24 lines
757 B
C++
24 lines
757 B
C++
#include "ggml-openvino.h"
|
|
#include "ggml-backend-impl.h"
|
|
#include "ggml-impl.h"
|
|
|
|
// backend API
|
|
GGML_API ggml_backend_t ggml_backend_openvino_init(int device) {}
|
|
|
|
GGML_API bool ggml_backend_is_openvino(ggml_backend_t backend) {}
|
|
|
|
// device buffer
|
|
GGML_API ggml_backend_buffer_type_t
|
|
ggml_backend_openvino_buffer_type(int device) {}
|
|
|
|
// split tensor buffer that splits matrices by rows across multiple devices
|
|
GGML_API ggml_backend_buffer_type_t
|
|
ggml_backend_openvino_split_buffer_type(const float *tensor_split) {}
|
|
|
|
// pinned host buffer for use with the CPU backend for faster copies between CPU
|
|
// and GPU
|
|
GGML_API ggml_backend_buffer_type_t
|
|
ggml_backend_openvino_host_buffer_type(void) {}
|
|
|
|
GGML_API ggml_backend_reg_t ggml_backend_openvino_reg(void) {}
|