sleap.gui.learning.datagen¶
Preview of training data (i.e., confidence maps and part affinity fields).
-
sleap.gui.learning.datagen.
make_datagen_results
(reader: sleap.nn.data.providers.LabelsReader, cfg: sleap.nn.config.training_job.TrainingJobConfig) → numpy.ndarray[source]¶ Gets (subset of) raw images used for training.
- TODO: Refactor so we can get this data without digging into details of the
the specific pipelines (e.g., key for confmaps depends on head type).
-
sleap.gui.learning.datagen.
show_datagen_preview
(labels: sleap.io.dataset.Labels, config_info_list: List[sleap.gui.learning.configs.ConfigFileInfo])[source]¶ Shows window(s) with preview images of training data for model configs.