Skip to content

feat: update llama.cpp to ggml-org/llama.cpp@9e3b928fd #2282

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
abetlen merged 1 commit into from
Jun 7, 2026
+9 −1
Merged

feat: update llama.cpp to ggml-org/llama.cpp@9e3b928fd #2282

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0

## [Unreleased]

- feat: update llama.cpp to ggml-org/llama.cpp@9e3b928fd
- feat(example): add OpenAI-compatible embeddings endpoint by @abetlen in #2281

## [0.3.27]
Expand Down
7 changes: 7 additions & 0 deletions llama_cpp/llama_cpp.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -949,6 +949,10 @@ class llama_sampler_seq_config(ctypes.Structure):
# // ref: https://github.com/ggml-org/llama.cpp/pull/14363
# struct llama_sampler_seq_config * samplers;
# size_t n_samplers;
#
# // a source/target/parent context
# // can be utilized in various ways, for example by sharing results or llama_memory between 2 contexts
# struct llama_context * ctx_other;
# };
class llama_context_params(ctypes.Structure):
"""Parameters for llama_context
Expand Down Expand Up @@ -989,6 +993,7 @@ class llama_context_params(ctypes.Structure):
kv_unified (bool): use a unified buffer across the input sequences when computing the attention
samplers (ctypes.POINTER(llama_sampler_seq_config)): backend sampler chain configuration
n_samplers (int): number of backend sampler chain configurations
ctx_other (llama_context_p): source, target, or parent context
"""

if TYPE_CHECKING:
Expand Down Expand Up @@ -1027,6 +1032,7 @@ class llama_context_params(ctypes.Structure):
kv_unified: bool
samplers: ctypes.POINTER(llama_sampler_seq_config)
n_samplers: int
ctx_other: llama_context_p

_fields_ = [
("n_ctx", ctypes.c_uint32),
Expand Down Expand Up @@ -1064,6 +1070,7 @@ class llama_context_params(ctypes.Structure):
("kv_unified", ctypes.c_bool),
("samplers", ctypes.POINTER(llama_sampler_seq_config)),
("n_samplers", ctypes.c_size_t),
("ctx_other", llama_context_p_ctypes),
]


Expand Down
2 changes: 1 addition & 1 deletion vendor/llama.cpp
Open in desktop
Submodule llama.cpp updated 39 files
+1 −1 common/arg.cpp
+1 −1 common/common.cpp
+49 −40 common/sampling.cpp
+1 −1 common/sampling.h
+50 −41 common/speculative.cpp
+2 −0 conversion/__init__.py
+10 −0 conversion/gemma.py
+3 −2 conversion/mistral.py
+1 −1 convert_hf_to_gguf.py
+24 −0 gguf-py/gguf/constants.py
+8 −0 gguf-py/gguf/tensor_mapping.py
+4 −0 include/llama.h
+5 −0 src/llama-arch.cpp
+3 −0 src/llama-arch.h
+37 −18 src/llama-context.cpp
+2 −1 src/llama-context.h
+2 −0 src/llama-cparams.h
+2 −0 src/llama-ext.h
+18 −13 src/llama-graph.cpp
+1 −0 src/llama-graph.h
+4 −0 src/llama-hparams.cpp
+4 −0 src/llama-hparams.h
+2 −2 src/llama-kv-cache-dsa.cpp
+15 −3 src/llama-kv-cache-iswa.cpp
+3 −1 src/llama-kv-cache-iswa.h
+122 −21 src/llama-kv-cache.cpp
+10 −3 src/llama-kv-cache.h
+2 −0 src/llama-kv-cells.h
+2 −0 src/llama-memory-hybrid-iswa.cpp
+2 −0 src/llama-memory-hybrid.cpp
+4 −0 src/llama-memory.h
+64 −23 src/llama-model.cpp
+5 −0 src/llama-model.h
+200 −0 src/models/gemma4-assistant.cpp
+18 −4 src/models/gemma4.cpp
+13 −0 src/models/models.h
+3 −3 tests/test-llama-archs.cpp
+17 −8 tools/server/server-context.cpp
+1 −1 tools/server/server-task.cpp
Loading