Releases: oneapi-src/oneDAL
Releases · oneapi-src/oneDAL
Intel® oneAPI Data Analytics Library 2021.1
The release contains all functionality of Intel® DAAL. See Intel® DAAL release notes for more details.
What's New
Library Engineering:
- Renamed the library from
Intel® Data Analytics Acceleration Library
toIntel® oneAPI Data Analytics Library
and changed the package names to reflect this. - Deprecated 32-bit version of the library.
- Introduced Intel GPU support for both
OpenCL
andLevel Zero
backends. - Introduced
Unified Shared Memory
(USM
) support
Introduced new Intel® oneDAL and daal4py functionality:
- GPU:
- Batch algorithms:
K-means
,Covariance, PCA
,Logistic Regression
,Linear Regression
,Random Forest Classification
andRegression
,Gradient Boosting Classification
andRegression
,kNN
,SVM
,DBSCAN
andLow-order moments
- Online algorithms:
Covariance
,PCA
,Linear Regression
andLow-order moments
- Added
Data Management
functionality to supportDPC++ APIs
: a new table type for representation ofSYCL-based
numeric tables (SyclNumericTable
) and an optimizedCSV data source
- Batch algorithms:
Improved Intel® oneDAL and daal4py performance for the following algorithms:
- CPU:
Logistic Regression
training and predictionk-Nearest Neighbors
prediction withBrute Force
methodLogistic Loss
andCross Entropy objective functions
Added Technical Preview Features in Graph Analytics:
- CPU:
- Undirected graph without edge and vertex weights (
undirected_adjacency_array_graph
), where vertex indices can only be of type int32 Jaccard Similarity Coefficients
for all pairs of vertices, a batch algorithm that processes the graph by blocks
- Undirected graph without edge and vertex weights (
Aligned the library with Intel® oneDAL Specification 1.0 for the following algorithms:
- CPU/GPU:
K-means
,PCA
,kNN
Introduced new functionality for scikit-learn patching through daal4py:
- CPU:
- Acceleration of
NearestNeighbors
andKNeighborsRegressor
scikit-learn estimators withBrute Force
andK-D tree
methods - Acceleration of
TSNE
scikit-learn estimator
- Acceleration of
- GPU:
- Intel GPU support in scikit-learn for
DBSCAN
,K-means
,Linear
andLogistic Regression
- Intel GPU support in scikit-learn for
Improved performance of the following scikit-learn estimators via scikit-learn patching:
- CPU:
LogisticRegression
fit, predict and predict_proba methodsKNeighborsClassifier
predict, predict_proba and kneighbors methods with“brute”
method
Known Issues
Intel® oneDAL DPC++ APIs
does not work onGEN12
graphics withOpenCL
backend. UseLevel Zero
backend for such cases.train_test_split
indaal4py
patches forScikit-learn
can produce incorrect shuffling on Windows*
Intel® DAAL 2020 Update 3
What's New in Intel® DAAL 2020 Update 3:
Introduced new Intel® DAAL and daal4py functionality:
- Brute Force method for
k-Nearest Neighbors
classification algorithm, which for datasets with more than 13 features demonstrates a better performance than the existing K-D tree method k-Nearest Neighbors
search for K-D tree and Brute Force methods with computation of distances to nearest neighbors and their indices
Extended existing Intel® DAAL and daal4py functionality:
- Voting methods for prediction in
k-Nearest Neighbors
classification and search: based on inverse-distance and uniform weighting - New parameters in
Decision Forest
classification and regression: minObservationsInSplitNode, minWeightFractionInLeafNode, minImpurityDecreaseInSplitNode, maxLeafNodes with best-first strategy and sample weights - Support of Support Vector Machine (
SVM
) decision function for Multi-class Classifier
Improved Intel® DAAL and daal4py performance for the following algorithms:
SVM
training and predictionDecision Forest
classification trainingRBF
andLinear
kernel functions
Introduced new daal4py functionality:
- Conversion of trained
XGBoost
* andLightGBM
* models into a daal4py Gradient Boosted Trees model for fast prediction - Support of
Modin
* DataFrame as an input
Introduced new functionality for scikit-learn patching through daal4py:
- Acceleration of
KNeighborsClassifier
scikit-learn estimator with Brute Force and K-D tree methods - Acceleration of
RandomForestClassifier
andRandomForestRegressor
scikit-learn estimators - Sparse input support for
KMeans
and Support Vector Classification (SVC
) scikit-learn estimators - Prediction of probabilities for
SVC
scikit-learn estimator - Support of ‘normalize’ parameter for
Lasso
andElasticNet
scikit-learn estimators
Improved performance of the following functionality for scikit-learn patching through daal4py:
train_test_split()
- Support Vector Classification (
SVC
) fit and prediction
Dependencies
fix one-algorithm build and spicific prediction case after probabilit… …y feature (#177)
DAAL 2020
DAAL 2019 Update 4
Revision: 33235
Linux* (32-bit and 64-bit binary): l_daal_oss_p_2019.4.007.tgz
macOS* (32-bit and 64-bit binary): m_daal_oss_p_2019.4.007.tgz
Note: Please, use Git client with enabled Git LFS module to clone repository if you want to get sources. We are working with GitHub support to enable correct work of archives ”Source code (zip)" and "Source code (tar.gz)".