alibi
alibi copied to clipboard
SeldonIO
Algorithms for explaining machine learning models
Add support for Pandas data frames in any relevant explainers. (See #596)
Quick question, is [IntegratedGradients.explain(..)](https://github.com/SeldonIO/alibi/blob/master/alibi/explainers/integrated_gradients.py#L777) thread-safe, i.e. could it be invoked from multiple threads safely? I assume no, as a call to `ig.explain` would modify the model to capture the gradients,...
I have an xgboost model I am trying to run couterfactuals on using the CounterfactualProto explainer. Given the RuntimeError I see below, I tired to disable the eager execution using...
The counterfactual reinforcement learning explainers don't currently allow the user to specify the device that alibi uses to train the actor and critic models on. Instead, the behaviour is to...
This PR is a draft proposal for specifying the return schemas of `Explanation` objects resulting from calls to `explain` methods on `Explainer` objects. # Motivation Currently finding out the exact...
I have a TensorFlow based Onnx classifier model. It's a binary classifier with the prediction function as "model.classify('image')". But I am getting the following errors while passing it to Alibi...
I am getting this weird error while using **AnchorImage** for object detection model as well as image classification model error says `boolean index did not match indexed array along dimension...
~We should go over the API docs and check that everything is rendered and documented correctly. Additionally, there are a bunch of methods and functions which should be made private...
Every `Explanation` has a wide variety of available fields in the `data` dictionary depending on the algorithm, but some fields are common across algorithm types yet have different names and...
Many `alibi` implementations rely heavily on user callbacks, allowing to customize the behaviour of explainers. On the other hand, for application developers it is sometimes necessary to be able to...