diff --git a/backends/vulkan/runtime/api/containers/Tensor.cpp b/backends/vulkan/runtime/api/containers/Tensor.cpp
index 900854ccd75..8c76c11532b 100644
--- a/backends/vulkan/runtime/api/containers/Tensor.cpp
+++ b/backends/vulkan/runtime/api/containers/Tensor.cpp
@@ -478,13 +478,6 @@ vTensor::vTensor(
   if (storage_type != utils::kBuffer) {
     set_logical_limits(storage_.image_extents_);
   }
-
-  if (dtype == vkapi::kHalf) {
-    VK_CHECK_COND(
-        api::context()->adapter_ptr()->supports_16bit_storage_buffers(),
-        "Half dtype is only available if the physical device supports float16 "
-        "storage buffers!");
-  }
 }
 
 // NOLINTNEXTLINE
diff --git a/backends/vulkan/runtime/graph/ops/PrepackNode.cpp b/backends/vulkan/runtime/graph/ops/PrepackNode.cpp
index e27723468ab..bf501296b1b 100644
--- a/backends/vulkan/runtime/graph/ops/PrepackNode.cpp
+++ b/backends/vulkan/runtime/graph/ops/PrepackNode.cpp
@@ -68,6 +68,8 @@ api::StagingBuffer PrepackNode::create_staging_buffer(ComputeGraph* graph) {
 void PrepackNode::encode(ComputeGraph* graph) {
   api::Context* const context = graph->context();
 
+  context->check_device_capabilities(shader_);
+
   vTensorPtr packed = graph->get_tensor(packed_);
   api::StagingBuffer staging = create_staging_buffer(graph);
 
diff --git a/backends/vulkan/test/op_tests/utils/gen_computegraph.py b/backends/vulkan/test/op_tests/utils/gen_computegraph.py
index 472127ffe2e..6f93e662076 100644
--- a/backends/vulkan/test/op_tests/utils/gen_computegraph.py
+++ b/backends/vulkan/test/op_tests/utils/gen_computegraph.py
@@ -667,7 +667,6 @@ def gen_op_check_fn(self) -> str:
             op_check_fn = self.gen_decl(f"prepacked_check_{op_name}") + " {\n"
 
         op_check_fn_body = ""
-        op_check_fn_body += self.gen_conditional_skips()
         op_check_fn_body += self.gen_graph_build_code()
         op_check_fn_body += self.gen_graph_exec_code()