MCPcopy Create free account

hub / github.com/WU-CVGL/Omni-Scene / functions

Functions190 in github.com/WU-CVGL/Omni-Scene

⨍ Functions 190◇ Types & classes 33↳ Endpoints 1
↓ 15 callersMethodtime
(self, tag: str, num_calls: int = 1)
model/utils/benchmarker.py:16
↓ 10 callersFunctioninterpolate_extrinsics
Interpolate extrinsics by rotating around their "focus point," which is the least-squares intersection between the look vectors of the initial and
model/utils/interpolation.py:208
↓ 5 callersMethod__init__
(self, use_dropout=True)
model/losses.py:40
↓ 5 callersMethod__init__
( self, in_features, hidden_features=None, out_features=None,
model/volume/vit.py:226
↓ 5 callersMethodrender
( self, gaussians: Float[Tensor, "B N F"], c2w: Float[Tensor, "B V 4 4"], fo
model/gaussian.py:156
↓ 4 callersMethoddownsample_2d
(self, tensor, h=None, w=None)
model/pixel/attention_processor.py:32
↓ 4 callersMethoddump
(self, path: Path)
model/utils/benchmarker.py:25
↓ 4 callersFunctionload_info
(info)
data/transforms/loading.py:14
↓ 4 callersFunctionmaybe_resize
(tensor, tgt_reso, interp_mode="bilinear")
model/utils/image.py:38
↓ 3 callersMethodextract_img_feat
Extract features of images.
model/omni_gs.py:66
↓ 3 callersMethodforward
(self, input, target)
model/losses.py:69
↓ 3 callersFunctiongenerate_coordinate_frame
Generate a coordinate frame given perpendicular, unit-length Y and Z vectors.
model/utils/interpolation.py:55
↓ 3 callersMethodget_data
(self, batch)
model/omni_gs.py:101
↓ 3 callersMethodget_reference_points
Get the reference points used in SCA and TSA. Args: H, W: spatial shape of tpv. Z: height of pillar. devi
model/volume/tpvformer_encoder.py:175
↓ 2 callersMethod__init__
( self, in_channels: int, out_channels: int, dropout: float = 0.0, num
model/pixel/blocks.py:236
↓ 2 callersFunctiondepths_to_colors
(depths, concat="width", max_val=None)
tools/visualization.py:27
↓ 2 callersMethoddevice
(self)
model/omni_gs.py:84
↓ 2 callersFunctionextrinsics_to_pivot_parameters
Convert the extrinsics to a representation with 5 degrees of freedom: 1. Distance from pivot point in the "X" (look cross pivot axis) direction.
model/utils/interpolation.py:107
↓ 2 callersMethodforward_demo
(self, batch)
model/omni_gs.py:398
↓ 2 callersFunctionget_ray_directions
Get ray directions for all pixels in camera coordinate. Reference: https://www.scratchapixel.com/lessons/3d-basic-rendering/ r
model/utils/ops.py:178
↓ 2 callersFunctionget_rays
( directions: Float[Tensor, "... 3"], c2w: Float[Tensor, "... 4 4"], keepdim=False, noise_scal
model/utils/ops.py:218
↓ 2 callersFunctionget_rel_pos
Get relative positional embeddings according to the relative positions of query and key sizes. Args: q_size (int): size of query
model/volume/vit.py:58
↓ 2 callersFunctionload_conditions
(img_paths, reso)
data/transforms/loading.py:32
↓ 2 callersFunctionnormalize_tensor
(x,eps=1e-10)
model/losses.py:144
↓ 1 callersFunctionHWC3
(x)
model/utils/image.py:6
↓ 1 callersMethod__init__
(self, embed_dims=256, num_cams=6, pc_range=None,
model/volume/image_cross_attention.py:34
↓ 1 callersMethod__init__
(self, ch, sr_ratio=1)
model/pixel/attention_processor.py:20
↓ 1 callersFunctionadd_decomposed_rel_pos
Args: attn (Tensor): attention map. q (Tensor): query q in the attention layer with shape (B, q_h * q_w, C).
model/volume/vit.py:93
↓ 1 callersMethodbackward
(ctx, g)
model/utils/ops.py:50
↓ 1 callersFunctionbuild_rotation
(r)
model/gaussian.py:47
↓ 1 callersFunctionbuild_scaling_rotation
(s, r)
model/gaussian.py:73
↓ 1 callersFunctioncompute_lpips
( ground_truth: Float[Tensor, "batch channel height width"], predicted: Float[Tensor, "batch channel h
tools/metrics.py:29
↓ 1 callersFunctioncompute_pcc
( ground_truth: Float[Tensor, "batch height width"], predicted: Float[Tensor, "batch height width"], )
tools/metrics.py:42
↓ 1 callersFunctioncompute_psnr
( ground_truth: Float[Tensor, "batch channel height width"], predicted: Float[Tensor, "batch channel h
tools/metrics.py:13
↓ 1 callersFunctioncompute_ssim
( ground_truth: Float[Tensor, "batch channel height width"], predicted: Float[Tensor, "batch channel h
tools/metrics.py:79
↓ 1 callersMethodconfigure_optimizers
(self, lr)
model/omni_gs.py:145
↓ 1 callersFunctionconvert_blender_to_opencv
(camera_matrix)
model/utils/camera.py:45
↓ 1 callersFunctionconvert_pose
(C2W)
model/utils/ops.py:303
↓ 1 callersFunctioncreate_camera_to_world_matrix
(elevation, azimuth, cam_dist=1.0)
model/utils/camera.py:6
↓ 1 callersFunctioncreate_logger
(log_file=None, is_main_process=False, log_level=logging.INFO)
train.py:24
↓ 1 callersFunctioncreate_logger
(log_file=None, is_main_process=False, log_level=logging.INFO)
evaluate.py:31
↓ 1 callersFunctioncreate_logger
(log_file=None, is_main_process=False, log_level=logging.INFO)
demo.py:30
↓ 1 callersFunctiondot
(x, y)
model/utils/ops.py:14
↓ 1 callersFunctioneuler_to_matrix
( rotations: Float[Tensor, "*batch 3"], pattern: str, )
model/utils/interpolation.py:96
↓ 1 callersMethodforward
Forward training function.
model/omni_gs.py:155
↓ 1 callersMethodforward_test
(self, batch)
model/omni_gs.py:338
↓ 1 callersFunctiongenerate_rotation_coordinate_frame
Generate a coordinate frame where the Y direction is normal to the plane defined by unit vectors a and b. The other axes are arbitrary.
model/utils/interpolation.py:64
↓ 1 callersFunctionget_abs_pos
(abs_pos, has_cls_token, hw)
model/volume/vit.py:37
↓ 1 callersFunctionget_cam_info_gaussian
(c2w, fovx, fovy, znear, zfar)
model/utils/ops.py:356
↓ 1 callersMethodget_cross_view_ref_points
(tpv_h, tpv_w, tpv_z, num_points_in_pillar)
model/volume/tpvformer_encoder.py:88
↓ 1 callersFunctionget_lpips
(device: torch.device)
tools/metrics.py:24
↓ 1 callersFunctionget_pcc
(device: torch.device)
tools/metrics.py:38
↓ 1 callersFunctionget_projection_matrix_gaussian
(znear, zfar, fovX, fovY, device="cuda")
model/utils/ops.py:311
↓ 1 callersMethodget_reference_points
Get the reference points used in spatial cross-attn and self-attn. Args: H, W: spatial shape of tpv plane. Z: hight of
model/volume/volume_gs_decoder.py:61
↓ 1 callersMethodget_sampling_offsets_and_attention
( self, queries: List[Tensor])
model/volume/cross_view_hybrid_attention.py:115
↓ 1 callersMethodget_sampling_offsets_and_attention
(self, queries)
model/volume/image_cross_attention.py:307
↓ 1 callersFunctiongray_to_colormap
Transfer gray map to matplotlib colormap
tools/visualization.py:8
↓ 1 callersMethodinit_weight
Default initialization for Parameters of Module.
model/volume/image_cross_attention.py:62
↓ 1 callersMethodinit_weights
(self)
model/volume/vit.py:426
↓ 1 callersMethodinit_weights
Initialize the transformer weights.
model/volume/tpvformer_encoder.py:75
↓ 1 callersFunctioninterpolate_circular
( a: Float[Tensor, "*#batch"], b: Float[Tensor, "*#batch"], t: Float[Tensor, "*#batch"], )
model/utils/interpolation.py:162
↓ 1 callersFunctioninterpolate_pivot_parameters
( initial: Float[Tensor, "*#batch 5"], final: Float[Tensor, "*#batch 5"], t: Float[Tensor, " time_
model/utils/interpolation.py:190
↓ 1 callersFunctionintersect_rays
Compute the least-squares intersection of rays. Uses the math from here: https://math.stackexchange.com/a/1762491/286022
model/utils/interpolation.py:19
↓ 1 callersFunctioninverse_sigmoid
(x)
evaluate.py:25
↓ 1 callersFunctioninverse_sigmoid
(x)
model/gaussian.py:28
↓ 1 callersMethodload_from_pretrained
(self, name="vgg_lpips")
model/losses.py:54
↓ 1 callersFunctionmain
(args)
train.py:44
↓ 1 callersFunctionmain
(args)
evaluate.py:49
↓ 1 callersFunctionmain
(args)
demo.py:48
↓ 1 callersFunctionmatrix_to_euler
( rotations: Float[Tensor, "*batch 3 3"], pattern: str, )
model/utils/interpolation.py:85
↓ 1 callersFunctionmaybe_resize
(img, tgt_reso, ck)
data/transforms/loading.py:34
↓ 1 callersFunctionnormalize
(a: Float[Tensor, "*#batch dim"])
model/utils/interpolation.py:51
↓ 1 callersFunctionpivot_parameters_to_extrinsics
( parameters: Float[Tensor, "*#batch 5"], pivot_coordinate_frame: Float[Tensor, "*#batch 3 3"], pi
model/utils/interpolation.py:136
↓ 1 callersMethodplucker_embedder
( self, rays_o, rays_d )
model/omni_gs.py:91
↓ 1 callersMethodpoint_sampling
(self, reference_points, pc_range, img_metas)
model/volume/tpvformer_encoder.py:211
↓ 1 callersMethodreshape_output
(self, output: Tensor, lens: List[int])
model/volume/cross_view_hybrid_attention.py:139
↓ 1 callersMethodreshape_output
(self, output, lens)
model/volume/image_cross_attention.py:340
↓ 1 callersMethodreshape_reference_points
(self, reference_points)
model/volume/image_cross_attention.py:330
↓ 1 callersFunctionsave_ply
(gaussians, path, crop_range=[-50.0, -50.0, -3.0, 50.0, 50.0, 12.0], compatible=True)
evaluate.py:231
↓ 1 callersMethodsave_ply
(self, gaussians, path, compatible=True)
model/gaussian.py:256
↓ 1 callersMethodsave_val_results
(self, batch_gt, render_pkg_fuse, render_pkg_pixel, render_pkg_volume, gaussians_all,
model/omni_gs.py:484
↓ 1 callersMethodsetup_functions
(self)
model/gaussian.py:139
↓ 1 callersFunctionspatial_average
(x, keepdim=True)
model/losses.py:149
↓ 1 callersFunctionstrip_lowerdiag
(L)
model/gaussian.py:31
↓ 1 callersFunctionstrip_symmetric
(sym)
model/gaussian.py:43
↓ 1 callersMethodvalidation_step
(self, batch, val_result_savedir)
model/omni_gs.py:329
↓ 1 callersFunctionwindow_partition
Args: x: (B, H, W, C) window_size (int): window size Returns: windows: (num_windows*B, window_size, window_size, C)
model/volume/vit.py:127
↓ 1 callersFunctionwindow_unpartition
Args: windows: (num_windows*B, window_size, window_size, C) window_size (int): Window size H (int): Height of image
model/volume/vit.py:150
FunctionRGB2SH
(rgb)
model/gaussian.py:20
FunctionSH2RGB
(sh)
model/gaussian.py:24
Method__call__
( self, attn: Attention, hidden_states: torch.FloatTensor, encoder_hidden_stat
model/pixel/attention_processor.py:53
Method__call__
( self, attn: Attention, hidden_states: torch.FloatTensor, encoder_hidden_stat
model/pixel/attention_processor.py:144
Method__getitem__
(self, index)
data/dataloader.py:140
Method__init__
(self, backbone=None, neck=None, pixel_gs=None,
model/omni_gs.py:27
Method__init__
(self, *args, **kwargs)
model/gaussian.py:85
Method__init__
( self, device, resolution: list = [512, 512], znear: float = 0.1, zf
model/gaussian.py:118
Method__init__
(self, use_second_derivative=False, sigma_image=False, near=0.1, far=1000.0)
model/losses.py:13
Method__init__
(self)
model/losses.py:86
Method__init__
(self, chn_in, chn_out=1, use_dropout=False)
model/losses.py:97
Method__init__
(self, requires_grad=False, pretrained=True)
model/losses.py:105
next →1–100 of 190, ranked by callers