2.1.0 Release

The new release incorporates the following changes:

• Work around a compiler bug in GCC 7 (#302)
• Fix errors reported by Python linter (#312)
• [CI] Upgrade to CentOS 7 + GCC 8 (#313)
• Improve vector leaf performance (#311)
• Handle the case where left_categories={} (#303)
• Use JSON serializer to enable model introspection (#296)
• Fix NativeLibLoader on Java 9+ (#315)

2.0.0 Release

The new release incorporates the following major changes:

• Remove dmlc-core dependency (#285, #287, #288, #289, #290, #291, #292, #293). #284 explains why we are jumping to a new major version.
• Rename CHECK and LOG macros to avoid name collisions (#295)
• Suppress warning about build time when parallel_comp is specified (#297)
• Handle LightGBM models with '=' in feature names (#298)
• Remove OpenMP dependency (#300)

1.3.0 release

This version incorporates one new major feature:

• [EXPERIMENTAL] Add General Tree Inference Library (GTIL), reference inference backend (#274). GTIL is useful in cases where it is infeasible to build the tree models as native shared libs.

It also contains the following bug fixes:

• Hotfix for serializer (#273)

1.2.0 release

This version incorporates two major new features:

• Faster model import for sklearn tree models (#264)
• Binary serializer to a file stream (#270)

Release 0.93

[CI] Use mainline branch

Release 0.92

Release 0.92

Treelite 0.32

Usability improvements

• Add __version__ attribute (#17)
• Shape results appropriately for multiclass classification (#16)
• Add default LOG handler, to make C API easier to use (#19)
• Allow batches smaller than nthread

Bug fixes

• Python 2.7 compatibility fix (#14)
• handle categorical splits with empty left_categories
• Fix create_shared() to be compatible with shells zsh and fish (#18)

New feature

• (Experimental) Export model as XGBoost model file, as per request in #14.

Treelite 0.31

• Implement a custom thread pool to improve prediction performance
  • Amortize thread launch cost
  • Pin down threads to physical cores and prevent migration
• Accommodate infinite thresholds in LightGBM models

Treelite 0.30

Major re-factoring of the backend code

• New code generator, based on Abstract Syntax Trees (AST)
  • Build AST first
  • Iteratively transform AST
  • Finally convert AST into C/JAVA program
• Improved abstraction and modularity (no more spaghetti code!)
  • It's now much easier to "edit" tree prediction logic: you'd just need to operate on a tree graph. Re-factoring was a worthwhile investment for future research. For instance, we could create a new node type in AST representing an AVX operation. A group of conditional nodes can be converted into the AVX node with equivalent semantics.
• Support for Java: see example at https://github.com/hcho3/treelite-java-example

BREAKING CHANGES

• Prediction library generated by previous versions are NOT compatible with the current version. You will have to run export_lib() or compile() again.

Treelite 0.21

New features

• Now categorical features with more than 64 categories are supported.
• It is now possible to specify variants of GCC and Clang for export_lib: gcc-5, gcc-7, arm-linux-gnueabi-gcc, and so forth.

Bug fixes

• Fix segmentation fault when parallel_comp is set to a high value