netmap.model.zinbautoencoder.get_thetas_grouping¶
- netmap.model.zinbautoencoder.get_thetas_grouping(model, data_tensor, grouping)[source]¶
Return the mean predicted theta per observation group.
- Parameters:
model (ZINBAutoencoder) – A trained ZINB autoencoder.
data_tensor (torch.Tensor) – Input data on CPU; moved to CUDA internally.
grouping (array-like) – Group label per cell, length
n_cells.
- Returns:
Mapping
{group_label: numpy.ndarray}of mean theta per gene.- Return type: