]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
model : full modern bert support (#18330)
authorRyan Mangeno <redacted>
Thu, 19 Feb 2026 07:52:21 +0000 (02:52 -0500)
committerGitHub <redacted>
Thu, 19 Feb 2026 07:52:21 +0000 (08:52 +0100)
commitc0d04303400e64a798506e3f2342940ae268db15
treee10296443d7f3e8f2c92489fee8dd937cc954c70
parent3bb2fcc8567e139a0ef70b8d43f82a3130147c00
model : full modern bert support (#18330)

* full modern bert support

* added gelu op in rank pooling for modern bert

* still working on stuff, added mean calculation before classifier head

* Update convert_hf_to_gguf.py

Co-authored-by: Sigbjørn Skjæret <redacted>
* first layer is dense, as per modern bert research paper

* Update src/llama-graph.cpp

Co-authored-by: Sigbjørn Skjæret <redacted>
* fixed set input for mean pooling to check if pooling type is ranking since modern bert does mean & rank

* Update src/llama-graph.cpp

Co-authored-by: Sigbjørn Skjæret <redacted>
* Update convert_hf_to_gguf.py

Co-authored-by: Sigbjørn Skjæret <redacted>
---------

Co-authored-by: Sigbjørn Skjæret <redacted>
12 files changed:
convert_hf_to_gguf.py
gguf-py/gguf/constants.py
gguf-py/gguf/tensor_mapping.py
src/llama-arch.cpp
src/llama-arch.h
src/llama-context.cpp
src/llama-graph.cpp
src/llama-graph.h
src/llama-model-saver.cpp
src/llama-model.cpp
src/llama-model.h
src/models/modern-bert.cpp