diff --git a/ggml/src/ggml-cuda/fattn.cu b/ggml/src/ggml-cuda/fattn.cu
index 79ee204d195..b24479ec705 100644
--- a/ggml/src/ggml-cuda/fattn.cu
+++ b/ggml/src/ggml-cuda/fattn.cu
@@ -287,6 +287,7 @@ static void ggml_cuda_flash_attn_ext_vec(ggml_backend_cuda_context & ctx, ggml_t
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_Q8_0, GGML_TYPE_BF16)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_BF16, GGML_TYPE_BF16)
 
+ #ifndef GGML_USE_HIP
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_TQ3_0, GGML_TYPE_TQ3_0)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_F16,  GGML_TYPE_TQ3_0)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_Q4_0, GGML_TYPE_TQ3_0)
@@ -297,11 +298,14 @@ static void ggml_cuda_flash_attn_ext_vec(ggml_backend_cuda_context & ctx, ggml_t
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_TQ3_0, GGML_TYPE_Q4_0)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_TQ3_0, GGML_TYPE_Q8_0)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_TQ3_0, GGML_TYPE_BF16)
+#endif // GGML_USE_HIP
 #else
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_F16,  GGML_TYPE_F16)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_Q4_0, GGML_TYPE_Q4_0)
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_Q8_0, GGML_TYPE_Q8_0)
+#ifndef GGML_USE_HIP
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_TQ3_0, GGML_TYPE_TQ3_0)
+#endif // GGML_USE_HIP
     FATTN_VEC_CASES_ALL_D(GGML_TYPE_BF16, GGML_TYPE_BF16)
 #endif // GGML_CUDA_FA_ALL_QUANTS
 
diff --git a/ggml/src/ggml-cuda/gated_delta_net.cu b/ggml/src/ggml-cuda/gated_delta_net.cu
index ac8c8c205a4..ded13ee1b66 100644
--- a/ggml/src/ggml-cuda/gated_delta_net.cu
+++ b/ggml/src/ggml-cuda/gated_delta_net.cu
@@ -1,5 +1,7 @@
 #include "gated_delta_net.cuh"
+#ifndef GGML_USE_HIP
 #include <cuda_fp16.h>
+#endif
 #include <type_traits>
 
 // Tree-mode parent index sentinel: a node whose parent is the pre-block state
diff --git a/ggml/src/ggml-cuda/vendors/hip.h b/ggml/src/ggml-cuda/vendors/hip.h
index 898fec31e36..f187dfbc9e6 100644
--- a/ggml/src/ggml-cuda/vendors/hip.h
+++ b/ggml/src/ggml-cuda/vendors/hip.h
@@ -45,6 +45,7 @@
 #define cublasGemmEx hipblasGemmEx
 #define cublasGemmBatchedEx hipblasGemmBatchedEx
 #define cublasGemmStridedBatchedEx hipblasGemmStridedBatchedEx
+#define cublasSgemmStridedBatched hipblasSgemmStridedBatched
 #define cublasHandle_t hipblasHandle_t
 #define cublasSetMathMode(handle, mode) CUBLAS_STATUS_SUCCESS
 #define cublasSetStream hipblasSetStream
@@ -140,6 +141,9 @@
 #define cudaGraphExecUpdate hipGraphExecUpdate
 #define cudaStreamCaptureModeRelaxed hipStreamCaptureModeRelaxed
 #define cudaStreamBeginCapture hipStreamBeginCapture
+#define cudaStreamCaptureStatus hipStreamCaptureStatus
+#define cudaStreamCaptureStatusNone hipStreamCaptureStatusNone
+#define cudaStreamIsCapturing hipStreamIsCapturing
 #define cudaGraph_t hipGraph_t
 #define cudaStream_t hipStream_t
 #define cudaSuccess hipSuccess