload_state_dict

(self, ckpt, **kw)

src/model/encoder/backbone/backbone_croco.py:109

Methodload_state_dict

(self, ckpt, **kw)

src/model/encoder/backbone/backbone_croco_multiview.py:100

Methodlog_hyperparams

src/misc/LocalLogger.py:27

Methodlog_metrics

(self, metrics, step)

src/misc/LocalLogger.py:31

src/misc/LocalLogger.py:19

Methodnum_context_views

src/dataset/view_sampler/view_sampler_arbitrary.py:72

Methodnum_context_views

src/dataset/view_sampler/view_sampler.py:56

Methodnum_context_views

src/dataset/view_sampler/view_sampler_bounded.py:133

Methodnum_context_views

src/dataset/view_sampler/view_sampler_all.py:32

Methodnum_context_views

src/dataset/view_sampler/view_sampler_evaluation.py:78

Methodnum_target_views

src/dataset/view_sampler/view_sampler_arbitrary.py:76

Methodnum_target_views

src/dataset/view_sampler/view_sampler.py:51

Methodnum_target_views

src/dataset/view_sampler/view_sampler_bounded.py:137

Methodnum_target_views

src/dataset/view_sampler/view_sampler_all.py:36

Methodnum_target_views

src/dataset/view_sampler/view_sampler_evaluation.py:82

Functionobtain_pointcloud_center

Args: - pts (BxNx3 torch.array): pointmap of absolute coordinates - valid_mask (BxN torch.array): mask specifying valid pixels.

src/geometry/ptc_geometry.py:247

Methodon_test_end

src/model/model_wrapper.py:357

Methodon_test_end

src/evaluation/pose_evaluator.py:173

Functionopencv_to_colmap_intrinsics

Modify camera intrinsics to follow a different convention. Coordinates of the center of the top-left pixels are by default: - (0.5, 0.5)

src/geometry/ptc_geometry.py:234

src/model/encoder/backbone/croco/blocks.py:25

← previous next →401–500 of 553, ranked by callers