netmap.model.nbautoencoder.get_thetas¶

netmap.model.nbautoencoder.get_thetas(model, data_tensor)[source]¶

Return the mean predicted dispersion (theta) across all cells.

Parameters:

model (NegativeBinomialAutoencoder) – A trained NB autoencoder.
data_tensor (torch.Tensor) – Input data on CPU; moved to CUDA internally.

Returns:

Mean theta value per gene, shape (input_dim,).

Return type: