]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
model : jina-embeddings-v3 support (#13693)
authorSigbjørn Skjæret <redacted>
Thu, 28 Aug 2025 13:49:50 +0000 (15:49 +0200)
committerGitHub <redacted>
Thu, 28 Aug 2025 13:49:50 +0000 (15:49 +0200)
commit84ab83cc0b4b7e769451ee48e4c7d1acef91ef25
tree583742aa4ef02d9b8734627cac771384d787e6f4
parent55042b3692cb1467c9ee15c62c4a9fbf180f89e3
model : jina-embeddings-v3 support (#13693)

* initial jina-embeddings-v3 support

* initial jina-embeddings-v3 support

* initial jina-embeddings-v3 support

* fix vocab parsing with only tokenizer.json

* set mask token lstrip attribute

* additional unk_token_id fallback just in case [no ci]

* revert vocab_size() change [no ci]

* merge tensor loading into general bert

* rope

* add lora embedding and loading (non-functional)

* export separate lora ggufs instead

* add adapter metadata api

* use std::string

* convert_hf_to_lora compatibility

* fix assert

* apply suggestions from review

* apply suggestion from review
14 files changed:
common/arg.cpp
common/common.cpp
common/common.h
convert_hf_to_gguf.py
gguf-py/gguf/constants.py
include/llama.h
src/llama-adapter.cpp
src/llama-adapter.h
src/llama-arch.cpp
src/llama-arch.h
src/llama-model.cpp
src/llama-model.h
src/llama-vocab.cpp
tools/server/server.cpp