[REVIEW] Kernel shap improvements #5187
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
I'll take a look at CI failures. |
vinaydes
changed the title
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## branch-23.04 #5187 +/- ##
===============================================
Coverage ? 67.12%
===============================================
Files ? 192
Lines ? 12396
Branches ? 0
===============================================
Hits ? 8321
Misses ? 4075
Partials ? 0 Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
CI is successful, ready to merge. |
dantegd
approved these changes
Feb 7, 2023
|
Changes look great! Just merging branch-23.04 into the PR and will merge after CI runs. There's a new CI check that fails if a PR is 5+ commits behind the target branch, we might increase that threshold. |
dantegd
added
improvement
|
/merge |
AyodeAwe
pushed a commit
to AyodeAwe/cuml
that referenced
this pull request
Feb 13, 2023
Removed slow modulo operator by minor change in index arithmetic. This gave me following performance improvement for a test case:
| | branch-23.02 |kernel-shap-improvments | Gain |
|-------------------------|------------------|-------------------------|------|
| sampled_rows_kernel | 663 | 193 | 3.4x |
| exact_rows_kernel | 363 | 236 | 1.5x |
All times in microseconds.
Code used for benchmarking:
```python
from sklearn.datasets import make_classification
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestRegressor as rf
from cuml.explainer import KernelExplainer
import numpy as np
data, labels = make_classification(n_samples=1000, n_features=20, n_informative=20, random_state=42,
n_redundant=0, n_repeated=0)
X_train, X_test, y_train, y_test = train_test_split(data, labels, train_size=998,
random_state=42) #sklearn train_test_split
y_train = np.ravel(y_train)
y_test = np.ravel(y_test)
model = rf(random_state=42).fit(X_train, y_train)
cu_explainer = KernelExplainer(model=model.predict, data=X_train, is_gpu_model=False, random_state=42, nsamples=100)
cu_shap_values = cu_explainer.shap_values(X_test)
print('cu_shap:', cu_shap_values)
```
Authors:
- Vinay Deshpande (https://github.com/vinaydes)
- Dante Gama Dessavre (https://github.com/dantegd)
Approvers:
- Dante Gama Dessavre (https://github.com/dantegd)
URL: rapidsai#5187
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Removed slow modulo operator by minor change in index arithmetic. This gave me following performance improvement for a test case:
All times in microseconds.
Code used for benchmarking: