MCPcopy Create free account

hub / github.com/Lu-Feng/DHE-VPR / types & classes

Types & classes25 in github.com/Lu-Feng/DHE-VPR

⨍ Functions 131◇ Types & classes 25
↓ 2 callersClassDropPath
Obtained from: github.com:rwightman/pytorch-image-models Drop paths (Stochastic Depth) per sample (when applied in main path of residual blo
model/cct/stochastic_depth.py:28
↓ 1 callersClassAttention
Obtained from timm: github.com:rwightman/pytorch-image-models
model/cct/transformers.py:7
↓ 1 callersClassCCT
model/cct/cct.py:32
↓ 1 callersClassFlatten
network.py:39
↓ 1 callersClassGeM
network.py:26
↓ 1 callersClassL2Norm
network.py:46
↓ 1 callersClassMaskedAttention
model/cct/transformers.py:38
↓ 1 callersClassMaskedTransformerEncoderLayer
Inspired by torch.nn.TransformerEncoderLayer and timm.
model/cct/transformers.py:103
↓ 1 callersClassRAMEfficient2DMatrix
This class behaves similarly to a numpy.ndarray initialized with np.zeros(), but is implemented to save RAM when the rows within the 2D array
datasets_ws.py:359
↓ 1 callersClassRAMEfficient4DMatrix
This class behaves similarly to a numpy.ndarray initialized with np.zeros(), but is implemented to save RAM when the rows within the 3D array
datasets_ws.py:380
↓ 1 callersClassTokenizer
model/cct/tokenizer.py:6
↓ 1 callersClassTransformerClassifier
model/cct/transformers.py:133
↓ 1 callersClassTransformerEncoderLayer
Inspired by torch.nn.TransformerEncoderLayer and timm.
model/cct/transformers.py:73
ClassBaseDataset
Dataset with images from database and queries, used for inference (testing and building cache).
datasets_ws.py:63
ClassDatasetQP
dataset_qp.py:8
ClassEmbedder
model/cct/embedder.py:4
ClassFeaturesExtractor
The FeaturesExtractor is composed of two parts: the backbone encoder and the pooling/aggregation layer. The pooling/aggregation layer is used
network.py:53
ClassGeolocDataset
dataset_geoloc.py:12
ClassHomographyRegression
network.py:86
ClassInfiniteDataLoader
commons.py:38
ClassMaskedTransformerClassifier
model/cct/transformers.py:231
ClassNetwork
Overview of the network: name input output FeaturesExtractor: (2B x 3 x H x W)
network.py:119
ClassPCADataset
datasets_ws.py:51
ClassTextTokenizer
model/cct/tokenizer.py:52
ClassTripletsDataset
Dataset used for training, it is used to compute the triplets with TripletsDataset.compute_triplets() with various mining methods. If is_infe
datasets_ws.py:143