From 60909fa38fc385716226f5b81178316a8eaec0d4 Mon Sep 17 00:00:00 2001 From: abetlen Date: Wed, 8 Apr 2026 01:53:14 -0700 Subject: [PATCH 1/2] feat: Update llama.cpp to ggerganov/llama.cpp@3bd9aa1f9 --- llama_cpp/llama_cpp.py | 8 ++++++-- vendor/llama.cpp | 2 +- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/llama_cpp/llama_cpp.py b/llama_cpp/llama_cpp.py index 204f1e626..0a66a5d85 100644 --- a/llama_cpp/llama_cpp.py +++ b/llama_cpp/llama_cpp.py @@ -88,7 +88,8 @@ def _warn_deprecated(symbol: str, hint: str) -> None: # GGML_TYPE_IQ1_M = 29, # GGML_TYPE_MXFP4 = 39, # GGML_TYPE_NVFP4 = 40, -# GGML_TYPE_COUNT, +# GGML_TYPE_Q1_0 = 41, +# GGML_TYPE_COUNT = 42, # }; GGML_TYPE_F32 = 0 GGML_TYPE_F16 = 1 @@ -120,7 +121,8 @@ def _warn_deprecated(symbol: str, hint: str) -> None: GGML_TYPE_IQ1_M = 29 GGML_TYPE_MXFP4 = 39 GGML_TYPE_NVFP4 = 40 -GGML_TYPE_COUNT = 41 +GGML_TYPE_Q1_0 = 41 +GGML_TYPE_COUNT = 42 # from ggml-backend.h # typedef bool (*ggml_backend_sched_eval_callback)(struct ggml_tensor * t, bool ask, void * user_data); @@ -406,6 +408,7 @@ def _warn_deprecated(symbol: str, hint: str) -> None: # LLAMA_FTYPE_MOSTLY_TQ2_0 = 37, // except 1d tensors # LLAMA_FTYPE_MOSTLY_MXFP4_MOE = 38, // except 1d tensors # LLAMA_FTYPE_MOSTLY_NVFP4 = 39, // except 1d tensors +# LLAMA_FTYPE_MOSTLY_Q1_0 = 40, // except 1d tensors # # LLAMA_FTYPE_GUESSED = 1024, // not specified in the model file # }; @@ -446,6 +449,7 @@ def _warn_deprecated(symbol: str, hint: str) -> None: LLAMA_FTYPE_MOSTLY_TQ2_0 = 37 LLAMA_FTYPE_MOSTLY_MXFP4_MOE = 38 LLAMA_FTYPE_MOSTLY_NVFP4 = 39 +LLAMA_FTYPE_MOSTLY_Q1_0 = 40 LLAMA_FTYPE_GUESSED = 1024 # enum llama_rope_scaling_type { diff --git a/vendor/llama.cpp b/vendor/llama.cpp index f49e91787..3bd9aa1f9 160000 --- a/vendor/llama.cpp +++ b/vendor/llama.cpp @@ -1 +1 @@ -Subproject commit f49e9178767d557a522618b16ce8694f9ddac628 +Subproject commit 3bd9aa1f9250cd15f5371f3622d73d954b68a747 From c2d3c7dd3a1715853da181f7483a050714940d8a Mon Sep 17 00:00:00 2001 From: abetlen Date: Wed, 8 Apr 2026 01:58:39 -0700 Subject: [PATCH 2/2] docs: Update changelog for llama.cpp bump --- CHANGELOG.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/CHANGELOG.md b/CHANGELOG.md index 9bc8d6654..fbe5b6b6f 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -7,6 +7,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ## [Unreleased] +- feat: Update llama.cpp to ggerganov/llama.cpp@3bd9aa1f9 and sync Python bindings + ## [0.3.20] - refactor: Replace deprecated llama.cpp references in library, docs, and examples by @abetlen in #2170