We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent 8178622 commit 1ce8fb4Copy full SHA for 1ce8fb4
src/llama.cpp
@@ -5972,6 +5972,7 @@ static bool llm_load_tensors(
5972
model.buft_layer[i] = llama_default_buffer_type_cpu(true);
5973
}
5974
5975
+#ifndef GGML_USE_KOMPUTE
5976
if (split_mode == LLAMA_SPLIT_MODE_LAYER) {
5977
// calculate the split points
5978
int device_count = llama_get_device_count(model);
@@ -6009,7 +6010,9 @@ static bool llm_load_tensors(
6009
6010
} else {
6011
model.buft_output = llama_default_buffer_type_cpu(true);
6012
- } else {
6013
+ } else
6014
+#endif
6015
+ {
6016
ggml_backend_buffer_type_t split_buft;
6017
if (split_mode == LLAMA_SPLIT_MODE_ROW) {
6018
split_buft = llama_default_buffer_type_split(model, main_gpu, tensor_split);
0 commit comments