Docs: Clarify "ignore bias" in Composite How-To

chr5tphr / zennit

Zennit is a high-level framework in Python using PyTorch for explaining/exploring neural networks using attribution methods like LRP.

Other

188 stars 32 forks source link

Closed jackmcrider closed 2 years ago

jackmcrider commented 2 years ago

avoid ambiguity by explicitly stating "setting bias to zero" instead of "ignore bias"