NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Packagist, #3232 most downloaded on Packagist
A high-level machine learning and deep learning library for the PHP language.
Latest stable
3.3.0
Last release 27 days ago
20 Oct 2026
Ships unpredictably
gaps range from 8 days to 12 months
Nearly every release is documented
notes for 59 of the last 60 stable releases
Nothing withdrawn
no release was ever pulled
9 years old
105 releases · first in 2018
Drag across the ruler to compare, or tap a release to open it.
Removed previously deprecated portions of the API
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Update to Flysystem 2.1 and above
Fix Missing Extension exception class filename
Optimized CART node splitting for low variance continuous features
Deprecated Igbinary serializer and move to Extras package
Fix t-SNE momentum gain bus error when using Tensor extension
Fix CART feature importances purity increase overflow
Deprecated the Ranking interface
Note: This update includes changes to class definitions which may break previously persisted models if upgraded. See the caveats section of the persistence documentation for more info.
Categorized and annotated internal API
Now compatible with PHP 8 GD Image types
Optimized CART quantile-based node splitting
Allow list of empty regex patterns in Regex Filter
Added Dataset join() method and deprecated augment()
Note: This update includes changes to class definitions which may break previously persisted models if upgraded. See the caveats section of the persistence documentation for more info.
Fix KNN Imputer spatial tree dependency injection
Compensate for zero vectors in Cosine kernel
Optimized Cosine distance for sparse vectors
Optimized Cosine distance kernel
Fixed Grid Search best hyper-parameters method
Fixed Image Resizer placeholder image
CV Report Generators now return Report objects
Improved logging for Verbose Learners
Added Multibyte Text Normalizer transformer
Embedders now adopt the Transformer API
Persistent Model now implements Verbose interface
Random Forest now handles imbalanced datasets
Nothing published for this version
Added SiLU self-stabilizing neural network activation function
Added phpbench performance benchmarks
Added Tensor extension compatibility
Radius Neighbors allows user-definable anomaly class
Added Gaussian MLE anomaly detector
Added feature importances to Gradient Boost
Added Regression and Classification Loss interfaces
Added Parallel interface for multiprocessing
K Means now uses mini batch GD instead of SGD
Added LODA online anomaly detector
Added transform labels method to Labeled Dataset
Added Model Orchestra meta estimator
Added Support Vector Machine classifier and regressor
AdaBoost now uses SAMME multiclass algorithm
Added Gaussian Mixture clusterer
Added K-d Neighbors classifier and regressor
Random Forest now supports Extra Trees
Nothing published for this version
Nothing published for this version
One column per quarter.