Local Ai
b9148
Release b9148 of llama.cpp adds a non-backtracking tokenizer handler for Qwen3.5 to handle Unicode regex patterns for letters and combining marks, addressing stack overflow issues on long inputs. The
Release b9148 of llama.cpp adds a non-backtracking tokenizer handler for Qwen3.5 to handle Unicode regex patterns for letters and combining marks, addressing stack overflow issues on long inputs. The release includes test vocabulary files and regression test cases to ensure robust Unicode tokenization.
Source: llama.cpp Releases | 2026-05-14