Added

• O200K encoding

Focused on internal refactoring and API stabilization to enhance usability and facilitate future development.

Breaking Changes

• Rename Enconding to EncodingConfig.
• Rework encodings to enhance extensibility:
  • Add a new Encoding interface.
  • Introduce default implementations of Encoding: CL100KBase, P50KBase, R50KBase, and P50KEdit.
• Rename Tokenizer.encoding and Tokenizer.encodingForModel to Tokenizer.of with overrides.

Changed

• JVM: use local PBE loader by default
• Encoding: enable custom encoding

Initial release.

Added

• Encodings: CL100K_BASE , R50K_BASE, P50K_BASE and P50K_EDIT
• Custom encoding support