Tobias Pitters
Tobias Pitters
This should be fixed for some time. Can anyone give any feedback if this happens with the latest shap version? No reproducible example is provided here.
hey, thanks for opening the issue. Can you pls provide a minimal example to reproduce the issue? We would need a dummy dataset and some code that fits the model...
related to #3419
The tolerance that is acceptable is 0.01, so your example lies within that. See [here](https://github.com/shap/shap/blob/master/shap/explainers/_deep/deep_utils.py#L5)
@karllandheer any objections to closing this issue?
Thanks for the report [tn84](https://github.com/tn84). Is it possible for you to provide the model defintion and a dataset that reproduces this issue? This is another issue with the deepexplainer that...
> Sure, I created a deep explainer label. Yep, to investigate further we really need a MRE: please see [writing helpful bug reports](https://shap.readthedocs.io/en/latest/contributing.html#writing-helpful-bug-reports). Thanks a lot! I got an example...
@swap357 thanks for the swift and exhaustive reply. That works for me and is good enough. I suppose `llvmlite==0.45.1` does not support Python 3.14 though? Probably not a big deal,...
Thanks for the report. This should be a relatively easy fix by adding the correct custom gradient to `op_handlers`. Could somebody please provide an example so that we can reliably...
Could you please provide some more context on this? We need a reproducible example to investigate this further