matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-14 01:37:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: update vendored code to commit 46e3556 (#8308)

Browse Source
This commit is contained in:
Jeffrey Morgan
2025-01-08 11:22:01 -08:00
committed by GitHub
parent 57f038ec7b
commit 1deafd8254
305 changed files with 16048 additions and 12926 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
llama/patches/0009-add-unpad-operator.patch
View File

@@ -15,7 +15,7 @@ Subject: [PATCH] add unpad operator
8 files changed, 220 insertions(+), 2 deletions(-)
diff --git a/ggml/include/ggml.h b/ggml/include/ggml.h
index b0c1ac9c..091e6e6b 100644
index c714fc8c..1bc50fca 100644
--- a/ggml/include/ggml.h
+++ b/ggml/include/ggml.h
@@ -499,6 +499,7 @@ extern "C" {
@@ -26,7 +26,7 @@ index b0c1ac9c..091e6e6b 100644
GGML_OP_ARANGE,
GGML_OP_TIMESTEP_EMBEDDING,
GGML_OP_ARGSORT,
@@ -1718,6 +1719,15 @@ extern "C" {
@@ -1735,6 +1736,15 @@ extern "C" {
int p0,
int p1);
@@ -43,7 +43,7 @@ index b0c1ac9c..091e6e6b 100644
// timesteps: [N,]
// return: [N, dim]
diff --git a/ggml/src/ggml-cpu/ggml-cpu.c b/ggml/src/ggml-cpu/ggml-cpu.c
index 67e67a08..bebff207 100644
index b7fefb9d..b307d554 100644
--- a/ggml/src/ggml-cpu/ggml-cpu.c
+++ b/ggml/src/ggml-cpu/ggml-cpu.c
@@ -10588,6 +10588,59 @@ static void ggml_compute_forward_pad_reflect_1d(
@@ -126,7 +126,7 @@ index 67e67a08..bebff207 100644
case GGML_OP_TIMESTEP_EMBEDDING:
case GGML_OP_ARGSORT:
diff --git a/ggml/src/ggml-cuda/ggml-cuda.cu b/ggml/src/ggml-cuda/ggml-cuda.cu
index 8fd7c1a3..7c351b89 100644
index bb425ee8..1e7c2a22 100644
--- a/ggml/src/ggml-cuda/ggml-cuda.cu
+++ b/ggml/src/ggml-cuda/ggml-cuda.cu
@@ -2085,6 +2085,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg
@@ -139,7 +139,7 @@ index 8fd7c1a3..7c351b89 100644
case GGML_OP_ARANGE:
ggml_cuda_op_arange(ctx, dst);
break;
@@ -3012,6 +3015,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
@@ -3013,6 +3016,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
case GGML_OP_GROUP_NORM:
case GGML_OP_UPSCALE:
case GGML_OP_PAD:
@@ -211,7 +211,7 @@ index 8fd386b0..e2ededc3 100644
void ggml_cuda_op_pad(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
+void ggml_cuda_op_unpad(ggml_backend_cuda_context & ctx, ggml_tensor * dst);
diff --git a/ggml/src/ggml-metal/ggml-metal.m b/ggml/src/ggml-metal/ggml-metal.m
index 28f590f9..787fc713 100644
index a85502ee..84e027eb 100644
--- a/ggml/src/ggml-metal/ggml-metal.m
+++ b/ggml/src/ggml-metal/ggml-metal.m
@@ -311,6 +311,7 @@ static void ggml_backend_metal_device_rel(struct ggml_backend_metal_device_conte
@@ -332,7 +332,7 @@ index 8ba43904..204c93e6 100644
device char * dst,
constant int64_t & ne0,
diff --git a/ggml/src/ggml.c b/ggml/src/ggml.c
index 51cc8566..0e74e554 100644
index 2bbe5f48..7ffcd907 100644
--- a/ggml/src/ggml.c
+++ b/ggml/src/ggml.c
@@ -954,6 +954,7 @@ static const char * GGML_OP_NAME[GGML_OP_COUNT] = {
@@ -369,7 +369,7 @@ index 51cc8566..0e74e554 100644
static_assert(GGML_OP_POOL_COUNT == 2, "GGML_OP_POOL_COUNT != 2");
@@ -4180,6 +4182,25 @@ struct ggml_tensor * ggml_pad_reflect_1d(
@@ -4214,6 +4216,25 @@ struct ggml_tensor * ggml_pad_reflect_1d(
return result;
}