[vulkan] Add nonVarTypeModeGuard to vulkan tests and speed_benchmark_…

…torch (pytorch#52535) Summary: Pull Request resolved: pytorch#52535 Test Plan: Imported from OSS Reviewed By: ailzhang Differential Revision: D26580994 Pulled By: SS-JIA fbshipit-source-id: 94f091432265cf6607b73c34846c07273d47c70b
zmmwl · Feb 25, 2021 · c2558b4 · c2558b4
1 parent e94940b
commit c2558b4
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 0 deletions.
diff --git a/aten/src/ATen/test/vulkan_api_test.cpp b/aten/src/ATen/test/vulkan_api_test.cpp
@@ -70,6 +70,7 @@ TEST(VulkanAPITest, adaptive_avg_pool2d) {
   if (!at::is_vulkan_available()) {
     return;
   }
+  at::AutoNonVariableTypeMode nonVarTypeModeGuard(true);
 
   const auto in_cpu = at::rand({5, 7, 47, 31}, at::TensorOptions(at::kCPU).dtype(at::kFloat));
   const auto out_cpu = at::adaptive_avg_pool2d(in_cpu, {3, 3});
@@ -617,6 +618,7 @@ TEST(VulkanAPITest, reshape) {
   if (!at::is_vulkan_available()) {
     return;
   }
+  at::AutoNonVariableTypeMode nonVarTypeModeGuard(true);
 
   const auto in_cpu = at::rand({47, 11, 83, 97}, at::device(at::kCPU).dtype(at::kFloat));
   const auto in_vulkan = in_cpu.vulkan();
@@ -638,6 +640,7 @@ TEST(VulkanAPITest, reshape_) {
   if (!at::is_vulkan_available()) {
     return;
   }
+  at::AutoNonVariableTypeMode nonVarTypeModeGuard(true);
 
   const auto cpu = at::rand({59, 41, 19, 67}, at::device(at::kCPU).dtype(at::kFloat));
   const auto vulkan = cpu.vulkan();

diff --git a/binaries/compare_models_torch.cc b/binaries/compare_models_torch.cc
@@ -224,6 +224,7 @@ int main(int argc, char** argv) {
   float tolerance = 0;
   ss >> tolerance;
 
+  at::AutoNonVariableTypeMode nonVarTypeModeGuard(true);
   torch::autograd::AutoGradMode guard(false);
   torch::jit::GraphOptimizerEnabledGuard no_optimizer_guard(false);
   auto module = torch::jit::load(FLAGS_model);

diff --git a/binaries/speed_benchmark_torch.cc b/binaries/speed_benchmark_torch.cc
@@ -209,6 +209,7 @@ int main(int argc, char** argv) {
 
   std::vector<c10::IValue> inputs = create_inputs();
 
+  at::AutoNonVariableTypeMode nonVarTypeModeGuard(true);
   torch::autograd::AutoGradMode guard(false);
   torch::jit::GraphOptimizerEnabledGuard no_optimizer_guard(false);
   auto module = torch::jit::load(FLAGS_model);