Fix build

Author: howard0su

ggml-cuda.h

@@ -15,7 +15,7 @@ void   ggml_cuda_mul_mat(const struct ggml_tensor * src0, const struct ggml_tens
 void * ggml_cuda_host_malloc(size_t size);
 void   ggml_cuda_host_free(void * ptr);
 
-void ggml_cuda_transform_tensor(struct ggml_tensor * tensor);
+void ggml_cuda_transform_tensor(const void * data, struct ggml_tensor * tensor);
 
 #ifdef  __cplusplus
 }