Fix some build error for ROCM

1. Fix terrible logic in CMakeLists.txt
2. using the correct typedef for hip
Signed-off-by: onepick <jiajuku12@163.com>

csrc/ktransformers_ext/CMakeLists.txt

@@ -248,9 +248,9 @@ if (WIN32)
     include_directories("$ENV{CUDA_PATH}/include")
     add_compile_definitions(KTRANSFORMERS_USE_CUDA=1)
 elseif (UNIX)
-    if (NOT KTRANSFORMERS_USE_MUSA)
-        # find_package(CUDA REQUIRED)
-        # include_directories("${CUDA_INCLUDE_DIRS}")
+    if (KTRANSFORMERS_USE_CUDA)
+        find_package(CUDA REQUIRED)
+        include_directories("${CUDA_INCLUDE_DIRS}")
         include(CheckLanguage)
         check_language(CUDA)
         if(CMAKE_CUDA_COMPILER)
@@ -324,7 +324,7 @@ target_link_libraries(${PROJECT_NAME} PRIVATE llama)
 if(WIN32)
     target_link_libraries(${PROJECT_NAME} PRIVATE "$ENV{CUDA_PATH}/lib/x64/cudart.lib")#CUDA::cudart
 elseif(UNIX)
-    if(NOT KTRANSFORMERS_USE_MUSA)
+    if(KTRANSFORMERS_USE_CUDA)
         target_link_libraries(${PROJECT_NAME} PRIVATE "${CUDAToolkit_LIBRARY_DIR}/libcudart.so")
     endif()
     if (KTRANSFORMERS_USE_ROCM)

csrc/ktransformers_ext/cuda/custom_gguf/dequant.cu

@@ -17,7 +17,7 @@
 #include <c10/cuda/CUDAGuard.h>
 
 #ifdef __HIP_PLATFORM_AMD__
-typedef hip_bfloat16 nv_bfloat16;
+typedef __hip_bfloat16 nv_bfloat16;
 #endif
 
 __global__ void dequantize_q8_0_fp32_kernel(const int8_t* data, float* output, const int blk_size, const int ele_per_blk, const int num_blocks) {
@@ -879,4 +879,4 @@ torch::Tensor dequantize_iq4_xs(const int8_t* data, const int num_bytes, const i
     }
     cudaDeviceSynchronize();
     return output;
-}
\ No newline at end of file
+}

csrc/ktransformers_ext/ext_bindings.cpp

@@ -9,7 +9,9 @@
  **/
 // Python bindings
 #include "cpu_backend/cpuinfer.h"
+#ifndef KTRANSFORMERS_USE_ROCM
 #include "device_launch_parameters.h"
+#endif
 #include "llamafile/flags.h"
 #include "operators/kvcache/kvcache.h"
 #include "operators/llamafile/linear.h"