Types & classes — github.com/apple/ml-sharp

hub / github.com/apple/ml-sharp / types & classes

Types & classes48 in github.com/apple/ml-sharp

⨍ Functions 179◇ Types & classes 48 ↳ Endpoints 2

↓ 6 callersClassActivationPair

A pair of forward and inverse activation functions.

src/sharp/utils/math.py:25

↓ 4 callersClassGaussians3D

Represents a collection of 3D Gaussians.

src/sharp/utils/gaussians.py:26

↓ 2 callersClassFeatureFusionBlock2d

Feature fusion for DPT.

src/sharp/models/blocks.py:127

↓ 2 callersClassImageFeatures

Image feature extracted from decoder.

src/sharp/models/gaussian_decoder.py:86

↓ 2 callersClassMultiresConvDecoder

Decoder for multi-resolution encodings.

src/sharp/models/decoders/multires_conv_decoder.py:22

↓ 2 callersClassRenderingOutputs

Outputs of 3D Gaussians renderer.

src/sharp/utils/gsplat.py:21

↓ 2 callersClassResidualBlock

Generic implementation of residual blocks. This implements a generic residual block from He et al. - Identity Mappings in Deep Residual

src/sharp/models/blocks.py:49

↓ 2 callersClassSceneMetaData

Meta data about Gaussian scene.

src/sharp/utils/gaussians.py:46

↓ 2 callersClassSkipConvBackbone

A wrapper around a conv layer that behaves like a BaseBackbone.

src/sharp/models/gaussian_decoder.py:93

↓ 1 callersClassCameraInfo

Camera info for a pinhole camera.

src/sharp/utils/camera.py:23

↓ 1 callersClassDepthAlignment

Depth alignment in a dedicated nn.Module. Wrap scale_map_estimator to perform the conditional logic in a separated torch module outside the f

src/sharp/models/predictor.py:22

↓ 1 callersClassDirectPredictionHead

Decodes features into delta values using convolutions.

src/sharp/models/heads.py:15

↓ 1 callersClassFocusRange

Parametrizes a range of depth / disparity values.

src/sharp/utils/camera.py:32

↓ 1 callersClassGaussianBaseValues

Base values for gaussian predictor. We predict x and y in normalized device coordinates (NDC) where (-1, -1) is the top left corner and (1, 1

src/sharp/models/initializer.py:33

↓ 1 callersClassGaussianComposer

Converts base values and deltas into Gaussians.

src/sharp/models/composer.py:31

↓ 1 callersClassGaussianDensePredictionTransformer

Dense Prediction Transformer for Gaussian. Reuse monodepth decoded features for processing.

src/sharp/models/gaussian_decoder.py:125

↓ 1 callersClassInitializerOutput

Output of initializer.

src/sharp/models/initializer.py:51

↓ 1 callersClassLearnedAlignment

Aligns tensors using a UNet.

src/sharp/models/alignment.py:43

↓ 1 callersClassMonodepthDensePredictionTransformer

Dense Prediction Transformer for monodepth. Attach the disparity prediction head for monodepth prediction.

src/sharp/models/monodepth.py:30

↓ 1 callersClassMonodepthOutput

Output of the monodepth model.

src/sharp/models/monodepth.py:152

↓ 1 callersClassMonodepthParams

Parameters for monodepth network.

src/sharp/models/params.py:106

↓ 1 callersClassMonodepthWithEncodingAdaptor

Monodepth model with feature maps.

src/sharp/models/monodepth.py:167

↓ 1 callersClassMultiLayerInitializer

Initialize Gaussians with multilayer representation. The returned tensors have the shape batch_size x dim x num_layers x height x width

src/sharp/models/initializer.py:64

↓ 1 callersClassPinholeCameraModel

Camera model that focuses on point.

src/sharp/utils/camera.py:271

↓ 1 callersClassPredictorParams

Parameters for predictors with default values.

src/sharp/models/params.py:164

↓ 1 callersClassProjectionModule

Apply projection of features.

src/sharp/models/encoders/monodepth_encoder.py:62

↓ 1 callersClassRGBGaussianPredictor

Predicts 3D Gaussians from images.

src/sharp/models/predictor.py:64

↓ 1 callersClassSlidingPyramidNetwork

Sliding Pyramid Network. An encoder aimed at creating multi-resolution encodings from Vision Transformers. Reference: Bochkovskii et al. - "

src/sharp/models/encoders/spn_encoder.py:30

↓ 1 callersClassTimmViT

Contains TIMM implementation for Vanilla ViT.

src/sharp/models/encoders/vit_encoder.py:19

↓ 1 callersClassUNetDecoder

Decoder of UNet model.

src/sharp/models/decoders/unet_decoder.py:23

↓ 1 callersClassUNetEncoder

Encoder of UNet model.

src/sharp/models/encoders/unet_encoder.py:23

↓ 1 callersClassViTConfig

Configuration for ViT.

src/sharp/models/presets/vit.py:18

ClassAffineRangeNormalizer

Perform linear mapping to map input_range to output_range. Output_range defaults to (0, 1).

src/sharp/models/normalizers.py:41

ClassAlignmentParams

Parameters for depth alignment.

src/sharp/models/params.py:36

ClassBaseDecoder

Base decoder class.

src/sharp/models/decoders/base_decoder.py:14

ClassBaseEncoder

Base encoder class.

src/sharp/models/encoders/base_encoder.py:13

ClassClampWithPushback

Implementation of clamp_with_pushback function.

src/sharp/utils/math.py:106

ClassDeltaFactor

Factors to multiply deltas with before activation. These factors effectively selectively reduce the learning rate.

src/sharp/models/params.py:55

ClassGSplatRenderer

Module to render 3D Gaussians to images using gsplat.

src/sharp/utils/gsplat.py:48

ClassGaussianDecoderParams

Parameters for backbone with default values.

src/sharp/models/params.py:132

ClassInitializerParams

Parameters for initializer.

src/sharp/models/params.py:70

ClassMeanStdNormalizer

Normalizing image input by mean and std.

src/sharp/models/normalizers.py:15

ClassMobileNetNormalizer

Image normalization in mobilenet.

src/sharp/models/normalizers.py:75

ClassMonodepthAdaptorParams

Parameters for monodepth network feature adaptor.

src/sharp/models/params.py:124

ClassMonodepthFeatureEncoder

A wrapper around monodepth network to extract features.

src/sharp/models/encoders/monodepth_encoder.py:81

ClassOutputWriter

Protocol for writing output to disk.

src/sharp/utils/io.py:170

ClassTrajectoryParams

Parameters for trajectory.

src/sharp/utils/camera.py:41

ClassVideoWriter

Output writer for video output.

src/sharp/utils/io.py:182