There is a "gotcha" relating to scipy imports, that can cause an AttributeError under certain conditions. Scipy states in the top level package that: Using any of these subpackages requires an explicit import. For example, import scipy.cluster. However, in this repo scipy submodules are not consis...

shap\utils_clustering.py:35: NumbaDeprecationWarning: The 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is currently False, but it will be changed to True in Numba 0.59.0. See https://numba.readthedocs.io/en/stable/reference/d...

Copied from dsgibbons#98 We have seen GitHub Actions randomly fail occasionally due to floating-point precision issues. See, for example: https://github.com/dsgibbons/shap/actions/runs/5085485586/attempts/1 https://github.com/dsgibbons/shap/actions/runs/5135384415/attempts/1 In both cases, the tes...

Some PRs are unable to be merged at present: the python unit tests do not run unless the path filter conditions are met, but the PR cannot be completed as the tests are configured as required status checks in the repository settings. Example PR #2949 : This could be resolved by: Changing the branc...

This is a meta-issue to track progress towards merging the SHAP Community Fork into this upstream project. See discussion: #2942 TODO: Fix failing CodeQL checks: dsgibbons#104 Prepare branch that removes fork-specific changes: dsgibbons#103 Complete main outstanding PRs in dsgibbons/shap: dsgi...

TF version 2.8.2 SHAP version 0.41 I tried to apply SHAP for my binary classifier of deep residual shrinkage network (DRSN), from tensorflow.keras import backend as K def abs_backend(inputs): return K.abs(inputs) def expand_dim_backend(inputs): return K.expand_dims(K.expand_dims(inputs,1),1) def sig...

I am attempting to use the KernelExplainer to explain a function that uses 5 separate XGBoost models to determine a final value. The structure of this function looks like this: def model_func(data): if type(data)==np.ndarray: data = pd.DataFrame(data, columns=features).astype(float)...

X.shape, y.shape ((1341026, 81), (1341026,)) Model: "sequential" Layer (type) Output Shape Param # dense (Dense) (None, 78) 6396 dense_1 (Dense) (None, 39) 3081 dense_2 (Dense) (None, 19)...

Warning appear just by importing shap: import shap #> C:\Users\user\anaconda3\envs\env\Lib\site-packages\shap\utils\_clustering.py:34: NumbaDeprecationWarning: �[1mThe 'nopython' keyword argument was not supplied to the 'numba.jit' decorator. The implicit default value for this argument is curre...

Is there a way to set the base_values before calculating the shap values for a TreeExplainer in the shap package? I'm using a gpu xgbooster and I want to be able to set the base values. Thanks in advance!