你能不能提示我，在SHAP库中的
shap.GradientExplainer()
要放什么参数（使用Tensorflow 2.0）。
There is the only 例子 in internet linked with images classification. In the 例子, the function accepts 2 inputs as the 1st argument:
model.layers[7].input, model.layers[-1].output
The full code snippet:
# explain how the input to the 7th layer of the model explains the top two classes
def map2layer(x, layer):
feed_dict = dict(zip([model.layers[0].input], [preprocess_input(x.copy())]))
return K.get_session().run(model.layers[layer].input, feed_dict)
e = shap.GradientExplainer((model.layers[7].input, model.layers[-1].output), map2layer(preprocess_input(X.copy()), 7))
shap_values,indexes = e.shap_values(map2layer(to_explain, 7), ranked_outputs=2)
但我的模型是顺序式的see here.
当我试图把2个输入作为第一个参数时
shap_values = explainer.shap_values((model.input, model.output))
，发生错误
Layer sequential expects 1 inputs, but it received 2 input tensors.
When I use 1 input
shap_values = explainer.shap_values(model.output)
，又出现了一个错误。
Input 0 of layer sequential is incompatible with the layer: expected axis -1 of input shape to have value 46 but received input with shape `[None, 1]`
然后，我尝试从数据集中输入数据。
shap_values = explainer.shap_values(df3.to_numpy()[:100,:])