Oossa

llama.cpp b12425 adds Qwen3-ForcedAligner support

The release reads Qwen3-VL’s output-head size from its tensor, allowing llama.cpp to handle Qwen3-ForcedAligner’s smaller timestamp output.

NoteBy Published by Oossa: 1 min read

The llama.cpp project has released version b12425 with a change that allows it to read Qwen3-VL’s output-head size directly from the model tensor. The change supports Qwen3-ForcedAligner, whose output has 5,000 timestamp classes—fewer than the model’s vocabulary.

Why it matters

People running Qwen3-ForcedAligner with llama.cpp can use a model whose timestamp output is smaller than its vocabulary; the release notes do not specify setup requirements.

Was this article useful?
Share

Read next

Oossa · Newsletter

The week in AI, explained

Every Monday: the stories worth knowing, in plain language. Free, no spam.