lightning_pose.models

lightning_pose.models Package

Pose estimation model classes, re-exported at the package level.

Functions

check_if_semi_supervised([losses_to_use])	Determine from the losses config whether the model is semi-supervised.
get_model(cfg, data_module, loss_factories)	Create model: regression or heatmap based, supervised or semi-supervised.

Classes

HeatmapTracker	Base model that produces heatmaps of keypoints from images.
SemiSupervisedHeatmapTracker	Model produces heatmaps of keypoints from labeled/unlabeled images.
HeatmapTrackerMHCRNN	Multi-headed Convolutional RNN network that handles context frames.
SemiSupervisedHeatmapTrackerMHCRNN	Model produces heatmaps of keypoints from labeled/unlabeled images.
HeatmapTrackerMultiviewTransformer	Transformer network that handles multi-view datasets.
SemiSupervisedHeatmapTrackerMultiviewTransformer	Semi-supervised HeatmapTrackerMultiviewTransformer that supports unsupervised losses.
RegressionTracker	Base model that produces (x, y) predictions of keypoints from images.
SemiSupervisedRegressionTracker	Model produces vectors of keypoints from labeled/unlabeled images.

lightning_pose.models.base Module

Base class for backbone that acts as a feature extractor.

Functions

check_if_semi_supervised([losses_to_use])	Determine from the losses config whether the model is semi-supervised.
get_context_from_sequence(img_seq, ...)	Build overlapping context windows from a sequence of frames or feature maps.

Classes

BaseFeatureExtractor	Object that contains the base resnet feature extractor.
BaseSupervisedTracker	Base class for supervised trackers.
SemiSupervisedTrackerMixin	Mixin class providing training step function for semi-supervised models.

lightning_pose.models.factory Module

Factory function for creating pose estimation models from config.

Functions

get_model(cfg, data_module, loss_factories)

Create model: regression or heatmap based, supervised or semi-supervised.

Subpackages

• lightning_pose.models.heads