is_layer_at_idx not implemented

[adrhill]

The function is_layer_at_idx currently doesn't return anything when it should be returning a boolean value after checking the layer:

innvestigate/src/innvestigate/utils/keras/checks.py

Lines 446 to 448 in c6ce893

	def is_layer_at_idx(layer, index, ignore_reshape_layers=True):
	"""Checks if layer is a layer at index index, by repeatedly applying is_input_layer()."""
	kgraph = get_kgraph()

This is used in all analyzers that inherit from LRP and use the kwargs until_layer_rule and until_layer_idx:

innvestigate/src/innvestigate/analyzer/relevance_based/relevance_analyzer.py

Lines 399 to 410 in c3295be

	# apply rule to first self._until_layer_idx layers
	if self._until_layer_rule is not None and self._until_layer_idx is not None:
	for i in range(self._until_layer_idx + 1):
	rules.insert(
	0,
	(
	lambda layer, foo, bound_i=i: kchecks.is_layer_at_idx(
	layer, bound_i
	),
	self._until_layer_rule,
	),
	)

As Python evaluates bool(None) as False instead of throwing an error, the until_layer_rule never gets applied.

Currently only LRPSequentialPresetBFlatUntilIdx and user-defined composite LRP methods are affected, so this bug should be fixed before merging #261.
Until then PR #263 adds a NotImplementedError.

[HugoTex98]

Hi @adrhill !

Given the example in https://github.com/adrhill/innvestigate/blob/adrhill/custom-composite/examples/notebooks/imagenet_composite_lrp.ipynb, without the application of until_layer_rule , the only rule being applied to the model is the one specified in rule?

[adrhill]

Hi HugoTex98,

that notebook is work in progress from PR #261 , which hasn't been merged yet due to this issue.

That being said, you can still use LRP without the until_layer_rule:

analyzer = LRP(
    model_wo_softmax, 
    rule="Z",
    input_layer_rule="Bounded",
    bn_layer_rule="Z"
)

[HugoTex98]

Thanks @adrhill!

Btw, do you have any idea of a possible release date for this LRP with the until_layer_rule?

[adrhill]

It's high up on my to-do list, but it will have to wait until the TF2 release, which should be ready soon.