feat: improve thought segments syntax extraction - #636

Merged

giladgd merged 25 commits into

gilad/adaptToBreakingChanges

Aug 9, 2026

Commits on Jul 25, 2026

fix: adapt to breaking `llama.cpp` changes
giladgd
committed
0fcf724
fix: map GgmlType configs directly from native code
giladgd
committed
f3e4ad6
feat: expose the model's architecture directly on the model instance
giladgd
committed
b059872

Commits on Aug 4, 2026

Commits on Aug 7, 2026

docs: add type links
giladgd
committed
5e17a4a

Commits on Aug 8, 2026

fix: Qwen chat wrapper auto thought segment opening, optimize checkpoints, bug fixes
giladgd
committed
614420e
fix: safer cleanup flows
giladgd
committed
76a28b5
refactor: Vulkan device memory readings
giladgd
committed
a27bee1
feat: expose download speed and ETA for a model downloader
giladgd
committed
f5d5fe4
feat: export gguf file quant names
giladgd
committed
ad977f3
test: fix tests
giladgd
committed
30ee8ea
build: fix docs build
giladgd
committed
6e5a899
fix: adapt to breaking `llama.cpp` changes
giladgd
committed
fecd41e
fix: vulkan info
giladgd
committed
b64874e
test: fix tests
giladgd
committed
2c3d52c
build: update npm on the riscv runner
giladgd
committed
10554ca
build: adapt to a recent npm change
giladgd
committed
2139dd1
test: fix tests
giladgd
committed
c79d666
test: fix test
giladgd
committed
5b3ccf6

Commits on Aug 9, 2026

fix(`LlamaContextSequence`): make `.dispose()` return a promise
giladgd
committed
57fbc4f
chore: update `lifecycle-utils`
giladgd
committed
66d55c1