v2.5.0

2.5.0 (2023-09-26)

Bug Fixes

• adapt to llama.cpp interface change (#49) (9db72b0)

Features

• add FalconChatPromptWrapper (#53) (656bf3c)
• fall back to build from source if prebuilt binary loading fails (#54) (d99e3b0)
• load conversation history into a LlamaChatSession (#51) (4e274ce)
• only build one binary for all node versions (#50) (1e617cd)

v2.4.1

2.4.1 (2023-09-15)

v2.4.0

2.4.0 (2023-09-09)

Features

• use cmake instead of node-gyp (#37) (01b89ce)

v2.3.2

2.3.2 (2023-09-02)

Bug Fixes

• load image urls properly also outside GitHub (#35) (cf1f5f1)

v2.3.1

2.3.1 (2023-09-02)

v2.3.0

2.3.0 (2023-09-02)

Bug Fixes

• handle stop words remainder properly in a chat session (#32) (9bdef11)
• move default export to be the last one in package.json (#31) (dd49959)

Features

• threads count setting on a model (#33) (47c3c5f)

v2.2.0

2.2.0 (2023-09-01)

Features

• export class options types (#29) (74be398)
• improve error message when llama.cpp source is not downloaded (#27) (7837af7)
• make contributions and support more efficient via GitHub templates (#28) (5fc0d18)

v2.1.2

2.1.2 (2023-08-28)

Bug Fixes

• import (#21) (fd332e1)

v2.1.1

2.1.1 (2023-08-28)

v2.1.0

2.1.0 (2023-08-28)

Features

• add grammar support (#13) (c28d2de)
• add support for metal and cuda in the build command (#17) (1043596)