Functions — github.com/Pointcept/PointTransformerV3

CodeHub MCPcopy Create free account

hub / github.com/Pointcept/PointTransformerV3 / functions

Functions50 in github.com/Pointcept/PointTransformerV3

⨍ Functions 50 ◇ Types & classes 13

↓ 13 callersMethodadd

(self, module, name=None)

↓ 11 callersMethod__init__

(self, patch_size, num_heads)

↓ 3 callersFunctionoffset2bincount

↓ 3 callersMethodxyz2key

(self, x, y, z, depth)

serialization/z_order.py:40

↓ 2 callersFunctionhilbert_encode

(grid_coord: torch.Tensor, depth: int = 16)

serialization/default.py:54

↓ 2 callersFunctionright_shift

Right shift an array of binary values. Parameters: ----------- binary: An ndarray of binary values. k: The number of bits to

serialization/hilbert.py:12

↓ 2 callersMethodsparsify

Point Cloud Serialization Point cloud is sparse, here we use "sparsify" to specifically refer to preparing "spconv.Spars

↓ 2 callersFunctionz_order_encode

(grid_coord: torch.Tensor, depth: int = 16)

serialization/default.py:41

↓ 1 callersFunctionbatch2offset

↓ 1 callersFunctionbinary2gray

Convert an array of binary values into Gray codes. This uses the classic X ^ (X >> 1) trick to compute the Gray code. Parameters: -

serialization/hilbert.py:46

↓ 1 callersMethoddecode_lut

(self, device=torch.device("cpu"))

serialization/z_order.py:34

↓ 1 callersFunctionencode

(grid_coord, batch=None, depth=16, order="z")

serialization/default.py:9

↓ 1 callersMethodencode_lut

(self, device=torch.device("cpu"))

serialization/z_order.py:28

↓ 1 callersMethodget_padding_and_inverse

↓ 1 callersMethodget_rel_pos

(self, point, order)

↓ 1 callersFunctiongray2binary

Convert an array of Gray codes back into binary values. Parameters: ----------- gray: An ndarray of gray codes. axis: The ax

serialization/hilbert.py:69

↓ 1 callersFunctionhilbert_decode

(code: torch.Tensor, depth: int = 16)

serialization/default.py:58

↓ 1 callersMethodkey2xyz

(self, key, depth)

serialization/z_order.py:52

↓ 1 callersFunctionoffset2batch

↓ 1 callersMethodserialization

Point Cloud Serialization relay on ["grid_coord" or "coord" + "grid_size", "batch", "feat"]

↓ 1 callersFunctionz_order_decode

(code: torch.Tensor, depth)

serialization/default.py:48

Method__getitem__

(self, *args, **kwargs)

(self, *args, **kwargs)

(self, *args, **kwargs)

( self, num_features, norm_layer, context_channels=256, condition

( self, channels, num_heads, patch_size, qkv_bias=True,

( self, in_channels, hidden_channels=None, out_channels=None, act

( self, channels, num_heads, patch_size=48, mlp_ratio=4.0,

( self, in_channels, out_channels, stride=2, norm_layer=None,

( self, in_channels, skip_channels, out_channels, norm_layer=None

( self, in_channels, embed_channels, norm_layer=None, act_layer=N

( self, in_channels=6, order=("z", "z-trans", "hilbert", "hilbert-trans"),

serialization/z_order.py:13

(code, depth=16, order="z")

serialization/default.py:28

Decode an array of Hilbert integers into locations in a hypercube. This is a vectorized-ish version of the Hilbert curve implementation by John

serialization/hilbert.py:201

Decode an array of locations in a hypercube into a Hilbert integer. This is a vectorized-ish version of the Hilbert curve implementation by Joh

serialization/hilbert.py:91

(self, point: Point)

(self, point: Point)

(self, point: Point)

A data_dict is a dictionary containing properties of a batched point cloud. It should contain the following properties for PTv3:

Functionkey2xyz

r"""Decodes the shuffled key to :attr:`x`, :attr:`y`, :attr:`z` coordinates and the batch index based on pre-computed look up tables. Args

serialization/z_order.py:104

Functionxyz2key

r"""Encodes :attr:`x`, :attr:`y`, :attr:`z` coordinates to the shuffled keys based on pre-computed look up tables. The speed of this function is

serialization/z_order.py:66