seqgra.evaluator.explainer.feedback module¶

class FeedbackExplainer(model, input_size, class_num=1000, lr=0.1, lambd=0.01, max_iters=30)[source]¶

Bases: seqgra.evaluator.explainer.backprop.SaliencyExplainer

explain(inp, ind=None)[source]¶

first_forward(self, x)[source]¶

mask_forward(self, x)[source]¶

replace_first(m)[source]¶

replace_mask(m)[source]¶