Third-party notices
KaLM-Reranker
The GGUF weight is a text-only Q8_0 conversion of the corresponding
KaLM-Embedding/KaLM-Reranker-V1-* model. The source model declares the
Apache License 2.0. See the base_model link in the model card for the
original model documentation, authorship and citation information.
llama.cpp
The runtime patch bundle modifies
ggml-org/llama.cpp, based on commit
277a105dc8f8643dab54331926a9830860a03292. llama.cpp and the derived patch
contents are distributed under the MIT License reproduced in
llama.cpp/LICENSE.
The patch bundle is provided because stock llama.cpp does not yet recognize
the t5gemma2 GGUF architecture used by this release.