- Upgrade llama.cpp to 60ed04cf to support qwen1.5.
- Code refine.
- Support embedding.
- Support GPU building.
- Add a RAG example.
- Upgrade llama.cpp to 8c0e8f4e.
- Upgrade Dart to 3.3.0.
- Upgrade dependencies and fix issues.
- Initial version.
- Integrate with
native_assets_cli. - Native helper classes.
- Porting token sampling from
common/sampling.cpp. - Passing LLM params to the isolate.
- Token string as raw bytes stream.