ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

model : add NeoBERT (#14164)

Browse source 
* convert neobert model to gguf

* add inference graph

* fix flake8 lint

* followed reviewer suggestions

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* follow reviewers suggestions

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

* override NeoBERT feed-forward length

---------

Co-authored-by: dinhhuy <huy.dinh@brains-tech.co.jp>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Đinh Trọng Huy 2025-06-16 21:53:41 +09:00 • committed by GitHub
parent 7d6d91babf
commit ad590be98c
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
6 changed files with 222 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

14
gguf-py/gguf/constants.py
View file

@ -291,6 +291,7 @@ class MODEL_ARCH(IntEnum):
BERT = auto()
NOMIC_BERT = auto()
NOMIC_BERT_MOE = auto()
NEO_BERT = auto()
JINA_BERT_V2 = auto()
BLOOM = auto()
STABLELM = auto()
@ -573,6 +574,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = {
MODEL_ARCH.BERT: "bert",
MODEL_ARCH.NOMIC_BERT: "nomic-bert",
MODEL_ARCH.NOMIC_BERT_MOE: "nomic-bert-moe",
MODEL_ARCH.NEO_BERT: "neo-bert",
MODEL_ARCH.JINA_BERT_V2: "jina-bert-v2",
MODEL_ARCH.BLOOM: "bloom",
MODEL_ARCH.STABLELM: "stablelm",
@ -1081,6 +1083,18 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.FFN_UP_EXP,
MODEL_TENSOR.LAYER_OUT_NORM,
],
MODEL_ARCH.NEO_BERT: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.ATTN_NORM,
MODEL_TENSOR.ATTN_QKV,
MODEL_TENSOR.ATTN_OUT,
MODEL_TENSOR.FFN_NORM,
MODEL_TENSOR.FFN_DOWN,
MODEL_TENSOR.FFN_UP,
MODEL_TENSOR.ENC_OUTPUT_NORM,
MODEL_TENSOR.CLS,
MODEL_TENSOR.CLS_OUT,
],
MODEL_ARCH.JINA_BERT_V2: [
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.TOKEN_EMBD_NORM,