MCPcopy Create free account

hub / github.com/devinli123/MV-SAM3D / functions

Functions930 in github.com/devinli123/MV-SAM3D

⨍ Functions 930◇ Types & classes 130
↓ 148 callersMethodreshape
(self, *shape)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:265
↓ 135 callersMethodcpu
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:236
↓ 124 callersMethodto
(self, dtype: torch.dtype)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:202
↓ 58 callersMethodreplace
( self, feats: torch.Tensor, coords: Optional[torch.Tensor] = None )
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:272
↓ 56 callersMethodfloat
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:250
↓ 48 callersMethoddetach
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:254
↓ 32 callersMethoddim
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:158
↓ 30 callersMethodunbind
(self, dim: int)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:269
↓ 28 callersFunctiontree_tensor_map
(fn, tree, *rest)
sam3d_objects/data/utils.py:58
↓ 26 callersMethodtype
(self, dtype)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:232
↓ 25 callersMethodcuda
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:241
↓ 20 callersFunctionlog_analysis
(msg)
run_inference_weighted.py:3347
↓ 15 callersMethodset_view
Set current view being processed.
sam3d_objects/pipeline/multi_view_weighted.py:61
↓ 14 callersMethodwrite
(self, x)
sam3d_objects/model/backbone/tdfy_dit/representations/gaussian/general_utils.py:130
↓ 13 callersMethodnormalize
(self, value: torch.Tensor)
sam3d_objects/data/dataset/tdfy/pose_target.py:780
↓ 12 callersMethodclose
(self)
sam3d_objects/utils/cross_attention_logger.py:101
↓ 12 callersMethodload_state_dict
(self, state_dict)
sam3d_objects/model/backbone/tdfy_dit/representations/octree/octree_dfs.py:270
↓ 11 callersMethodfrom_instance_pose
(cls, instance_pose: InstancePose)
sam3d_objects/data/dataset/tdfy/pose_target.py:697
↓ 11 callersMethodto_instance_pose
(cls, pose_target: PoseTarget)
sam3d_objects/data/dataset/tdfy/pose_target.py:709
↓ 9 callersMethodset_num_views
(self, num_views: int)
sam3d_objects/utils/cross_attention_logger.py:95
↓ 9 callersMethodstart_stage
(self, stage: str)
sam3d_objects/utils/cross_attention_logger.py:83
↓ 8 callersMethod__elemwise__
( self, other: Union[torch.Tensor, "SparseTensor"], op: callable )
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:349
↓ 8 callersMethodpreprocess_image
( self, image: Union[Image.Image, np.ndarray], preprocessor )
sam3d_objects/pipeline/inference_pipeline.py:912
↓ 7 callersFunction_apply_metric_to_ssi
Normalize pointmap using Scale-Shift Invariant (SSI) normalization. Args: pointmap: Pointmap tensor of shape (H, W, 3) or (3, H,
sam3d_objects/data/dataset/tdfy/img_and_mask_transforms.py:780
↓ 7 callersMethodget_spatial_cache
Get a spatial cache.
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:445
↓ 7 callersMethodinstantiate_and_load_from_pretrained
( self, config, ckpt_path, state_dict_fn=None, state_dict_key="state_d
sam3d_objects/pipeline/inference_pipeline.py:261
↓ 7 callersMethodregister_spatial_cache
Register a spatial cache. The spatial cache can be any thing you want to cache. The registery and retrieval of the cache is b
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:434
↓ 6 callersMethod__init__
( self, channels: int, out_channels: Optional[int] = None, norm_type: Literal[
sam3d_objects/model/backbone/tdfy_dit/models/sparse_structure_vae.py:28
↓ 6 callersMethod_apply_transform
(self, input: torch.Tensor, transform)
sam3d_objects/data/dataset/tdfy/preprocessor.py:199
↓ 6 callersMethodfull
(aabb, dim, value, dtype=torch.float32, device=None)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:314
↓ 6 callersMethodinverse
Apply inverse remapping to recover original point coordinates.
sam3d_objects/model/backbone/dit/embedder/point_remapper.py:50
↓ 6 callersFunctionlinear_approximation_step
(x_t, dt, velocity)
sam3d_objects/model/backbone/generator/flow_matching/solver.py:9
↓ 6 callersFunctionscaled_dot_product_attention
Apply scaled dot product attention. Args: qkv (torch.Tensor): A [N, L, 3, H, C] tensor containing Qs, Ks, and Vs.
sam3d_objects/model/backbone/tdfy_dit/modules/attention/full_attn.py:42
↓ 6 callersMethodstep
(self, dynamics_fn, x_t, t, dt, *args, **kwargs)
sam3d_objects/model/backbone/generator/flow_matching/solver.py:60
↓ 6 callersFunctionweighted_fusion_sparse
Perform weighted fusion of sparse predictions. Args: predictions: List of [B, L_latent, C] or [L_latent, C] tensors weig
sam3d_objects/pipeline/multi_view_weighted.py:767
↓ 5 callersMethod_down_sample_img
(img_3chw: torch.Tensor)
sam3d_objects/pipeline/inference_pipeline_pointmap.py:481
↓ 5 callersMethod_moduledict_to_dict
(self, module)
sam3d_objects/model/backbone/tdfy_dit/modules/transformer/modulated.py:280
↓ 5 callersFunctioncompose_transform
Args: scale: (..., 3) tensor of scale factors rotation: (..., 3, 3) tensor of rotation matrices translation: (..., 3) ten
sam3d_objects/data/dataset/tdfy/transforms_3d.py:20
↓ 5 callersMethodconcatenate_tensor
(self, tensor_dict, latent_names)
sam3d_objects/model/backbone/tdfy_dit/modules/attention/modules.py:291
↓ 5 callersMethoddevice
(self)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:198
↓ 5 callersFunctionget_child
(obj: Any, *keys: Iterable[Any])
sam3d_objects/data/utils.py:110
↓ 5 callersMethodssi_to_metric
(scale: torch.Tensor, shift: torch.Tensor)
sam3d_objects/data/dataset/tdfy/pose_target.py:375
↓ 4 callersMethod_apply_transform
(self, input: torch.Tensor, transform)
sam3d_objects/pipeline/inference_pipeline.py:888
↓ 4 callersMethod_broadcast_postcompose
Assumes scale, rotation, translation are of shape: B, K, C --- B: batch size K: number of obj
sam3d_objects/data/dataset/tdfy/pose_target.py:26
↓ 4 callersMethod_generate_noise
(self, x_shape, x_device)
sam3d_objects/model/backbone/generator/flow_matching/model.py:99
↓ 4 callersMethod_linear
( module: nn.Linear, x: Union[SparseTensor, torch.Tensor] )
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/attention/modules.py:92
↓ 4 callersMethod_moduledict_to_dict
(self, module)
sam3d_objects/model/backbone/tdfy_dit/modules/attention/modules.py:275
↓ 4 callersFunctioncompile_wrapper
( fn, *, mode="max-autotune", fullgraph=True, dynamic=False, name=None )
sam3d_objects/pipeline/inference_pipeline_pointmap.py:73
↓ 4 callersMethodcompute_pointmap
(self, image, pointmap=None)
sam3d_objects/pipeline/inference_pipeline_pointmap.py:242
↓ 4 callersFunctionget_strength
(strength, interval, t)
sam3d_objects/model/backbone/generator/classifier_free_guidance.py:151
↓ 4 callersFunctionimshow
(image, ax)
notebook/inference.py:382
↓ 4 callersMethodis_mm_dit
(self, model_name="ss_generator")
sam3d_objects/pipeline/inference_pipeline.py:734
↓ 4 callersMethodmerge_image_and_mask
Merge mask into image's alpha channel (RGBA), Keep single-view path consistent with multi-view processing to avoid mask issues due to
sam3d_objects/pipeline/inference_pipeline.py:664
↓ 4 callersFunctionnorm_layer
Return a normalization layer.
sam3d_objects/model/backbone/tdfy_dit/models/sparse_structure_vae.py:15
↓ 4 callersFunctionpose_decoder
( pose_target_convention, )
sam3d_objects/pipeline/inference_utils.py:224
↓ 4 callersMethodrun
Parameters: - image (Image): The input image to be processed. - seed (int, optional): The random seed for reproducibility. De
sam3d_objects/pipeline/inference_pipeline.py:551
↓ 4 callersMethodsample_sparse_structure
( self, ss_input_dict: dict, inference_steps=None, use_distillation=False, attention_logger: Optional[
sam3d_objects/pipeline/inference_pipeline.py:759
↓ 4 callersMethodsave_ply
(self, path)
sam3d_objects/model/backbone/tdfy_dit/representations/gaussian/gaussian_model.py:136
↓ 4 callersMethodseed
(self)
sam3d_objects/model/backbone/generator/base.py:22
↓ 4 callersMethodsolve_iter
(self, dynamics_fn, x_init, times, *args, **kwargs)
sam3d_objects/model/backbone/generator/flow_matching/solver.py:33
↓ 3 callersMethod__cal_layout
(self, coords, batch_size)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:145
↓ 3 callersMethod__cal_shape
(self, feats, coords)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/basic.py:139
↓ 3 callersMethod__init__
(self, dim: int, heads: int)
sam3d_objects/model/backbone/tdfy_dit/modules/attention/modules.py:15
↓ 3 callersMethod__init__
(self, channels: int, mlp_ratio: float = 4.0)
sam3d_objects/model/backbone/tdfy_dit/modules/transformer/blocks.py:60
↓ 3 callersMethod_linear_interp
Computes the location of zero-crossings on 'edges_x' using linear interpolation with 'edges_weight'.
sam3d_objects/model/backbone/tdfy_dit/representations/mesh/flexicubes/flexicubes.py:224
↓ 3 callersMethod_prepare_t
(self, steps=None)
sam3d_objects/model/backbone/generator/flow_matching/model.py:190
↓ 3 callersFunction_to_np
(x)
run_inference_weighted.py:2839
↓ 3 callersMethod_tree_flatten
(self, tree)
sam3d_objects/model/backbone/generator/flow_matching/model.py:291
↓ 3 callersMethodadd_view_attention
Add attention weights for a view. Args: view_idx: View index attention: [B, L_latent, L_cond] attent
sam3d_objects/utils/latent_weighting.py:510
↓ 3 callersFunctionas_keys
(path: ChildPathType)
sam3d_objects/data/utils.py:102
↓ 3 callersMethodcollect
Collect attention for the current view. Only collects the FIRST attention for each view (cond branch). Skips subsequ
sam3d_objects/pipeline/multi_view_weighted.py:65
↓ 3 callersFunctioncompute_iou
(render_mask_obj, mask_obj_gt, threshold=0.5)
sam3d_objects/pipeline/layout_post_optimization_utils.py:329
↓ 3 callersMethodcompute_weights
Compute final fusion weights based on weight_source config. Weight sources: - "entropy": w_v = softmax(-alpha_e * en
sam3d_objects/utils/latent_weighting.py:552
↓ 3 callersMethodconvert_to_fp16
Convert the torso of the model to float16.
sam3d_objects/model/backbone/tdfy_dit/models/structured_latent_vae/base.py:105
↓ 3 callersMethoddecode_slat
Decode the structured latent. Args: slat (sp.SparseTensor): The structured latent. formats (List[str]): The
sam3d_objects/pipeline/inference_pipeline.py:706
↓ 3 callersFunctionfilter_and_remove_prefix_state_dict_fn
(prefix: str)
sam3d_objects/model/io.py:104
↓ 3 callersMethodforward
(self, input: SparseTensor)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/norm.py:19
↓ 3 callersFunctionfull_class_name
(klass: Any)
sam3d_objects/config/utils.py:64
↓ 3 callersMethodget_multi_view_condition_input
Prepare conditions for multi-view input Args: condition_embedder: Condition embedder view_input_dict
sam3d_objects/pipeline/inference_pipeline.py:964
↓ 3 callersMethodimage_to_float
(self, image)
sam3d_objects/pipeline/inference_pipeline.py:906
↓ 3 callersFunctionmake_hook
(layer_idx)
sam3d_objects/pipeline/multi_view_weighted.py:389
↓ 3 callersFunctionmasked_sdpa
Mimic xFormers' memory_efficient_attention using PyTorch 2.0 scaled_dot_product_attention.
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/attention/masked_sdpa.py:30
↓ 3 callersMethodobject_pointcloud
Args: points_local: (N, 3) tensor of point coordinates colors: (N, 3) tensor of colors
sam3d_objects/utils/visualization/scene_visualizer.py:169
↓ 3 callersMethodpostprocess_slat_output
( self, outputs, with_mesh_postprocess, with_texture_baking, use_vertex_color )
sam3d_objects/pipeline/inference_pipeline.py:629
↓ 3 callersMethodpreprocess_image
( self, image: Union[Image.Image, np.ndarray], preprocessor, pointmap=None,
sam3d_objects/pipeline/inference_pipeline_pointmap.py:172
↓ 3 callersFunctionrender_frames
( sample, extrinsics, intrinsics, options={}, colors_overwrite=None, verbose=True,
sam3d_objects/model/backbone/tdfy_dit/utils/render_utils.py:53
↓ 3 callersMethodsample_slat
( self, slat_input: dict, coords: torch.Tensor, inference_steps=25, us
sam3d_objects/pipeline/inference_pipeline.py:838
↓ 3 callersFunctiontensor_to_o3d_pcd
(tensor)
sam3d_objects/pipeline/layout_post_optimization_utils.py:372
↓ 3 callersFunctionto_numpy
(x)
run_inference_weighted.py:1223
↓ 3 callersFunctiontree_reduce_unique
(fn, tree, ensure_unique=True, **kwargs)
sam3d_objects/data/utils.py:233
↓ 3 callersFunctionyaw_pitch_r_fov_to_extrinsics_intrinsics
(yaws, pitchs, rs, fovs)
sam3d_objects/model/backbone/tdfy_dit/utils/render_utils.py:14
↓ 3 callersFunctionzero_module
Zero out the parameters of a module and return it.
sam3d_objects/model/backbone/tdfy_dit/modules/utils.py:53
↓ 2 callersMethod__init__
(self, *args, strength_pm=0.0, **kwargs)
sam3d_objects/model/backbone/generator/classifier_free_guidance.py:170
↓ 2 callersMethod__init__
(self, factor: Union[int, Tuple[int, int, int], List[int]])
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/spatial.py:74
↓ 2 callersMethod__init__
(self, channels: int, mlp_ratio: float = 4.0)
sam3d_objects/model/backbone/tdfy_dit/modules/sparse/transformer/blocks.py:13
↓ 2 callersMethod__init__
( self, channels: int, num_heads: int, mlp_ratio: float = 4.0, attn_mo
sam3d_objects/model/backbone/tdfy_dit/modules/transformer/modulated.py:17
↓ 2 callersMethod__init__
(self, hidden_size, frequency_embedding_size=256)
sam3d_objects/model/backbone/tdfy_dit/models/sparse_structure_flow.py:26
↓ 2 callersMethod__init__
( self, resolution: int, in_channels: int, model_channels: int, cond_c
sam3d_objects/model/backbone/tdfy_dit/models/structured_latent_flow.py:78
↓ 2 callersMethod__init__
( self, resolution: int, model_channels: int, latent_channels: int, nu
sam3d_objects/model/backbone/tdfy_dit/models/structured_latent_vae/decoder_mesh.py:87
↓ 2 callersFunction_compute_dense_attention
Compute dense attention weights.
sam3d_objects/pipeline/multi_view_weighted.py:732
next →1–100 of 930, ranked by callers