scanvi_explainer.scanvi_deep.SCANVIDeep

scanvi_explainer.scanvi_deep.SCANVIDeep#

class scanvi_explainer.scanvi_deep.SCANVIDeep(model: SCANVI, train_size: float = 0.8, batch_size: int = 128)#

SCANVIDeep is an extension of DeepExplainer [Lundberg and Lee, 2017] for models trained using SCANVI [Xu et al., 2021].

Parameters:

Explainer – Main Explainer class from shap package

__init__(model: SCANVI, train_size: float = 0.8, batch_size: int = 128)#

Constructor setting up expected values.

Currently categorical not continuous covariates are not supported.

Parameters:

  • model – Trained SCANVI model

  • train_size – float Training size (background), by default 0.8

  • batch_size – int Number of cells used from each group, by default 128 To ignore the batch_size subsetting, set batch_size=-1

Methods

__init__(model[, train_size, batch_size])

Constructor setting up expected values.

add_handles(model, forward_handle, ...)

Add handles to all non-container layers in the model.

add_target_handle(layer)

explain_row(*row_args, max_evals, ...)

Explains a single row and returns the tuple (row_values, row_expected_values, row_mask_shapes, main_effects).

get_train_test()

gradient(idx, input_x, input_batch)

load(in_file[, model_loader, masker_loader, ...])

Load an Explainer from the given file stream.

memory_stats()

Helper function to track CUDA memory usage.

remove_attributes(model)

Removes the x and y attributes which were added by the forward handles Recursively searches for non-container layers

save(out_file[, model_saver, masker_saver])

Write the explainer to the given file stream.

shap_values([with_labels])

Estimate SHAP values

supports_model_with_masker(model, masker)

Determines if this explainer can handle the given model.
Contents