Mohammed Safvan 2177dbfbad fix(llama-helper): align llama-cpp token decoding [skip ci] ...

Refs #462

Addresses L-1, L-9

Replace the deprecated token_to_bytes usage with token_to_piece_bytes while preserving special-token decoding behavior and retrying when the initial token buffer is too small.

Pin llama-cpp-2 to the locked 0.1.146 version so the helper manifest matches the resolved dependency used by the workspace.

2026-05-30 12:47:21 +05:30

..

src

fix(llama-helper): align llama-cpp token decoding [skip ci]

2026-05-30 12:47:21 +05:30

Cargo.toml

fix(llama-helper): align llama-cpp token decoding [skip ci]

2026-05-30 12:47:21 +05:30