MCPcopy Create free account

hub / github.com/Yukun-Huang/DreamCube / functions

Functions331 in github.com/Yukun-Huang/DreamCube

↓ 36 callersFunctionsafe_setattr
(obj, attr, value, mode=None)
models/multiplane_sync/utils.py:4
↓ 24 callersFunctionvid
(f, y, x)
utils/pano_to_3d.py:588
↓ 19 callersMethodflip
(self, flip=True)
utils/cube.py:526
↓ 12 callersFunctionadaptive_convolution
进行基于 theta 角度自适应旋转的卷积
models/multiplane_sync/sync_conv2d.py:207
↓ 10 callersMethodsave
(self, save_path, cube_format='horizon')
utils/cube.py:536
↓ 8 callersMethodcube_all2all
(faces, cube_format, target_cube_format='horizon')
utils/cube.py:353
↓ 8 callersFunctionget_device
Get device of a Tensor
external/equilib/torch_utils/func.py:12
↓ 7 callersMethodnormalize
Args: depth: [B, ...], torch.FloatTensor, [0.0, +inf] scale: float min_value: [B], torch.FloatTensor, [0.
utils/depth.py:607
↓ 6 callersFunctionmatrix
(*rows, device=None)
utils/ops/resample_padding.py:11
↓ 6 callersFunctionpad_cube
Args: cube: torch.Tensor, [b*m, c, h, w] or [b, m, c, h, w] padding: int
utils/cube.py:18
↓ 5 callersFunctioncube_sync_attn_processor
(self)
models/multiplane_sync_legacy.py:289
↓ 5 callersFunctioncube_sync_attn_processor
(self)
models/multiplane_sync/sync_attn.py:28
↓ 5 callersFunctioncube_sync_conv2d_processor
(self, rot_inv_mode='none', enable_cube_padding=True, impl='cuda')
models/multiplane_sync/sync_conv2d.py:248
↓ 5 callersMethodexpand_tensor_or_array
Expand a tensor or array to a specified number of images.
utils/depth.py:31
↓ 5 callersMethodnumpy_to_pt
Convert a NumPy image to a PyTorch tensor.
utils/depth.py:58
↓ 4 callersFunctioncreate_rotation_matrices
Create rotation matrices from batch of rotations This methods creates a bx3x3 np.ndarray where `b` referes to the number of rotations (rots)
external/equilib/numpy_utils/rotation.py:141
↓ 4 callersFunctioncreate_rotation_matrices
Create rotation matrices from batch of rotations This methods creates a bx3x3 np.ndarray where `b` referes to the number of rotations (rots)
external/equilib/torch_utils/rotation.py:134
↓ 4 callersMethodcube_all2all_equilib
(faces, cube_format, target_cube_format='horizon', to_equilib=False, from_equilib=False)
utils/cube.py:380
↓ 4 callersFunctionkernel
(s, a)
external/equilib/grid_sample/torch/bicubic.py:10
↓ 4 callersFunctionkernel
( s: np.ndarray, a: float = -0.75, dtype: np.dtype = np.dtype(np.float32) )
external/equilib/grid_sample/numpy/bicubic.py:8
↓ 4 callersFunctionnumpy_grid_sample
( img: np.ndarray, grid: np.ndarray, out: np.ndarray, cube_face_id: np.ndarray )
external/equilib/cube2equi/numpy.py:208
↓ 4 callersMethodsave
(self, path)
utils/equi.py:81
↓ 4 callersMethodto_equirectangular
(self, h, w, mode='bilinear', backend='equilib')
utils/cube.py:396
↓ 4 callersMethodto_perspective
Converts a cubemap to a perspective image using a specified field of view (FOV) and orientation angles. Args: - fov: float - Fiel
utils/cube.py:412
↓ 3 callersFunctioncalculate_edge_weights
(tensor: torch.Tensor, one_hot: bool = False, two_hot: bool = False)
models/multiplane_sync/sync_conv2d.py:86
↓ 3 callersFunctioncube_sync_attn_processor
(self)
models/multiplane_sync/processors_videocrafter2.py:19
↓ 3 callersFunctioncube_sync_gn_processor
(self)
models/multiplane_sync_legacy.py:212
↓ 3 callersFunctioncube_sync_gn_processor
(self)
models/multiplane_sync/sync_norm.py:7
↓ 3 callersFunctioncube_unit_rays
(h: int, w: int, device)
utils/pano_to_3d.py:33
↓ 3 callersMethodencode_cube_images
(self, images: torch.Tensor)
models/dreamcube.py:72
↓ 3 callersFunctionequi_unit_rays
(h: int, w: int, device)
utils/pano_to_3d.py:19
↓ 3 callersFunctionimages_to_equi_and_dice
Args: images: np.ndarray (B, 6, H, W, C) | torch.Tensor (B, 6, C, H, W) Returns: equis: np.ndarray, (B, H, W, C) dice
app.py:21
↓ 3 callersFunctioninterp
(v0, v1, d, L)
external/equilib/grid_sample/numpy/bilinear.py:8
↓ 3 callersFunctionlinear_interp
(v0, v1, d, L)
external/equilib/grid_sample/torch/bilinear.py:8
↓ 3 callersFunctionnamed_params_and_buffers
(module)
utils/ops/misc.py:150
↓ 3 callersFunctionrotation_invariant_kernels
参数: weight (torch.Tensor): 输入的卷积核权重,形状为 [C_in, C_out, K, K] 返回: torch.Tensor: 旋转后的卷积核权重,形状相同。
models/multiplane_sync/sync_conv2d.py:11
↓ 3 callersFunctionrotation_matrix
(rad, ax)
external/py360convert/utils.py:231
↓ 3 callersMethodto_perspective
(self, fov, yaw, pitch, hw, mode='bilinear')
utils/equi.py:61
↓ 2 callersFunction_resample_padding_cuda
Fast CUDA implementation of 'resample_padding()' using custom ops.
utils/ops/resample_padding.py:165
↓ 2 callersFunctionadaptive_convolution
进行基于 theta 角度自适应旋转的卷积
models/multiplane_sync_legacy.py:174
↓ 2 callersFunctionapply_homography_to_image_points
Maps all pixel coordinates in an image using a homography matrix. Parameters: - height: image height - width: image width - homo
utils/pers.py:9
↓ 2 callersFunctionconvert_grid
( M: torch.Tensor, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2pers/torch.py:74
↓ 2 callersFunctionconvert_grid
( M: np.ndarray, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2pers/numpy.py:80
↓ 2 callersFunctioncube_h2list
(cube_h)
external/py360convert/utils.py:176
↓ 2 callersFunctioncube_hsplits
Returns list of horizontal splits (doesn't split batch)
external/equilib/equi2cube/torch.py:16
↓ 2 callersFunctioncube_hsplits
Returns list of horizontal splits (doesn't split batch)
external/equilib/equi2cube/numpy.py:12
↓ 2 callersMethodfrom_cubediffusion
Args: faces: np.ndarray, (6, H, W, C), uint8 cube_format: str, 'list', 'dict', 'dice', 'horizon'
utils/cube.py:310
↓ 2 callersMethodfrom_mp3d_skybox
(cls, mp3d_skybox_path, scene, view, subfolder='matterport_skybox_images')
utils/cube.py:296
↓ 2 callersMethodfrom_perspective
Args: images: np.ndarray, (M, H, W, C), uint8 fovs: np.ndarray, (M,), float thetas: np.ndarray, (M,), flo
utils/cube.py:323
↓ 2 callersFunctiongenerate_K_R
(fov, theta, phi, height, width, order='yx')
utils/pers.py:73
↓ 2 callersFunctiongenerate_mask
Generates a mask indicating whether transformed points are within the image bounds. Parameters: - transformed_points: np.ndarray - T
utils/pers.py:39
↓ 2 callersFunctionget_edge
(face, edge)
utils/pano_to_3d.py:544
↓ 2 callersFunctionget_examples
(example_dir: str = './assets')
app.py:114
↓ 2 callersFunctionhorizon_sample_camera
(n)
utils/camera.py:22
↓ 2 callersFunctionicosahedron_sample_camera
()
utils/camera.py:28
↓ 2 callersFunctioninterp2d
(q00, q10, q01, q11, dy, dx)
external/equilib/grid_sample/torch/bilinear.py:12
↓ 2 callersFunctionlist2horizon
(lists: List[List[torch.Tensor]])
external/equilib/cube2equi/torch.py:55
↓ 2 callersMethodload_image_canonical
( image: Union[torch.Tensor, np.ndarray, Image.Image], device: torch.device = torch.device("cp
utils/depth.py:143
↓ 2 callersFunctionmatmul
(m: torch.Tensor, G: torch.Tensor, R: torch.Tensor)
external/equilib/equi2pers/torch.py:67
↓ 2 callersFunctionmatmul
( m: np.ndarray, G: np.ndarray, R: np.ndarray, method: str = "faster" )
external/equilib/equi2pers/numpy.py:53
↓ 2 callersFunctionnpers2pers
(images: List[np.ndarray], fovs, thetas, phis, fov2, theta2, phi2, height2, width2, overlap=False)
utils/pers.py:129
↓ 2 callersMethodpack
(self, inputs: torch.Tensor)
models/dreamcube.py:49
↓ 2 callersFunctionprep_matrices
( height: int, width: int, batch: int, fov_x: float, skew: float = 0.0, dtype: torch.d
external/equilib/equi2pers/torch.py:42
↓ 2 callersFunctionprep_matrices
( height: int, width: int, batch: int, fov_x: float, skew: float = 0.0, dtype: np.dtyp
external/equilib/equi2pers/numpy.py:36
↓ 2 callersFunctionrandom_sample_camera
(n)
utils/camera.py:15
↓ 2 callersMethodrelease
(self)
utils/video.py:145
↓ 2 callersFunctionrgb_to_sh
(rgb: torch.Tensor)
utils/pano_to_3d.py:13
↓ 2 callersFunctionsetup_sync_conv2d_processor
(module, **kwargs)
models/multiplane_sync_legacy.py:436
↓ 2 callersFunctionsingle_list2horizon
(cube: List[torch.Tensor])
external/equilib/cube2equi/torch.py:15
↓ 2 callersFunctionsingle_list2horizon
(cube: List[np.ndarray])
external/equilib/cube2equi/numpy.py:13
↓ 2 callersFunctionsuppress_tracer_warnings
()
utils/ops/misc.py:70
↓ 2 callersFunctionswitch_custom_processors_for_vae
(model, enable_sync_gn: bool, enable_sync_conv2d: bool, enable_sync_attn: bool)
models/multiplane_sync/processors_sd.py:15
↓ 2 callersMethodto_PIL_image
(self, cube_format='horizon')
utils/cube.py:518
↓ 2 callersMethodto_PIL_image
(self)
utils/equi.py:75
↓ 2 callersMethodunpack
(self, outputs: torch.Tensor, shape: torch.Size, dtype: torch.dtype)
models/dreamcube.py:56
↓ 2 callersMethodwrite
(self, image: Image.Image)
utils/video.py:142
↓ 1 callersFunction_equirect_facetype
0F 1R 2B 3L 4U 5D
external/equilib/cube2equi/torch.py:135
↓ 1 callersFunction_equirect_facetype
0F 1R 2B 3L 4U 5D
external/equilib/cube2equi/numpy.py:123
↓ 1 callersFunction_find_compiler_bindir
()
utils/ops/custom_ops.py:28
↓ 1 callersFunction_get_mangled_gpu_name
()
utils/ops/custom_ops.py:43
↓ 1 callersFunction_init
()
utils/ops/resample_padding.py:25
↓ 1 callersFunction_resample_padding_ref
(x, sampling_rate, backwards=False)
utils/ops/resample_padding.py:62
↓ 1 callersFunction_should_use_custom_op
()
utils/ops/grid_sample_gradfix.py:36
↓ 1 callersFunctionapply_custom_processors_for_unet
( model, enable_sync_self_attn: bool = True, enable_sync_cross_attn: bool = False, enable_sync
models/multiplane_sync_legacy.py:502
↓ 1 callersFunctionapply_custom_processors_for_vae
( model, mode: str = 'all', enable_sync_gn: bool = True, enable_sync_conv2d: bool = False,
models/multiplane_sync_legacy.py:540
↓ 1 callersMethodas_horizon
(self)
utils/cube.py:392
↓ 1 callersFunctionbicubic
( img: torch.Tensor, grid: torch.Tensor, out: torch.Tensor )
external/equilib/grid_sample/torch/bicubic.py:25
↓ 1 callersFunctionbicubic
Bicubic Interpolation
external/equilib/grid_sample/numpy/bicubic.py:25
↓ 1 callersFunctionbilinear
( img: torch.Tensor, grid: torch.Tensor, out: torch.Tensor )
external/equilib/grid_sample/torch/bilinear.py:18
↓ 1 callersFunctionbilinear
Bilinear Interpolation NOTE: asserts are removed
external/equilib/grid_sample/numpy/bilinear.py:18
↓ 1 callersFunctionbuild_pipeline
( ckpt_path: str = 'KevinHuang/DreamCube', device: Optional[torch.device] = None, local_files_only
app.py:143
↓ 1 callersFunctioncalculate_edge_weights
(tensor: torch.Tensor, one_hot: bool = False)
models/multiplane_sync_legacy.py:66
↓ 1 callersMethodcheck_image_values_range
(image: torch.Tensor)
utils/depth.py:189
↓ 1 callersMethodcolormap
Converts a monochrome image into an RGB image by applying the specified colormap. This function mimics the behavior of matplotlib.col
utils/depth.py:239
↓ 1 callersFunctioncompute_theta_map
计算每个像素点相对于中心的角度矩阵 Returns: theta: [H, W] 角度矩阵
models/multiplane_sync_legacy.py:94
↓ 1 callersFunctioncompute_theta_map
计算每个像素点相对于中心的角度矩阵 Returns: theta: [H, W] 角度矩阵
models/multiplane_sync/sync_conv2d.py:117
↓ 1 callersFunctionconvert_grid
( xyz: torch.Tensor, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2cube/torch.py:85
↓ 1 callersFunctionconvert_grid
( xyz: np.ndarray, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2cube/numpy.py:88
↓ 1 callersFunctionconvert_grid
( M: torch.Tensor, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2equi/torch.py:23
↓ 1 callersFunctionconvert_grid
( M: np.ndarray, h_equi: int, w_equi: int, method: str = "robust" )
external/equilib/equi2equi/numpy.py:35
next →1–100 of 331, ranked by callers