Performance improvements in TSNE algorithm #1087
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
|
Rebase on latest develop and apply PEP8, please. |
|
tSNE related test fails on this job https://app.circleci.com/pipelines/github/intel/scikit-learn-intelex/5330/workflows/66a1fba7-78ef-4248-b4db-c4c74745e270/jobs/5758 |
…intel#1071) * Add workaround for n_iter_ SVM attribute * Pytest onedal dir * Exclude test_policy * fix result_option fail for kneighbors method in knn classifier
* Downgrade build version and set test version of numpy * Upgrade version to 1.20 for python<=3.9 * Set base numpy to be 1.19.* for build * Set numpy for future python version * Enabling full versions and >=3.10 python Co-authored-by: Nikolay Petrov <nikolay.a.petrov@intel.com>
* add system info * fix python activation on win * add system descriptor * fix descriptor calls * fix descriptor calls 2 * fix cpufeature installation on mac * change system descriptor url to oneapi-src/onedal * Add descriptor script * Remove extra variable
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
* add balanced branch of _compute_class_weight * Remove extra computation of weights
* link libonedal_dpc.so with libonedal_sycl.a due to mkl depency issue * only changing linker if linux * additional condition for windows linkage * removing multiply defined symbols * windows build completes, sycl errors, trying to start job * builds and 86/100 examples pass, cleaned up previous unnecessary additions * initial fix for Linux DPCPP failure * Flake8 appliance * add env variables for Linux DPCPP * Linux DPCPP ci update * Linux DPCPP check: update conda build step * moved compiler options setting into conda-recipe build * reverting linkage changes - resolved from mkl fpk update * address some comments updates passing -fsycl flag some refactoring * rebase * fix comments * Update CMakeLists.txt fix using gcc compiler * Apply pep8 * Apply pep8 n2 * Update CMakeLists.txt * resolving windows intel compiler issues * MD to MT flag for compiler and removing unneeded cmake verbose flag * updated gitignore * link libonedal_dpc.so with libonedal_sycl.a due to mkl depency issue * only changing linker if linux * additional condition for windows linkage * removing multiply defined symbols * windows build completes, sycl errors, trying to start job * builds and 86/100 examples pass, cleaned up previous unnecessary additions * initial fix for Linux DPCPP failure * Flake8 appliance * add env variables for Linux DPCPP * Linux DPCPP ci update * Linux DPCPP check: update conda build step * moved compiler options setting into conda-recipe build * reverting linkage changes - resolved from mkl fpk update * address some comments updates passing -fsycl flag some refactoring * rebase * fix comments * Update CMakeLists.txt fix using gcc compiler * Apply pep8 * Apply pep8 n2 * Update CMakeLists.txt * resolving windows intel compiler issues * MD to MT flag for compiler and removing unneeded cmake verbose flag * updated gitignore * Update setup.py * reverting MT flag back to MD * undoing last commit * reverting sycl linking addition * removing sycl addition pt. 2 * addressing Mac comments * link fsycl in cmake only when using dpc interface * FIX: daal4py spmd test * some misc fixes setup.py Co-authored-by: Samir Nasibli <samir.nasibli@intel.com> Co-authored-by: Alexander Andreev <alexander.andreev@intel.com>
inteldimitrius
force-pushed
the
dev/vsfedoto/tsne_opt
branch
from
be0c0ec
to
f4bb9f4
Compare
inteldimitrius
requested review from
maria-Petrova,
KulikovNikita and
napetrov
as code owners
Alexsandruss
requested changes
Dec 14, 2022
KulikovNikita
approved these changes
Dec 14, 2022
napetrov
approved these changes
Dec 14, 2022
Alexsandruss
approved these changes
Dec 14, 2022
maria-Petrova
approved these changes
Dec 15, 2022
napetrov
pushed a commit
that referenced
this pull request
Dec 15, 2022
|
Do you have some perf numbers of the improvements vs stock? |
Would be published earlier next year. |
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.
Those changes enable DAAL tSNE optimizations in sklearnex and fix some mainstream sklearn test failures in TSNE tests.