v2.8.0

2.8.0 (2023-11-06)

Features

• export llama.cpp release info (#91) (190ef96)

v2.7.5

2.7.5 (2023-11-05)

v2.7.4

2.7.4 (2023-10-25)

Bug Fixes

• do not download redundant node headers (#80) (ff1644d)
• improve cmake custom options handling (#80) (ff1644d)
• do not set CMAKE_GENERATOR_TOOLSET for CUDA (#80) (ff1644d)
• do not fetch information from GitHub when using a local git bundle (#80) (ff1644d)
• GBNF JSON schema string const formatting (#80) (ff1644d)

Features

• adapt to the latest llama.cpp interface (#80) (ff1644d)
• print helpful information to help resolve issues when they happen (#80) (ff1644d)
• make portable cmake on Windows more stable (#80) (ff1644d)
• update CMakeLists.txt to match llama.cpp better (#80) (ff1644d)

v2.7.3

2.7.3 (2023-10-13)

v2.7.2

2.7.2 (2023-10-12)

Features

• minor: save and load history to chat command (#71) (dc88531)

v2.7.1

2.7.1 (2023-10-11)

Bug Fixes

• GeneralChatPromptWrapper output (#70) (4ff8189)
• improve JSON schema validation error messages (#69) (c41da09)

v2.7.0

2.7.0 (2023-10-11)

Features

• add JSON schema grammar support (#68) (8ceac05)
• add promptWithMeta function to LlamaChatSession (#68) (8ceac05)

v2.6.3

2.6.3 (2023-10-10)

Bug Fixes

• improve CUDA compilation (#66) (8691585)

v2.6.2

2.6.2 (2023-10-09)

Bug Fixes

• add documentation to Google Search (#65) (eb61383)

v2.6.1

2.6.1 (2023-10-09)

Bug Fixes

• git bundle info for build (#64) (88e89d2)