mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-22 14:53:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

update llama.cpp submodule to d7fd29f (#5475)

Browse Source
This commit is contained in:
Jeffrey Morgan
2024-07-05 13:25:58 -04:00
committed by GitHub
parent d89454de80
commit 8f8e736b13
15 changed files with 150 additions and 422 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
llm/patches/09-pooling.diff
View File

@@ -1,8 +1,8 @@
diff --git a/llama.cpp b/llama.cpp
index 61948751..61fe7b57 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -7591,14 +7591,14 @@ struct llm_build_context {
diff --git a/src/llama.cpp b/src/llama.cpp
index 721b8f4e..cfe7ac40 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -8420,14 +8420,14 @@ struct llm_build_context {
}
struct ggml_tensor * build_inp_mean() {
@@ -19,7 +19,7 @@ index 61948751..61fe7b57 100644
cb(lctx.inp_cls, "inp_cls", -1);
ggml_set_input(lctx.inp_cls);
return lctx.inp_cls;
@@ -12062,19 +12062,16 @@ static void llama_set_inputs(llama_context & lctx, const llama_batch & batch) {
@@ -13847,19 +13847,16 @@ static void llama_set_inputs(llama_context & lctx, const llama_batch & batch) {
GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_mean->buffer));
float * data = (float *) lctx.inp_mean->data;
@@ -42,7 +42,7 @@ index 61948751..61fe7b57 100644
const uint64_t s = sum[i];
if (s > 0) {
div[i] = 1.0f/float(s);
@@ -12094,14 +12091,11 @@ static void llama_set_inputs(llama_context & lctx, const llama_batch & batch) {
@@ -13879,14 +13876,11 @@ static void llama_set_inputs(llama_context & lctx, const llama_batch & batch) {
GGML_ASSERT(ggml_backend_buffer_is_host(lctx.inp_cls->buffer));
uint32_t * data = (uint32_t *) lctx.inp_cls->data;