SHAP explainer for a sequential model with 3D input and 3D output

Question

I have a sequential binary classification model in Keras (Tensorflow 2.2.0) which takes the input of shape (n_samples, n_timesteps, n_features) and provides the output of sh