Obtain CSR matrix from DMatrix. #8269
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
4 tasks
* Obtain CSR matrix from DMatrix. * Obtain gradient index from Quantile DMatrix. This is mostly for testing at higher level. Right now we rely on training a booster to infer that the DMatrix is correctly constructed. With this PR, we can ease the testing process. Also, this has been requested before. The return value from Quantile DMatrix is histogram index instead of cut values as we shift the cut value to include min_values, which is not very useful for language bindings. Unlike most of XGBoost C functions, caller of C API is required to allocate the memory itself instead of using thread local memory from XGBoost. This is to avoid allocating a huge memory buffer that can not be freed until exiting the thread. External memory is not supported.
RAMitchell
reviewed
Sep 26, 2022
include/xgboost/c_api.h
Outdated
| XGB_DLL int XGDMatrixNumNonMissing(DMatrixHandle handle, bst_ulong *out); | ||
|
|
||
| /*! | ||
| * \brief Get the predictors from DMatrix as CSR matrix. If this is a quantized DMatrix, |
There was a problem hiding this comment.
Could it be useful to return quantised float values instead of integers?
We could return values such that bst.predict(quantile_dmat) == bst.predict(xgb.DMatrix(quantile_dmat.get_data()))
There was a problem hiding this comment.
Yes, we can. That's a good argument. I skipped restoring the values because it contains artificially created min and max values, which might be confusing.
There was a problem hiding this comment.
Changed the return value to be cut values along with the addition of suggested tests.
|
@RAMitchell Please take another look when you are available. |
RAMitchell
approved these changes
Sep 29, 2022
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.
This is mostly for testing at a higher level. Right now we rely on training a booster to infer that the DMatrix is correctly constructed. With this PR, we can ease the testing process. Also, this has been requested before.
The return value from Quantile DMatrix is histogram index instead of cut values as we shift the cut value to include
min_values, which is not very useful for language bindings.Unlike most of XGBoost C functions, the caller of C API is required to allocate the memory itself instead of using thread local memory from XGBoost. This is to avoid allocating a huge memory buffer that can not be freed until exiting the thread.
External memory is not supported.
Close #4759 .