Add support for AIX360 explanations #1094
Conversation
aix api patch
Clean up example readme
update version format to match with KFServing
update v1beta1 spec and tests for rebase
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
This change is |
|
Thanks for the awesome contribution! Can we add a e2e test like alibi explainer here ? https://github.com/kubeflow/kfserving/blob/master/test/e2e/explainer/test_tabular_explainer.py |
AIX beta controller rebase
fix containerv1 spec
remove testing logs
|
/retest |
update e2e test to match with the v1beta1 controller changes
| }, | ||
| "aix": { | ||
| "image" : "aipipeline/aix-explainer", | ||
| "defaultImageVersion": "0.4.0" |
There was a problem hiding this comment.
please also change here which is used for KFServing CI https://github.com/kubeflow/kfserving/blob/master/config/overlays/test/configmap/inferenceservice.yaml#L39
| def __init__(self, name: str, predictor_host: str, segm_alg: str, num_samples: str, | ||
| top_labels: str, min_weight: str, positive_only: str, explainer_type: str): | ||
| super().__init__(name) | ||
| print("INIT") |
| KFServing.wait_isvc_ready(service_name, namespace=KFSERVING_TEST_NAMESPACE, timeout_seconds=720) | ||
| except RuntimeError as e: | ||
| logging.info(KFServing.api_instance.get_namespaced_custom_object("serving.knative.dev", "v1", | ||
| KFSERVING_TEST_NAMESPACE, "services", service_name + "-predictor")) |
There was a problem hiding this comment.
oh the service name is actually, service_name + "-predictor-default", I was trying to remove -default but figured it breaks the backwards compatibility.
address comments on ci config
move aix example to use beta api
|
@yuzisun @animeshsingh |
That's awesome! |
Co-authored-by: Animesh Singh <singhan@us.ibm.com>
|
/retest |
|
/lgtm |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: animeshsingh The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
What this PR does / why we need it:
Adds a new Explainer framework called Artificial Intelligence eXplainability 360 (AIX360) from the Trusted AI project. This PR contains the first explainer using a package called LIME (local interpretable model-agnostic explanations). A sample has been created which gives explanations for a random forest model's predictions on the MNIST dataset.
Which issue(s) this PR fixes (optional, in
fixes #<issue number>(, fixes #<issue_number>, ...)format, will close the issue(s) when PR gets merged):Fixes #1093
Special notes for your reviewer:
Release note: