mirror of
https://git.adityakumar.xyz/llama.cpp.git
synced 2024-11-09 15:29:43 +00:00
ggml : fix Q4_3 cuBLAS
This commit is contained in:
parent
857308d1e8
commit
0e018fe008
2 changed files with 5 additions and 0 deletions
|
@ -312,6 +312,7 @@ add_library(ggml OBJECT
|
||||||
target_include_directories(ggml PUBLIC .)
|
target_include_directories(ggml PUBLIC .)
|
||||||
target_compile_features(ggml PUBLIC c_std_11) # don't bump
|
target_compile_features(ggml PUBLIC c_std_11) # don't bump
|
||||||
target_link_libraries(ggml PUBLIC Threads::Threads ${LLAMA_EXTRA_LIBS})
|
target_link_libraries(ggml PUBLIC Threads::Threads ${LLAMA_EXTRA_LIBS})
|
||||||
|
|
||||||
if (BUILD_SHARED_LIBS)
|
if (BUILD_SHARED_LIBS)
|
||||||
set_target_properties(ggml PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
set_target_properties(ggml PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||||
endif()
|
endif()
|
||||||
|
@ -324,6 +325,7 @@ add_library(llama
|
||||||
target_include_directories(llama PUBLIC .)
|
target_include_directories(llama PUBLIC .)
|
||||||
target_compile_features(llama PUBLIC cxx_std_11) # don't bump
|
target_compile_features(llama PUBLIC cxx_std_11) # don't bump
|
||||||
target_link_libraries(llama PRIVATE ggml ${LLAMA_EXTRA_LIBS})
|
target_link_libraries(llama PRIVATE ggml ${LLAMA_EXTRA_LIBS})
|
||||||
|
|
||||||
if (BUILD_SHARED_LIBS)
|
if (BUILD_SHARED_LIBS)
|
||||||
set_target_properties(llama PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
set_target_properties(llama PROPERTIES POSITION_INDEPENDENT_CODE ON)
|
||||||
target_compile_definitions(llama PRIVATE LLAMA_SHARED LLAMA_BUILD)
|
target_compile_definitions(llama PRIVATE LLAMA_SHARED LLAMA_BUILD)
|
||||||
|
|
3
ggml.c
3
ggml.c
|
@ -7992,6 +7992,9 @@ static void ggml_compute_forward_mul_mat_q_f32(
|
||||||
else if (type == GGML_TYPE_Q4_2) {
|
else if (type == GGML_TYPE_Q4_2) {
|
||||||
dequantize_row_q_cuda = dequantize_row_q4_2_cuda;
|
dequantize_row_q_cuda = dequantize_row_q4_2_cuda;
|
||||||
}
|
}
|
||||||
|
else if (type == GGML_TYPE_Q4_3) {
|
||||||
|
dequantize_row_q_cuda = dequantize_row_q4_3_cuda;
|
||||||
|
}
|
||||||
else {
|
else {
|
||||||
GGML_ASSERT(false);
|
GGML_ASSERT(false);
|
||||||
}
|
}
|
||||||
|
|
Loading…
Reference in a new issue