MCPcopy Create free account

hub / github.com/UVA-Computer-Vision-Lab/OmniShotCut / types & classes

Types & classes16 in github.com/UVA-Computer-Vision-Lab/OmniShotCut

⨍ Functions 95◇ Types & classes 16
↓ 3 callersClassNestedTensor
omnishotcut/util/misc.py:31
↓ 3 callersClassVideo_Augmentation_Transform
Clip-wise (video-constant) augmentation. All randomness sampled ONCE per call, applied identically to every frame.
omnishotcut/datasets/transforms.py:36
↓ 2 callersClassOmniShotCut
This is the OmniShotCut module that performs object detection
omnishotcut/architecture/model.py:36
↓ 1 callersClassBackbone
ResNet backbone with frozen BatchNorm.
omnishotcut/architecture/backbone.py:87
↓ 1 callersClassCutAnything_Dataloader
omnishotcut/datasets/dataloader.py:68
↓ 1 callersClassJoiner
omnishotcut/architecture/backbone.py:100
↓ 1 callersClassOmniShotCutModel
omnishotcut/__init__.py:21
↓ 1 callersClassPositionEmbeddingSine
This is a more standard version of the position embedding, very similar to the one used by the Attention is all you need paper, generalized t
omnishotcut/architecture/position_encoding.py:12
↓ 1 callersClassTransformer
omnishotcut/architecture/transformer.py:35
↓ 1 callersClassTransformerDecoder
omnishotcut/architecture/transformer.py:110
↓ 1 callersClassTransformerDecoderLayer
omnishotcut/architecture/transformer.py:213
↓ 1 callersClassTransformerEncoder
omnishotcut/architecture/transformer.py:85
↓ 1 callersClassTransformerEncoderLayer
omnishotcut/architecture/transformer.py:153
ClassBackboneBase
omnishotcut/architecture/backbone.py:58
ClassFrozenBatchNorm2d
BatchNorm2d where the batch statistics and the affine parameters are fixed. Copy-paste from torchvision.misc.ops with added eps before rqsrt
omnishotcut/architecture/backbone.py:19
ClassMLP
Very simple multi-layer perceptron (also called FFN)
omnishotcut/architecture/model.py:19