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 6be1fda commit ce2afc5Copy full SHA for ce2afc5
src/llama.cpp
@@ -6951,6 +6951,7 @@ static bool llm_load_tensors(
6951
model.buft_layer[i] = llama_default_buffer_type_cpu(true);
6952
}
6953
6954
+#ifndef GGML_USE_KOMPUTE
6955
if (split_mode == LLAMA_SPLIT_MODE_LAYER) {
6956
// calculate the split points
6957
int device_count = llama_get_device_count(model);
@@ -6988,7 +6989,9 @@ static bool llm_load_tensors(
6988
6989
} else {
6990
model.buft_output = llama_default_buffer_type_cpu(true);
6991
- } else {
6992
+ } else
6993
+#endif
6994
+ {
6995
ggml_backend_buffer_type_t split_buft;
6996
if (split_mode == LLAMA_SPLIT_MODE_ROW) {
6997
split_buft = llama_default_buffer_type_split(model, main_gpu, tensor_split);
0 commit comments