mirrors/ollama-for-amd
1
0
Fork 0
mirror of https://github.com/likelovewant/ollama-for-amd.git synced 2025-12-22 23:03:55 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama: update llama.cpp vendor code to commit d7cfe1ff (#9356)

Browse Source
This commit is contained in:
Jeffrey Morgan
2025-02-26 20:34:44 -08:00
committed by GitHub
parent 2db96c18e7
commit d7d7e99662
149 changed files with 18215 additions and 11009 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
llama/patches/0003-embeddings.patch
View File

@@ -9,10 +9,10 @@ Subject: [PATCH] embeddings
2 files changed, 5 insertions(+), 3 deletions(-)
diff --git a/src/llama-context.cpp b/src/llama-context.cpp
index 38a55fb2..b9c4a5bf 100644
index 671d2a81..47e79ed4 100644
--- a/src/llama-context.cpp
+++ b/src/llama-context.cpp
@@ -475,7 +475,7 @@ size_t llama_output_reserve(struct llama_context & lctx, size_t n_outputs) {
@@ -479,7 +479,7 @@ size_t llama_output_reserve(struct llama_context & lctx, size_t n_outputs) {
const auto n_embd = hparams.n_embd;
// TODO: use a per-batch flag for logits presence instead
@@ -22,10 +22,10 @@ index 38a55fb2..b9c4a5bf 100644
const size_t logits_size = has_logits ? n_vocab*n_outputs_max : 0;
diff --git a/src/llama.cpp b/src/llama.cpp
index ea78ea48..4eb3f6b9 100644
index 607f2786..ac85bfed 100644
--- a/src/llama.cpp
+++ b/src/llama.cpp
@@ -10876,7 +10876,6 @@ static int llama_decode_internal(
@@ -8652,7 +8652,6 @@ static int llama_decode_impl(
res = nullptr;
embd = nullptr;
} else if (cparams.embeddings) {
@@ -33,7 +33,7 @@ index ea78ea48..4eb3f6b9 100644
embd = nullptr;
for (int i = ggml_graph_n_nodes(gf) - 1; i >= 0; --i) {
if (strcmp(ggml_graph_node(gf, i)->name, "result_embd_pooled") == 0) {
@@ -10884,12 +10883,15 @@ static int llama_decode_internal(
@@ -8660,12 +8659,15 @@ static int llama_decode_impl(
break;
}
}