Add support for Qwen3-Reranker

[iamlemec]

Add support for Qwen3 reranking models. This is largely based on #14029 by @ngxson, with a few tweaks to reflect changes to the codebase in the interim.

This hardcodes the chat template provided by in the README.md, which I'm assuming is the intended usage. If folks want to be able to change that, then we'd need a new CLI option. The template uses string substitution rather than jinja, as it seems like jinja is only used for chat messages.

Edit: Here's an example usage similar to that used in the official Qwen repo. Note that \t separates queries from documents and \n separates different prompts.

build/bin/llama-embedding -m qwen3-reranker-0.6b-f32.gguf --embd-normalize -1 -p "What is the capital of China?\tThe capital of China is Beijing.\nExplain gravity\tGravity is a force that attracts two bodies towards each other."

Notice that we need to pass --embd-normalize -1 to disable normalization (the default is L2 norm).

[ggerganov]

Great - will take a look tomorrow. Would be useful to add a basic usage example in the OP of this PR.

[iamlemec]

Yup, will add a usage example up above. Actually encountering some numerical differences comparing the output here to the transformers numbers. Might want to hold off on review/merge until I can sort this out. (Or if you spot something amiss in the code, that would be great too.)

[iamlemec]

Ok, finally fixed it! Now we have numerical parity with the HF implementation. It turned out to be a small difference in the chat template. Should be ready for review @ggerganov.

[ggerganov]

I am wondering if it makes sense to remove pooling type RANK all together from libllama? Do you have any thoughts about if having a separate pooling class RANK is really necessary?

[iamlemec]

I think you could get really close to merging RANK with LAST. The main differentiator is in llm_graph_context::build_pooling where you apply cls_out to map from the last token of the last hidden state to the classification output (usual yes/no). Unlike the other pooling types, you actually need knowledge of the model weights to do the calculation.