Release 0.3.13

• feat: sync llama.cpp (#80) (4795559)

Release 0.3.12

• feat(ts): setup native log for all loaded mods (79d1a70)
• fix(cpp): add back deprecated isChatTemplateSupported model info (f92f23f)

Release 0.3.11

• feat: add toggleNativeLog / addNativeLogListener support (b1a50c4)
• fix(cpp): chatTemplates.toolUse & caps (56301f1)

Release 0.3.10

• feat: sync llama.cpp & llama.rn api changes (#79) (d0a86ab)
• fix(cpp): use size 4096 for get meta str (1bf44e5)

Release 0.3.9

• feat: implement lora params & methods (#78) (02f7b4e)
• feat(cpp): add nEmbd in model info (0c89d41)

Release 0.3.8

• feat: add static model info method (#76) (6396083)
• feat: sync llama.cpp (#69) (2cbf1aa)

Release 0.3.7

• feat: build cuda bin for linux (x86/arm) (#77) (f084d66)
• feat: use new linux arm runner for build build-linux-arm64 & off GGML_NATIVE (#75) (9c6199a)

Release 0.3.6

• feat: use string enum for pooling_type (7321702)
• feat: expose flash_attn / cache_type_k / cache_type_v (ff47142)
• feat: expose ubatch param (c571221)

Release 0.3.5

• feat: implement vocab_only mode for context (#67) (5c91e0d)
• feat(cpp): add timings in completion result (7a6ac29)

Release 0.3.4

• feat: add XTC / DRY sampler params (#65) (08c2b14)
• feat: sync llama.cpp (#62) (5cfa9d6)
• feat: update embedding (#64) (1fc4c92)
• feat: add getModelInfo method (#63) (845a9f0)