[codemod][fbcode/caffe2] Apply clang-format update fixes

Test Plan: Sandcastle and visual inspection.

Reviewed By: igorsugak

Differential Revision: D25849205

fbshipit-source-id: ef664c1ad4b3ee92d5c020a5511b4ef9837a09a0
diff --git a/test/cpp/jit/test_alias_analysis.cpp b/test/cpp/jit/test_alias_analysis.cpp
index e700ee5..58078c9 100644
--- a/test/cpp/jit/test_alias_analysis.cpp
+++ b/test/cpp/jit/test_alias_analysis.cpp
@@ -906,14 +906,15 @@
 }
 
 TEST(WildcardsTest, Basic) {
-  RegisterOperators reg({Operator(
-                             "prim::returns_wildcard(Tensor a) -> Tensor(*)",
-                             [](Stack* stack) {},
-                             aliasAnalysisFromSchema()),
-                         Operator(
-                             "prim::writes(Tensor(z!) a) -> Tensor(a)",
-                             [](Stack* stack) {},
-                             aliasAnalysisFromSchema())});
+  RegisterOperators reg(
+      {Operator(
+           "prim::returns_wildcard(Tensor a) -> Tensor(*)",
+           [](Stack* stack) {},
+           aliasAnalysisFromSchema()),
+       Operator(
+           "prim::writes(Tensor(z!) a) -> Tensor(a)",
+           [](Stack* stack) {},
+           aliasAnalysisFromSchema())});
   const auto returns_wildcard =
       Symbol::fromQualString("prim::returns_wildcard");
   const auto writes = Symbol::fromQualString("prim::writes");
diff --git a/test/cpp/jit/test_argument_spec.cpp b/test/cpp/jit/test_argument_spec.cpp
index bf40761..b653abe 100644
--- a/test/cpp/jit/test_argument_spec.cpp
+++ b/test/cpp/jit/test_argument_spec.cpp
@@ -50,21 +50,23 @@
   auto const GF = at::CUDA(at::kFloat);
   auto const GD = at::CUDA(at::kDouble);
 
-  auto list = createStack({var(CF, {1}, true),
-                           var(CD, {1, 2}, false),
-                           var(GF, {}, true),
-                           var(GD, {4, 5, 6}, false),
-                           undef()});
+  auto list = createStack(
+      {var(CF, {1}, true),
+       var(CD, {1, 2}, false),
+       var(GF, {}, true),
+       var(GD, {4, 5, 6}, false),
+       undef()});
 
   // make sure we have some non-standard strides
   list[1].toTensor().transpose_(0, 1);
 
   // same list but different backing values
-  auto list2 = createStack({var(CF, {1}, true),
-                            var(CD, {1, 2}, false),
-                            var(GF, {}, true),
-                            var(GD, {4, 5, 6}, false),
-                            undef()});
+  auto list2 = createStack(
+      {var(CF, {1}, true),
+       var(CD, {1, 2}, false),
+       var(GF, {}, true),
+       var(GD, {4, 5, 6}, false),
+       undef()});
   list2[1].toTensor().transpose_(0, 1);
 
   CompleteArgumentSpec a(true, list);
@@ -142,21 +144,23 @@
 
   ArgumentSpecCreator arg_spec_creator(*graph);
 
-  auto list = createStack({var(CF, {1}, true),
-                           var(CD, {1, 2}, false),
-                           var(GF, {}, true),
-                           var(GD, {4, 5, 6}, false),
-                           undef()});
+  auto list = createStack(
+      {var(CF, {1}, true),
+       var(CD, {1, 2}, false),
+       var(GF, {}, true),
+       var(GD, {4, 5, 6}, false),
+       undef()});
 
   // make sure we have some non-standard strides
   list[1].toTensor().transpose_(0, 1);
 
   // same list but different backing values
-  auto list2 = createStack({var(CF, {1}, true),
-                            var(CD, {1, 2}, false),
-                            var(GF, {}, true),
-                            var(GD, {4, 5, 6}, false),
-                            undef()});
+  auto list2 = createStack(
+      {var(CF, {1}, true),
+       var(CD, {1, 2}, false),
+       var(GF, {}, true),
+       var(GD, {4, 5, 6}, false),
+       undef()});
   list2[1].toTensor().transpose_(0, 1);
 
   ArgumentSpec a = arg_spec_creator.create(true, list);
diff --git a/test/cpp/jit/test_gpu.cpp b/test/cpp/jit/test_gpu.cpp
index f8d25c4..5086cb5 100644
--- a/test/cpp/jit/test_gpu.cpp
+++ b/test/cpp/jit/test_gpu.cpp
@@ -2681,12 +2681,13 @@
   using OpTuple = std::tuple<AtenFuncSig, BinaryOpType, std::string>;
 
   // see [Note: explicit tuple type for uniform initialization list]
-  std::vector<OpTuple> logic_ops{OpTuple{at::eq, BinaryOpType::Eq, "eq"},
-                                 OpTuple{at::ge, BinaryOpType::GE, "ge"},
-                                 OpTuple{at::gt, BinaryOpType::GT, "gt"},
-                                 OpTuple{at::le, BinaryOpType::LE, "le"},
-                                 OpTuple{at::lt, BinaryOpType::LT, "lt"},
-                                 OpTuple{at::ne, BinaryOpType::NE, "ne"}};
+  std::vector<OpTuple> logic_ops{
+      OpTuple{at::eq, BinaryOpType::Eq, "eq"},
+      OpTuple{at::ge, BinaryOpType::GE, "ge"},
+      OpTuple{at::gt, BinaryOpType::GT, "gt"},
+      OpTuple{at::le, BinaryOpType::LE, "le"},
+      OpTuple{at::lt, BinaryOpType::LT, "lt"},
+      OpTuple{at::ne, BinaryOpType::NE, "ne"}};
 
   std::for_each(logic_ops.begin(), logic_ops.end(), [](OpTuple& op) {
     test_op(
@@ -4184,13 +4185,14 @@
   sub_tv3->computeAt(sum_exp_rf_tv9, -1);
   sub_tv3_copy->computeAt(output_tv7, -1);
 
-  TensorView* tensors_to_parallelize[] = {max_val_tv1,
-                                          bcast_max_tv2,
-                                          sum_exp_tv5,
-                                          bcast_sum_tv6,
-                                          output_tv7,
-                                          max_val_rf_tv8,
-                                          sum_exp_rf_tv9};
+  TensorView* tensors_to_parallelize[] = {
+      max_val_tv1,
+      bcast_max_tv2,
+      sum_exp_tv5,
+      bcast_sum_tv6,
+      output_tv7,
+      max_val_rf_tv8,
+      sum_exp_rf_tv9};
 
   for (auto tv : tensors_to_parallelize) {
     tv->axis(-1)->parallelize(ParallelType::TIDx);
@@ -4318,13 +4320,14 @@
   sub_tv3->computeAt(sum_exp_rf_tv9, -1);
   sub_tv3_copy->computeAt(output_tv7, -1);
 
-  TensorView* tensors_to_parallelize[] = {max_val_tv1,
-                                          bcast_max_tv2,
-                                          sum_exp_tv5,
-                                          bcast_sum_tv6,
-                                          output_tv7,
-                                          max_val_rf_tv8,
-                                          sum_exp_rf_tv9};
+  TensorView* tensors_to_parallelize[] = {
+      max_val_tv1,
+      bcast_max_tv2,
+      sum_exp_tv5,
+      bcast_sum_tv6,
+      output_tv7,
+      max_val_rf_tv8,
+      sum_exp_rf_tv9};
 
   for (auto tv : tensors_to_parallelize) {
     tv->axis(0)->parallelize(ParallelType::BIDx);
@@ -5931,15 +5934,16 @@
   cache_x->setMemoryType(MemoryType::Shared);
   exp->setMemoryType(MemoryType::Shared);
 
-  std::vector<TensorView*> all_tensors({x,
-                                        cache_x,
-                                        max_val,
-                                        bcast_max,
-                                        x_max_sub,
-                                        exp,
-                                        sum_exp,
-                                        bcast_sum,
-                                        softmax});
+  std::vector<TensorView*> all_tensors(
+      {x,
+       cache_x,
+       max_val,
+       bcast_max,
+       x_max_sub,
+       exp,
+       sum_exp,
+       bcast_sum,
+       softmax});
 
   auto tidx = new Int();
   fusion.addInput(tidx);
@@ -6168,25 +6172,27 @@
   std::vector<TensorView*> common_tensors(
       {x_sum, x_sum_bcast, x_mean, var_sum, var_sum_bcast, var, var_eps, rvar});
 
-  std::vector<TensorView*> static_tensors({sx,
-                                           sx_cache,
-                                           sx_sum,
-                                           sx_mean_sub,
-                                           sx_mean_sub_pow,
-                                           sx_var_sum,
-                                           sx_norm,
-                                           sx_norm_gamma,
-                                           sx_norm_gamma_beta});
+  std::vector<TensorView*> static_tensors(
+      {sx,
+       sx_cache,
+       sx_sum,
+       sx_mean_sub,
+       sx_mean_sub_pow,
+       sx_var_sum,
+       sx_norm,
+       sx_norm_gamma,
+       sx_norm_gamma_beta});
 
-  std::vector<TensorView*> dynamic_tensors({dx,
-                                            dx_cache,
-                                            dx_sum,
-                                            dx_mean_sub,
-                                            dx_mean_sub_pow,
-                                            dx_var_sum,
-                                            dx_norm,
-                                            dx_norm_gamma,
-                                            dx_norm_gamma_beta});
+  std::vector<TensorView*> dynamic_tensors(
+      {dx,
+       dx_cache,
+       dx_sum,
+       dx_mean_sub,
+       dx_mean_sub_pow,
+       dx_var_sum,
+       dx_norm,
+       dx_norm_gamma,
+       dx_norm_gamma_beta});
 
   std::vector<TensorView*> all_tensors;
   all_tensors.insert(
@@ -6309,20 +6315,21 @@
   cache_x->setMemoryType(MemoryType::Shared);
   x_mean_sub->setMemoryType(MemoryType::Shared);
 
-  std::vector<TensorView*> all_tensors({x_sum,
-                                        x_mean,
-                                        cache_x,
-                                        x_sum_bcast,
-                                        x_mean_sub,
-                                        x_mean_sub_pow,
-                                        var_sum,
-                                        var_sum_bcast,
-                                        var,
-                                        var_eps,
-                                        rvar,
-                                        norm,
-                                        norm_gamma,
-                                        norm_gamma_beta});
+  std::vector<TensorView*> all_tensors(
+      {x_sum,
+       x_mean,
+       cache_x,
+       x_sum_bcast,
+       x_mean_sub,
+       x_mean_sub_pow,
+       var_sum,
+       var_sum_bcast,
+       var,
+       var_eps,
+       rvar,
+       norm,
+       norm_gamma,
+       norm_gamma_beta});
 
   auto tidx = new Int();
   fusion.addInput(tidx);
diff --git a/test/cpp/jit/test_misc.cpp b/test/cpp/jit/test_misc.cpp
index 5fb10d5..eaec2fe 100644
--- a/test/cpp/jit/test_misc.cpp
+++ b/test/cpp/jit/test_misc.cpp
@@ -1044,7 +1044,7 @@
   ids.clear();
   { // START: global test
     addGlobalCallback(RecordFunctionCallback(
-        [](const RecordFunction &
+        [](const RecordFunction&
            /* unused */) -> std::unique_ptr<at::ObserverContext> {
           auto ctx = std::make_unique<TestContext>();
           ctx->a = 123;
@@ -1070,7 +1070,7 @@
       const int test_val = 234;
       const std::string test_str = "test thread str";
       addThreadLocalCallback(RecordFunctionCallback(
-          [](const RecordFunction &
+          [](const RecordFunction&
              /* unused */) -> std::unique_ptr<at::ObserverContext> {
             auto ctx = std::make_unique<TestContext>();
             ctx->a = 234;
diff --git a/test/cpp/tensorexpr/test_conv.cpp b/test/cpp/tensorexpr/test_conv.cpp
index c047c7b..f49cac7 100644
--- a/test/cpp/tensorexpr/test_conv.cpp
+++ b/test/cpp/tensorexpr/test_conv.cpp
@@ -75,9 +75,10 @@
 
   at::Tensor result = at::empty_like(ref);
   te::SimpleIREvaluator cg(s, {inputB, filterB, conv});
-  cg.call({input.data_ptr<float>(),
-           filter.data_ptr<float>(),
-           result.data_ptr<float>()});
+  cg.call(
+      {input.data_ptr<float>(),
+       filter.data_ptr<float>(),
+       result.data_ptr<float>()});
 
   ASSERT_TRUE(at::allclose(ref, result, 1e-3, 1e-3));
 }
diff --git a/test/cpp/tensorexpr/test_llvm.cpp b/test/cpp/tensorexpr/test_llvm.cpp
index 5201416..7afb839 100644
--- a/test/cpp/tensorexpr/test_llvm.cpp
+++ b/test/cpp/tensorexpr/test_llvm.cpp
@@ -294,15 +294,15 @@
   std::vector<void*> args({v.data()});
   VarHandle x("x", kByte);
   VarHandle y("y", kHalf);
-  auto block =
-      Block::make({Let::make(x, 3),
-                   Let::make(y, 6.f),
-                   a.store(
-                       {0},
-                       Cast::make(
-                           kDouble,
-                           ExprHandle(2.f) +
-                               (x * ExprHandle(3.f) + y * ExprHandle(4.f))))});
+  auto block = Block::make(
+      {Let::make(x, 3),
+       Let::make(y, 6.f),
+       a.store(
+           {0},
+           Cast::make(
+               kDouble,
+               ExprHandle(2.f) +
+                   (x * ExprHandle(3.f) + y * ExprHandle(4.f))))});
 
   LLVMCodeGen cg(block, {a});
   ASSERT_EQ(cg.value<int>(args), 0);
diff --git a/test/cpp/tensorexpr/test_loopnest.cpp b/test/cpp/tensorexpr/test_loopnest.cpp
index 7294aa5..89ad7eb 100644
--- a/test/cpp/tensorexpr/test_loopnest.cpp
+++ b/test/cpp/tensorexpr/test_loopnest.cpp
@@ -2606,8 +2606,9 @@
       x,
       0,
       kTotalSize,
-      Block::make({Store::make(a_buf, {x}, x * 2),
-                   Store::make(b_buf, {x}, Load::make(a_buf, {x}, 1))}));
+      Block::make(
+          {Store::make(a_buf, {x}, x * 2),
+           Store::make(b_buf, {x}, Load::make(a_buf, {x}, 1))}));
   Block::make({f});
   Stmt* unrolled = nullptr;
   LoopNest::unroll(f, &unrolled);
@@ -2658,9 +2659,10 @@
       x,
       0,
       kTotalSize,
-      Block::make({Let::make(e, 7),
-                   Store::make(a_buf, {x}, e),
-                   Store::make(b_buf, {x}, e + 1)}));
+      Block::make(
+          {Let::make(e, 7),
+           Store::make(a_buf, {x}, e),
+           Store::make(b_buf, {x}, e + 1)}));
   Block::make({f});
   Stmt* unrolled = nullptr;
   LoopNest::unroll(f, &unrolled);
@@ -2700,9 +2702,9 @@
   BufHandle a_buf("A", {ExprHandle(kTotalSize)}, kInt);
   BufHandle b_buf("B", {ExprHandle(kTotalSize)}, kInt);
   VarHandle x("x", kInt);
-  auto for_body =
-      Block::make({Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
-                   Store::make(b_buf, {x}, x * 2)});
+  auto for_body = Block::make(
+      {Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
+       Store::make(b_buf, {x}, x * 2)});
   auto for_stmt = For::make(x, 50, 100, for_body);
   Block::make({for_stmt});
 
@@ -2768,9 +2770,9 @@
   BufHandle a_buf("A", {ExprHandle(kTotalSize)}, kInt);
   BufHandle b_buf("B", {ExprHandle(kTotalSize)}, kInt);
   VarHandle x("x", kInt);
-  auto for_body =
-      Block::make({Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
-                   Store::make(b_buf, {x}, x * 2)});
+  auto for_body = Block::make(
+      {Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
+       Store::make(b_buf, {x}, x * 2)});
   auto for_stmt = For::make(x, 0, 100, for_body);
   Block::make({for_stmt});
 
@@ -2803,9 +2805,9 @@
   BufHandle b_buf("B", {ExprHandle(kTotalSize)}, kInt);
   VarHandle x("x", kInt);
   VarHandle y("y", kInt);
-  auto for_body =
-      Block::make({Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
-                   Store::make(b_buf, {x}, x * 2)});
+  auto for_body = Block::make(
+      {Store::make(a_buf, {x}, Load::make(kInt, b_buf, {x}, 1), 1),
+       Store::make(b_buf, {x}, x * 2)});
   auto for_stmt = For::make(x, y, 100, for_body);
   Block::make({for_stmt});
 
diff --git a/test/cpp/tensorexpr/test_memdependency.cpp b/test/cpp/tensorexpr/test_memdependency.cpp
index cfb3388..7105b9d 100644
--- a/test/cpp/tensorexpr/test_memdependency.cpp
+++ b/test/cpp/tensorexpr/test_memdependency.cpp
@@ -1990,13 +1990,13 @@
     MemDependencyChecker analyzer({a}, {c});
     Store* initStore = Store::make(c, {x}, Load::make(a, {x}, 1), 1);
     ExprHandle conditionalLoad = Load::make(c, {0}, 1);
-    Stmt* stmt =
-        Block::make({For::make(x, 0, 10, initStore),
-                     Cond::make(
-                         CompareSelect::make(
-                             conditionalLoad, 5, CompareSelectOperation::kLT),
-                         Store::make(c, {0}, 5, 1),
-                         nullptr)});
+    Stmt* stmt = Block::make(
+        {For::make(x, 0, 10, initStore),
+         Cond::make(
+             CompareSelect::make(
+                 conditionalLoad, 5, CompareSelectOperation::kLT),
+             Store::make(c, {0}, 5, 1),
+             nullptr)});
 
     stmt->accept(&analyzer);
 
@@ -2177,11 +2177,12 @@
         Store::make(b, {x}, Add::make(Load::make(b, {x}, 1), 1), 3);
     For* secondLoop = For::make(x, 4, 7, secondStore);
 
-    Stmt* stmt = Block::make({firstLoop,
-                              secondLoop,
-                              Store::make(b, {4}, 100, 1),
-                              Store::make(b, {5}, 101, 1),
-                              Store::make(b, {6}, 102, 1)});
+    Stmt* stmt = Block::make(
+        {firstLoop,
+         secondLoop,
+         Store::make(b, {4}, 100, 1),
+         Store::make(b, {5}, 101, 1),
+         Store::make(b, {6}, 102, 1)});
 
     stmt->accept(&analyzer);
 
diff --git a/test/cpp/tensorexpr/test_registerizer.cpp b/test/cpp/tensorexpr/test_registerizer.cpp
index 0b8a5db..bfddb28 100644
--- a/test/cpp/tensorexpr/test_registerizer.cpp
+++ b/test/cpp/tensorexpr/test_registerizer.cpp
@@ -16,14 +16,14 @@
   KernelScope kernel_scope;
   BufHandle a("A", {1}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt =
-      Block::make({Store::make(a, {0}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {0}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {0}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               a, {0}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
 
   /*
    * A[0] = 0;
@@ -61,14 +61,14 @@
   KernelScope kernel_scope;
   BufHandle a("A", {10}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt =
-      Block::make({Store::make(a, {0}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {x}, Add::make(Load::make(a, {x}, 1), x), 1)}))});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {0}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               a, {x}, Add::make(Load::make(a, {x}, 1), x), 1)}))});
 
   /*
    * A[0] = 0;
@@ -108,14 +108,14 @@
   KernelScope kernel_scope;
   BufHandle a("A", {1}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt =
-      Block::make({Store::make(a, {0}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {x}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {0}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               a, {x}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
 
   /*
    * A[0] = 0;
@@ -518,14 +518,14 @@
   VarHandle N("N", kInt);
   BufHandle a("A", {N}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt =
-      Block::make({Store::make(a, {i}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {i}, Add::make(Load::make(a, {i}, 1), x), 1)}))});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {i}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               a, {i}, Add::make(Load::make(a, {i}, 1), x), 1)}))});
 
   /*
    * A[i] = 0;
@@ -720,14 +720,14 @@
   BufHandle a("A", {1}, kInt);
   BufHandle b("B", {10}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt =
-      Block::make({Store::make(a, {0}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           b, {x}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {0}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               b, {x}, Add::make(Load::make(a, {0}, 1), x), 1)}))});
 
   /*
    * A[0] = 0;
@@ -941,8 +941,9 @@
       x,
       0,
       10,
-      Block::make({Store::make(b, {0}, Add::make(Load::make(a, {0}, 1), x), 1),
-                   Store::make(a, {0}, Load::make(b, {0}, 1), 1)}))});
+      Block::make(
+          {Store::make(b, {0}, Add::make(Load::make(a, {0}, 1), x), 1),
+           Store::make(a, {0}, Load::make(b, {0}, 1), 1)}))});
 
   /*
    * for (int x = 0; x < 10; x++) {
@@ -988,15 +989,15 @@
   VarHandle x("x", kInt);
   LoopOptions loopOpts;
   loopOpts.set_gpu_block_index(0);
-  Stmt* stmt =
-      Block::make({Store::make(a, {0}, 0, 1),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {0}, Add::make(Load::make(a, {0}, 1), x), 1)}),
-                       loopOpts)});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {0}, 0, 1),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make(
+               {Store::make(a, {0}, Add::make(Load::make(a, {0}, 1), x), 1)}),
+           loopOpts)});
 
   /*
    * A[0] = 0;
@@ -2428,13 +2429,14 @@
   KernelScope kernel_scope;
   BufHandle a("A", {1}, kInt);
   VarHandle x("x", kInt);
-  Stmt* stmt = Block::make({Store::make(a, {1}, Load::make(a, {0}, 1), 1),
-                            Store::make(a, {0}, Load::make(a, {1}, 1), 1),
-                            Store::make(a, {0}, Load::make(a, {1}, 1), 1),
-                            For::make(x, 1, 10, Store::make(a, {x}, x, 1)),
-                            Store::make(a, {1}, Load::make(a, {0}, 1), 1),
-                            Store::make(a, {0}, Load::make(a, {1}, 1), 1),
-                            Store::make(a, {0}, Load::make(a, {1}, 1), 1)});
+  Stmt* stmt = Block::make(
+      {Store::make(a, {1}, Load::make(a, {0}, 1), 1),
+       Store::make(a, {0}, Load::make(a, {1}, 1), 1),
+       Store::make(a, {0}, Load::make(a, {1}, 1), 1),
+       For::make(x, 1, 10, Store::make(a, {x}, x, 1)),
+       Store::make(a, {1}, Load::make(a, {0}, 1), 1),
+       Store::make(a, {0}, Load::make(a, {1}, 1), 1),
+       Store::make(a, {0}, Load::make(a, {1}, 1), 1)});
 
   /*
    * A[1] = A[0];
@@ -2502,9 +2504,10 @@
       {Store::make(a, {0}, Add::make(Load::make(a, {0}, 1), 1), 1),
        Block::make(
            {Store::make(a, {0}, Add::make(Load::make(a, {0}, 1), 2), 1)}),
-       Block::make({Store::make(a, {0}, Add::make(Load::make(a, {0}, 1), 3), 1),
-                    Block::make({Store::make(
-                        a, {0}, Add::make(Load::make(a, {0}, 1), 4), 1)})})});
+       Block::make(
+           {Store::make(a, {0}, Add::make(Load::make(a, {0}, 1), 3), 1),
+            Block::make({Store::make(
+                a, {0}, Add::make(Load::make(a, {0}, 1), 4), 1)})})});
 
   /*
    * A[0] = (A[0]) + 1;
@@ -3441,14 +3444,14 @@
   BufHandle a("A", {10}, kInt);
   VarHandle x("x", kInt);
   VarHandle y("y", kInt);
-  Stmt* stmt =
-      Block::make({Let::make(y, 30),
-                   For::make(
-                       x,
-                       0,
-                       10,
-                       Block::make({Store::make(
-                           a, {y}, Add::make(x, Load::make(a, {y}, 1)), 1)}))});
+  Stmt* stmt = Block::make(
+      {Let::make(y, 30),
+       For::make(
+           x,
+           0,
+           10,
+           Block::make({Store::make(
+               a, {y}, Add::make(x, Load::make(a, {y}, 1)), 1)}))});
 
   /*
    * int y = 30;
diff --git a/test/cpp/tensorexpr/test_simplify.cpp b/test/cpp/tensorexpr/test_simplify.cpp
index 75df7ca..c34812e 100644
--- a/test/cpp/tensorexpr/test_simplify.cpp
+++ b/test/cpp/tensorexpr/test_simplify.cpp
@@ -3939,24 +3939,25 @@
     // Can't fuse, CompareSelect results are different.
     // Actually we totally could if we normalized CompareSelect results, but
     // TODO for later.
-    auto body = Block::make({Cond::make(
-                                 CompareSelect::make(
-                                     i,
-                                     10,
-                                     new IntImm(1),
-                                     new IntImm(0),
-                                     CompareSelectOperation::kLT),
-                                 Store::make(a, {0}, i, mask),
-                                 nullptr),
-                             Cond::make(
-                                 CompareSelect::make(
-                                     j,
-                                     10,
-                                     new IntImm(2),
-                                     new IntImm(0),
-                                     CompareSelectOperation::kLT),
-                                 Store::make(a, {1}, i, mask),
-                                 nullptr)});
+    auto body = Block::make(
+        {Cond::make(
+             CompareSelect::make(
+                 i,
+                 10,
+                 new IntImm(1),
+                 new IntImm(0),
+                 CompareSelectOperation::kLT),
+             Store::make(a, {0}, i, mask),
+             nullptr),
+         Cond::make(
+             CompareSelect::make(
+                 j,
+                 10,
+                 new IntImm(2),
+                 new IntImm(0),
+                 CompareSelectOperation::kLT),
+             Store::make(a, {1}, i, mask),
+             nullptr)});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
@@ -4155,20 +4156,21 @@
 
   {
     // Can fuse if the conditions simplify to the same thing.
-    auto body = Block::make({Cond::make(
-                                 CompareSelect::make(
-                                     i * 2,
-                                     ExprHandle(87) % ExprHandle(11),
-                                     CompareSelectOperation::kLT),
-                                 Store::make(a, {0}, i, mask),
-                                 nullptr),
-                             Cond::make(
-                                 CompareSelect::make(
-                                     i * 2,
-                                     ExprHandle(300) / ExprHandle(30),
-                                     CompareSelectOperation::kLT),
-                                 Store::make(a, {1}, i, mask),
-                                 nullptr)});
+    auto body = Block::make(
+        {Cond::make(
+             CompareSelect::make(
+                 i * 2,
+                 ExprHandle(87) % ExprHandle(11),
+                 CompareSelectOperation::kLT),
+             Store::make(a, {0}, i, mask),
+             nullptr),
+         Cond::make(
+             CompareSelect::make(
+                 i * 2,
+                 ExprHandle(300) / ExprHandle(30),
+                 CompareSelectOperation::kLT),
+             Store::make(a, {1}, i, mask),
+             nullptr)});
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
     ASSERT_EQ(block->nstmts(), 1);
@@ -4181,9 +4183,9 @@
   {
     // Can fuse non-CompareSelects.
     // if (i) { X } if (i) { Y } => if (i) { X; Y }
-    auto body =
-        Block::make({Cond::make(i, Store::make(a, {0}, i, mask), nullptr),
-                     Cond::make(i, Store::make(a, {1}, i, mask), nullptr)});
+    auto body = Block::make(
+        {Cond::make(i, Store::make(a, {0}, i, mask), nullptr),
+         Cond::make(i, Store::make(a, {1}, i, mask), nullptr)});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
@@ -4196,9 +4198,9 @@
 
   {
     // Sanity check wont fuse different non-CompareSelects.
-    auto body =
-        Block::make({Cond::make(i, Store::make(a, {0}, i, mask), nullptr),
-                     Cond::make(j, Store::make(a, {1}, i, mask), nullptr)});
+    auto body = Block::make(
+        {Cond::make(i, Store::make(a, {0}, i, mask), nullptr),
+         Cond::make(j, Store::make(a, {1}, i, mask), nullptr)});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
@@ -4210,9 +4212,9 @@
   {
     // Sanity check constant condition elimination still occurs when merging is
     // possible.
-    auto body =
-        Block::make({Cond::make(1, Store::make(a, {0}, i, mask), nullptr),
-                     Cond::make(1, Store::make(a, {1}, i, mask), nullptr)});
+    auto body = Block::make(
+        {Cond::make(1, Store::make(a, {0}, i, mask), nullptr),
+         Cond::make(1, Store::make(a, {1}, i, mask), nullptr)});
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
     ASSERT_EQ(block->nstmts(), 2);
@@ -4251,10 +4253,11 @@
 
   {
     // Merge two inner SyncThreads.
-    auto body = Block::make({Store::make(a, {0}, 1, 1),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             Store::make(a, {1}, 0, 1)});
+    auto body = Block::make(
+        {Store::make(a, {0}, 1, 1),
+         new SyncThreads(),
+         new SyncThreads(),
+         Store::make(a, {1}, 0, 1)});
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
     ASSERT_EQ(block->nstmts(), 3);
@@ -4278,13 +4281,14 @@
 
   {
     // Merge many inner SyncThreads.
-    auto body = Block::make({Store::make(a, {0}, 1, 1),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             Store::make(a, {1}, 0, 1)});
+    auto body = Block::make(
+        {Store::make(a, {0}, 1, 1),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads(),
+         Store::make(a, {1}, 0, 1)});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
@@ -4297,13 +4301,14 @@
 
   {
     // Merge multiple outer SyncThreads.
-    auto body = Block::make({new SyncThreads(),
-                             new SyncThreads(),
-                             Store::make(a, {1}, 0, 1),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads()});
+    auto body = Block::make(
+        {new SyncThreads(),
+         new SyncThreads(),
+         Store::make(a, {1}, 0, 1),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads()});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
@@ -4314,15 +4319,16 @@
 
   {
     // Merge multiple sections;
-    auto body = Block::make({Store::make(a, {0}, 1, 1),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             Store::make(a, {1}, 0, 1),
-                             Store::make(a, {2}, 0, 1),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             new SyncThreads(),
-                             Store::make(a, {3}, 0, 1)});
+    auto body = Block::make(
+        {Store::make(a, {0}, 1, 1),
+         new SyncThreads(),
+         new SyncThreads(),
+         Store::make(a, {1}, 0, 1),
+         Store::make(a, {2}, 0, 1),
+         new SyncThreads(),
+         new SyncThreads(),
+         new SyncThreads(),
+         Store::make(a, {3}, 0, 1)});
 
     Stmt* simplified = IRSimplifier::simplify(body);
     IS_NODE_WITH_NAME(Block, simplified, block);
diff --git a/test/cpp/tensorexpr/test_train.cpp b/test/cpp/tensorexpr/test_train.cpp
index ceac7e7..5bfe68d 100644
--- a/test/cpp/tensorexpr/test_train.cpp
+++ b/test/cpp/tensorexpr/test_train.cpp
@@ -304,14 +304,15 @@
 
     for (auto i = 0; i < 100; ++i) {
       std::generate(X_.begin(), X_.end(), gen);
-      cg.call({X_.data(),
-               W_ref_.data(),
-               W_.data(),
-               one_.data(),
-               K_.data(),
-               LR_.data(),
-               W_.data(),
-               N});
+      cg.call(
+          {X_.data(),
+           W_ref_.data(),
+           W_.data(),
+           one_.data(),
+           K_.data(),
+           LR_.data(),
+           W_.data(),
+           N});
     }
     // Less than 1% difference after running regression
     for (auto i = 0; i < W_.size(); ++i) {
diff --git a/test/cpp/tensorexpr/test_train_impl.cpp b/test/cpp/tensorexpr/test_train_impl.cpp
index b653f51..2beffa2 100644
--- a/test/cpp/tensorexpr/test_train_impl.cpp
+++ b/test/cpp/tensorexpr/test_train_impl.cpp
@@ -303,8 +303,9 @@
     },
     [](const std::vector<VTensor*>& inputs,
        const std::vector<VTensor*>& ginputs) -> std::vector<VTensor*> {
-      return {call("mul", {ginputs[0], inputs[1]})[0],
-              call("mul", {ginputs[0], inputs[0]})[0]};
+      return {
+          call("mul", {ginputs[0], inputs[1]})[0],
+          call("mul", {ginputs[0], inputs[0]})[0]};
     },
     [](const std::vector<VTensor*>& inputs)
         -> std::vector<std::vector<std::string>> {
@@ -329,8 +330,9 @@
        const std::vector<VTensor*>& ginputs) -> std::vector<VTensor*> {
       auto b_2 = call("mul", {inputs[1], inputs[1]})[0];
       auto a_div_b_2 = call("div", {inputs[0], b_2})[0];
-      return {call("div", {ginputs[0], inputs[1]})[0],
-              call("mul", {ginputs[0], call("neg", {a_div_b_2})[0]})[0]};
+      return {
+          call("div", {ginputs[0], inputs[1]})[0],
+          call("mul", {ginputs[0], call("neg", {a_div_b_2})[0]})[0]};
     },
     [](const std::vector<VTensor*>& inputs)
         -> std::vector<std::vector<std::string>> {
diff --git a/torch/csrc/jit/api/module.cpp b/torch/csrc/jit/api/module.cpp
index d74905b..3c53e54 100644
--- a/torch/csrc/jit/api/module.cpp
+++ b/torch/csrc/jit/api/module.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/api/module.h>
+
 #include <ATen/record_function.h>
 #include <c10/util/Exception.h>
 #include <torch/csrc/autograd/generated/variable_factories.h>
diff --git a/torch/csrc/jit/backends/backend_detail.cpp b/torch/csrc/jit/backends/backend_detail.cpp
index 9ec2129..32a1817 100644
--- a/torch/csrc/jit/backends/backend_detail.cpp
+++ b/torch/csrc/jit/backends/backend_detail.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/backends/backend_detail.h>
+
 #include <ATen/core/builtin_function.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/backends/backend_init.cpp b/torch/csrc/jit/backends/backend_init.cpp
index 596c19e..2f6ceeb 100644
--- a/torch/csrc/jit/backends/backend_init.cpp
+++ b/torch/csrc/jit/backends/backend_init.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/backends/backend_init.h>
+
 #include <torch/csrc/jit/backends/backend_detail.h>
 #include <torch/csrc/jit/backends/backend_resolver.h>
 #include <torch/csrc/jit/frontend/code_template.h>
@@ -130,11 +131,12 @@
   auto codegen_lambda = [=](const std::string& backend_name,
                             const Module& orig_module,
                             const py::dict& method_compile_spec) {
-    const c10::QualifiedName qual_backend_name({"__torch__",
-                                                "torch",
-                                                "classes",
-                                                detail::kBackendsNamespace,
-                                                backend_name});
+    const c10::QualifiedName qual_backend_name(
+        {"__torch__",
+         "torch",
+         "classes",
+         detail::kBackendsNamespace,
+         backend_name});
     // TODO: Validate method_compile_spec.
 
     // Clone orig_module to make sure backend transformation is
diff --git a/torch/csrc/jit/codegen/cuda/arith.cpp b/torch/csrc/jit/codegen/cuda/arith.cpp
index f881dc2..a9abd22 100644
--- a/torch/csrc/jit/codegen/cuda/arith.cpp
+++ b/torch/csrc/jit/codegen/cuda/arith.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/arith.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
 #include <torch/csrc/jit/codegen/cuda/type.h>
diff --git a/torch/csrc/jit/codegen/cuda/codegen.cpp b/torch/csrc/jit/codegen/cuda/codegen.cpp
index 459a6cc..c787404 100644
--- a/torch/csrc/jit/codegen/cuda/codegen.cpp
+++ b/torch/csrc/jit/codegen/cuda/codegen.cpp
@@ -293,9 +293,9 @@
       code_ << *op << gen(node->in());
     } else {
       if (node->getUnaryOpType() == UnaryOpType::Cast) {
-        const auto cast_str =
-            cast_func_str({node->in()->getDataType().value(),
-                           node->out()->getDataType().value()});
+        const auto cast_str = cast_func_str(
+            {node->in()->getDataType().value(),
+             node->out()->getDataType().value()});
         code_ << cast_str.value();
       } else {
         code_ << node->getUnaryOpType();
diff --git a/torch/csrc/jit/codegen/cuda/fusion.cpp b/torch/csrc/jit/codegen/cuda/fusion.cpp
index b6e6bd4..4a6fc58 100644
--- a/torch/csrc/jit/codegen/cuda/fusion.cpp
+++ b/torch/csrc/jit/codegen/cuda/fusion.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
+
 #include <torch/csrc/jit/codegen/cuda/codegen.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
diff --git a/torch/csrc/jit/codegen/cuda/index_compute.cpp b/torch/csrc/jit/codegen/cuda/index_compute.cpp
index 37eb8ca..eebf733 100644
--- a/torch/csrc/jit/codegen/cuda/index_compute.cpp
+++ b/torch/csrc/jit/codegen/cuda/index_compute.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/index_compute.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/codegen/cuda/arith.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
diff --git a/torch/csrc/jit/codegen/cuda/interface.cpp b/torch/csrc/jit/codegen/cuda/interface.cpp
index 09cddc4..b9a92ec 100644
--- a/torch/csrc/jit/codegen/cuda/interface.cpp
+++ b/torch/csrc/jit/codegen/cuda/interface.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/interface.h>
+
 #include <ATen/core/dispatch/OperatorOptions.h>
 #include <torch/csrc/jit/runtime/custom_operator.h>
 #include <torch/csrc/jit/runtime/register_ops_utils.h>
diff --git a/torch/csrc/jit/codegen/cuda/ir_cloner.cpp b/torch/csrc/jit/codegen/cuda/ir_cloner.cpp
index 7fdb9d0..72ae3d5 100644
--- a/torch/csrc/jit/codegen/cuda/ir_cloner.cpp
+++ b/torch/csrc/jit/codegen/cuda/ir_cloner.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/ir_cloner.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
 
diff --git a/torch/csrc/jit/codegen/cuda/ir_graphviz.cpp b/torch/csrc/jit/codegen/cuda/ir_graphviz.cpp
index 2aca8dd..ea3d31b 100644
--- a/torch/csrc/jit/codegen/cuda/ir_graphviz.cpp
+++ b/torch/csrc/jit/codegen/cuda/ir_graphviz.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/ir_graphviz.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
 #include <torch/csrc/jit/codegen/cuda/type.h>
@@ -372,8 +373,8 @@
   const bool is_output = outputs_.find(tv) != outputs_.end();
 
   const char* style = is_input ? "style=filled, fillcolor=palegreen"
-                               : is_output ? "style=filled, fillcolor=lightblue"
-                                           : "style=filled, fillcolor=beige";
+      : is_output              ? "style=filled, fillcolor=lightblue"
+                               : "style=filled, fillcolor=beige";
 
   graph_def_ << "    " << getid(tv) << " [label=\"" << label.str()
              << "\", shape=Mrecord, color=brown, " << style << "];\n";
diff --git a/torch/csrc/jit/codegen/cuda/ir_iostream.cpp b/torch/csrc/jit/codegen/cuda/ir_iostream.cpp
index 145a1b2..8108a3c 100644
--- a/torch/csrc/jit/codegen/cuda/ir_iostream.cpp
+++ b/torch/csrc/jit/codegen/cuda/ir_iostream.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
diff --git a/torch/csrc/jit/codegen/cuda/iter_visitor.cpp b/torch/csrc/jit/codegen/cuda/iter_visitor.cpp
index 4cbdba8..d1efc6d 100644
--- a/torch/csrc/jit/codegen/cuda/iter_visitor.cpp
+++ b/torch/csrc/jit/codegen/cuda/iter_visitor.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/iter_visitor.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
diff --git a/torch/csrc/jit/codegen/cuda/iter_visitor.h b/torch/csrc/jit/codegen/cuda/iter_visitor.h
index 7d8ba55..d70c4f9 100644
--- a/torch/csrc/jit/codegen/cuda/iter_visitor.h
+++ b/torch/csrc/jit/codegen/cuda/iter_visitor.h
@@ -68,8 +68,8 @@
 
   virtual std::vector<Statement*> next(Expr* expr) {
     FusionGuard::getCurFusion()->assertInFusion(expr, "Cannot traverse expr, ");
-    std::vector<Statement*> next_stmts{expr->inputs().begin(),
-                                       expr->inputs().end()};
+    std::vector<Statement*> next_stmts{
+        expr->inputs().begin(), expr->inputs().end()};
     return next_stmts;
   }
 
diff --git a/torch/csrc/jit/codegen/cuda/kernel.cpp b/torch/csrc/jit/codegen/cuda/kernel.cpp
index ef54151..1ad8d06 100644
--- a/torch/csrc/jit/codegen/cuda/kernel.cpp
+++ b/torch/csrc/jit/codegen/cuda/kernel.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/kernel.h>
+
 #include <torch/csrc/jit/codegen/cuda/dispatch.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/kernel_ir_printer.h>
diff --git a/torch/csrc/jit/codegen/cuda/kernel_cache.cpp b/torch/csrc/jit/codegen/cuda/kernel_cache.cpp
index 5bddc51..6b5585d 100644
--- a/torch/csrc/jit/codegen/cuda/kernel_cache.cpp
+++ b/torch/csrc/jit/codegen/cuda/kernel_cache.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/kernel_cache.h>
+
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_utils.h>
 #include <torch/csrc/jit/codegen/cuda/parser.h>
diff --git a/torch/csrc/jit/codegen/cuda/kernel_ir_printer.cpp b/torch/csrc/jit/codegen/cuda/kernel_ir_printer.cpp
index 4d7913e..4a20543 100644
--- a/torch/csrc/jit/codegen/cuda/kernel_ir_printer.cpp
+++ b/torch/csrc/jit/codegen/cuda/kernel_ir_printer.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/kernel_ir_printer.h>
+
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/type.h>
 
@@ -163,8 +164,9 @@
     os_ << *op << gen(node->in());
   } else {
     if (node->getUnaryOpType() == UnaryOpType::Cast) {
-      const auto cast_str = cast_func_str({node->in()->getDataType().value(),
-                                           node->out()->getDataType().value()});
+      const auto cast_str = cast_func_str(
+          {node->in()->getDataType().value(),
+           node->out()->getDataType().value()});
       os_ << cast_str.value();
     } else {
       os_ << node->getUnaryOpType();
diff --git a/torch/csrc/jit/codegen/cuda/lower2device.cpp b/torch/csrc/jit/codegen/cuda/lower2device.cpp
index 7f0dd63..fe36afb 100644
--- a/torch/csrc/jit/codegen/cuda/lower2device.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower2device.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower2device.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
diff --git a/torch/csrc/jit/codegen/cuda/lower_alias_memory.cpp b/torch/csrc/jit/codegen/cuda/lower_alias_memory.cpp
index f8c189c..12c5254 100644
--- a/torch/csrc/jit/codegen/cuda/lower_alias_memory.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower_alias_memory.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower_alias_memory.h>
+
 #include <torch/csrc/jit/codegen/cuda/expr_evaluator.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_utils.h>
diff --git a/torch/csrc/jit/codegen/cuda/lower_insert_syncs.cpp b/torch/csrc/jit/codegen/cuda/lower_insert_syncs.cpp
index 4431025..4326b83 100644
--- a/torch/csrc/jit/codegen/cuda/lower_insert_syncs.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower_insert_syncs.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower_insert_syncs.h>
+
 #include <torch/csrc/jit/codegen/cuda/expr_evaluator.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
diff --git a/torch/csrc/jit/codegen/cuda/lower_loops.cpp b/torch/csrc/jit/codegen/cuda/lower_loops.cpp
index e98761a..a291916 100644
--- a/torch/csrc/jit/codegen/cuda/lower_loops.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower_loops.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower_loops.h>
+
 #include <torch/csrc/jit/codegen/cuda/arith.h>
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
 #include <torch/csrc/jit/codegen/cuda/ir_utils.h>
diff --git a/torch/csrc/jit/codegen/cuda/lower_utils.cpp b/torch/csrc/jit/codegen/cuda/lower_utils.cpp
index 4d59b35..cfb2fce 100644
--- a/torch/csrc/jit/codegen/cuda/lower_utils.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower_utils.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower_utils.h>
+
 #include <torch/csrc/jit/codegen/cuda/arith.h>
 #include <torch/csrc/jit/codegen/cuda/ir_iostream.h>
 #include <torch/csrc/jit/codegen/cuda/ir_utils.h>
@@ -488,12 +489,13 @@
 }
 
 const std::unordered_map<ParallelType, int, TypeHash>
-    ParallelTypeBitmap::pt_to_offset_{{ParallelType::BIDx, 0},
-                                      {ParallelType::BIDy, 1},
-                                      {ParallelType::BIDz, 2},
-                                      {ParallelType::TIDx, 3},
-                                      {ParallelType::TIDy, 4},
-                                      {ParallelType::TIDz, 5}};
+    ParallelTypeBitmap::pt_to_offset_{
+        {ParallelType::BIDx, 0},
+        {ParallelType::BIDy, 1},
+        {ParallelType::BIDz, 2},
+        {ParallelType::TIDx, 3},
+        {ParallelType::TIDy, 4},
+        {ParallelType::TIDz, 5}};
 
 const std::unordered_map<int, ParallelType> ParallelTypeBitmap::offset_to_pt_ =
     {{0, ParallelType::BIDx},
diff --git a/torch/csrc/jit/codegen/cuda/lower_validation.cpp b/torch/csrc/jit/codegen/cuda/lower_validation.cpp
index 1ec32be..1967a66 100644
--- a/torch/csrc/jit/codegen/cuda/lower_validation.cpp
+++ b/torch/csrc/jit/codegen/cuda/lower_validation.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/lower_validation.h>
+
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/iter_visitor.h>
 #include <torch/csrc/jit/codegen/cuda/lower_utils.h>
diff --git a/torch/csrc/jit/codegen/cuda/mutator.cpp b/torch/csrc/jit/codegen/cuda/mutator.cpp
index affaddc..7e7dcf7 100644
--- a/torch/csrc/jit/codegen/cuda/mutator.cpp
+++ b/torch/csrc/jit/codegen/cuda/mutator.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/mutator.h>
+
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/ir_all_nodes.h>
 
diff --git a/torch/csrc/jit/codegen/cuda/partition.cpp b/torch/csrc/jit/codegen/cuda/partition.cpp
index a686567..88aaa3e 100644
--- a/torch/csrc/jit/codegen/cuda/partition.cpp
+++ b/torch/csrc/jit/codegen/cuda/partition.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/partition.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/codegen/cuda/parser.h>
diff --git a/torch/csrc/jit/codegen/cuda/runtime/grid_reduction.cu b/torch/csrc/jit/codegen/cuda/runtime/grid_reduction.cu
index 15e2041..8900ab8 100644
--- a/torch/csrc/jit/codegen/cuda/runtime/grid_reduction.cu
+++ b/torch/csrc/jit/codegen/cuda/runtime/grid_reduction.cu
@@ -63,9 +63,10 @@
 // Returns dim3 of each reduction segment.
 template <bool X_BLOCK, bool Y_BLOCK, bool Z_BLOCK, typename _dim3>
 __device__ dim3 dimension_of_reduction_segment(const _dim3& grid_dim) {
-  return dim3{X_BLOCK ? grid_dim.x : 1,
-              Y_BLOCK ? grid_dim.y : 1,
-              Z_BLOCK ? grid_dim.z : 1};
+  return dim3{
+      X_BLOCK ? grid_dim.x : 1,
+      Y_BLOCK ? grid_dim.y : 1,
+      Z_BLOCK ? grid_dim.z : 1};
 }
 
 // Returns the number of blocks in each reduction segment.
@@ -123,9 +124,10 @@
 // Returns dim3 of each reduction block.
 template <bool X_THREAD, bool Y_THREAD, bool Z_THREAD, typename _dim3>
 __device__ dim3 dimension_of_reduction_block(const _dim3& block_dim) {
-  return dim3{X_THREAD ? block_dim.x : 1,
-              Y_THREAD ? block_dim.y : 1,
-              Z_THREAD ? block_dim.z : 1};
+  return dim3{
+      X_THREAD ? block_dim.x : 1,
+      Y_THREAD ? block_dim.y : 1,
+      Z_THREAD ? block_dim.z : 1};
 }
 
 // Returns the number of threads of each reduction block.
diff --git a/torch/csrc/jit/codegen/cuda/shape_inference.cpp b/torch/csrc/jit/codegen/cuda/shape_inference.cpp
index 7b49cce..5430d0e 100644
--- a/torch/csrc/jit/codegen/cuda/shape_inference.cpp
+++ b/torch/csrc/jit/codegen/cuda/shape_inference.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/shape_inference.h>
+
 #include <c10/core/ScalarType.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
 #include <torch/csrc/jit/ir/constants.h>
diff --git a/torch/csrc/jit/codegen/cuda/transform_iter.cpp b/torch/csrc/jit/codegen/cuda/transform_iter.cpp
index 479177d..c7c356b 100644
--- a/torch/csrc/jit/codegen/cuda/transform_iter.cpp
+++ b/torch/csrc/jit/codegen/cuda/transform_iter.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/transform_iter.h>
+
 #include <torch/csrc/jit/codegen/cuda/ir_utils.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/codegen/cuda/transform_replay.cpp b/torch/csrc/jit/codegen/cuda/transform_replay.cpp
index 4a57a0f..120904a 100644
--- a/torch/csrc/jit/codegen/cuda/transform_replay.cpp
+++ b/torch/csrc/jit/codegen/cuda/transform_replay.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/cuda/transform_replay.h>
+
 #include <torch/csrc/jit/codegen/cuda/arith.h>
 #include <torch/csrc/jit/codegen/cuda/fusion.h>
 #include <torch/csrc/jit/codegen/cuda/instrumentation.h>
diff --git a/torch/csrc/jit/codegen/fuser/cpu/fused_kernel.cpp b/torch/csrc/jit/codegen/fuser/cpu/fused_kernel.cpp
index 4f4aa0d..6d249ee 100644
--- a/torch/csrc/jit/codegen/fuser/cpu/fused_kernel.cpp
+++ b/torch/csrc/jit/codegen/fuser/cpu/fused_kernel.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/fuser/cpu/fused_kernel.h>
+
 #include <c10/util/Exception.h>
 #include <c10/util/Optional.h>
 #include <torch/csrc/jit/codegen/fuser/compiler.h>
diff --git a/torch/csrc/jit/codegen/fuser/cuda/fused_kernel.cpp b/torch/csrc/jit/codegen/fuser/cuda/fused_kernel.cpp
index 9971cb7..03ae998 100644
--- a/torch/csrc/jit/codegen/fuser/cuda/fused_kernel.cpp
+++ b/torch/csrc/jit/codegen/fuser/cuda/fused_kernel.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/codegen/fuser/cuda/fused_kernel.h>
+
 #include <torch/csrc/jit/codegen/fuser/compiler.h>
 
 #include <ATen/ATen.h>
diff --git a/torch/csrc/jit/frontend/builtin_functions.cpp b/torch/csrc/jit/frontend/builtin_functions.cpp
index 22bd5b6..a8d0795 100644
--- a/torch/csrc/jit/frontend/builtin_functions.cpp
+++ b/torch/csrc/jit/frontend/builtin_functions.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/builtin_functions.h>
+
 #include <torch/csrc/api/include/torch/jit.h>
 #include <torch/csrc/jit/frontend/code_template.h>
 #include <torch/csrc/jit/frontend/resolver.h>
diff --git a/torch/csrc/jit/frontend/concrete_module_type.cpp b/torch/csrc/jit/frontend/concrete_module_type.cpp
index be3c173..b9ef3ec 100644
--- a/torch/csrc/jit/frontend/concrete_module_type.cpp
+++ b/torch/csrc/jit/frontend/concrete_module_type.cpp
@@ -256,8 +256,8 @@
   TORCH_INTERNAL_ASSERT(type);
   functionAttributes_.emplace(
       std::move(name),
-      ConcreteModuleTypeBuilder::FunctionAttribute{type->expect<FunctionType>(),
-                                                   std::move(pyFunction)});
+      ConcreteModuleTypeBuilder::FunctionAttribute{
+          type->expect<FunctionType>(), std::move(pyFunction)});
 }
 
 void ConcreteModuleTypeBuilder::addBuiltinFunction(
diff --git a/torch/csrc/jit/frontend/error_report.cpp b/torch/csrc/jit/frontend/error_report.cpp
index d148214..0c20014 100644
--- a/torch/csrc/jit/frontend/error_report.cpp
+++ b/torch/csrc/jit/frontend/error_report.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/error_report.h>
+
 #include <c10/util/Optional.h>
 #include <torch/csrc/jit/frontend/tree.h>
 #include <torch/csrc/utils/memory.h>
diff --git a/torch/csrc/jit/frontend/exit_transforms.cpp b/torch/csrc/jit/frontend/exit_transforms.cpp
index 60e3bb4..bb94f4d 100644
--- a/torch/csrc/jit/frontend/exit_transforms.cpp
+++ b/torch/csrc/jit/frontend/exit_transforms.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/exit_transforms.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/jit/frontend/error_report.h>
 #include <torch/csrc/jit/ir/ir.h>
diff --git a/torch/csrc/jit/frontend/function_schema_parser.cpp b/torch/csrc/jit/frontend/function_schema_parser.cpp
index e7d9da1..2daeabc 100644
--- a/torch/csrc/jit/frontend/function_schema_parser.cpp
+++ b/torch/csrc/jit/frontend/function_schema_parser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/function_schema_parser.h>
+
 #include <ATen/core/Reduction.h>
 #include <c10/util/string_utils.h>
 #include <torch/csrc/jit/frontend/lexer.h>
diff --git a/torch/csrc/jit/frontend/parser.cpp b/torch/csrc/jit/frontend/parser.cpp
index 2e1b2c9..87ae08a 100644
--- a/torch/csrc/jit/frontend/parser.cpp
+++ b/torch/csrc/jit/frontend/parser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/parser.h>
+
 #include <c10/util/Optional.h>
 #include <torch/csrc/jit/frontend/lexer.h>
 #include <torch/csrc/jit/frontend/parse_string_literal.h>
@@ -267,8 +268,9 @@
       auto kind = L.cur().kind;
       auto pos = L.cur().range;
       L.next();
-      auto unary_kind =
-          kind == '*' ? TK_STARRED : kind == '-' ? TK_UNARY_MINUS : kind;
+      auto unary_kind = kind == '*' ? TK_STARRED
+          : kind == '-'             ? TK_UNARY_MINUS
+                                    : kind;
       auto subexp = parseExp(unary_prec);
       // fold '-' into constant numbers, so that attributes can accept
       // things like -1
diff --git a/torch/csrc/jit/frontend/schema_matching.cpp b/torch/csrc/jit/frontend/schema_matching.cpp
index f431dd3..86c1c2f 100644
--- a/torch/csrc/jit/frontend/schema_matching.cpp
+++ b/torch/csrc/jit/frontend/schema_matching.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/schema_matching.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/jit/frontend/builtin_functions.h>
 #include <torch/csrc/jit/frontend/error_report.h>
@@ -454,9 +455,10 @@
     return_field_names =
         fmap(returns, [&](const Argument& r) { return r.name(); });
   }
-  return MatchedSchema{std::move(positional_inputs),
-                       std::move(return_types),
-                       std::move(return_field_names)};
+  return MatchedSchema{
+      std::move(positional_inputs),
+      std::move(return_types),
+      std::move(return_field_names)};
 }
 
 MatchedSchema matchSchema(
diff --git a/torch/csrc/jit/frontend/schema_type_parser.cpp b/torch/csrc/jit/frontend/schema_type_parser.cpp
index 05b7bf2..1785a24 100644
--- a/torch/csrc/jit/frontend/schema_type_parser.cpp
+++ b/torch/csrc/jit/frontend/schema_type_parser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/schema_type_parser.h>
+
 #include <ATen/core/alias_info.h>
 #include <ATen/core/interned_strings.h>
 #include <ATen/core/jit_type.h>
diff --git a/torch/csrc/jit/frontend/script_type_parser.cpp b/torch/csrc/jit/frontend/script_type_parser.cpp
index f4c1fa2..64837e6 100644
--- a/torch/csrc/jit/frontend/script_type_parser.cpp
+++ b/torch/csrc/jit/frontend/script_type_parser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/script_type_parser.h>
+
 #include <torch/csrc/jit/frontend/parser.h>
 #include <torch/csrc/jit/ir/ir.h>
 #include <torch/custom_class.h>
diff --git a/torch/csrc/jit/frontend/sugared_value.h b/torch/csrc/jit/frontend/sugared_value.h
index 6f8b760..85bf1d2 100644
--- a/torch/csrc/jit/frontend/sugared_value.h
+++ b/torch/csrc/jit/frontend/sugared_value.h
@@ -539,8 +539,8 @@
       size_t n_binders) override {
     TORCH_INTERNAL_ASSERT(args.size() == 0 && kwargs.size() == 0);
     Value* dtype_const = m.graph()->insertConstant(dtype_, loc);
-    std::vector<NamedValue> kwargs_{self_,
-                                    NamedValue(loc, "dtype", dtype_const)};
+    std::vector<NamedValue> kwargs_{
+        self_, NamedValue(loc, "dtype", dtype_const)};
     Value* casted_val = m.graph()->insert(
         /*opname=*/Symbol::fromQualString("aten::to"),
         /*args=*/args,
diff --git a/torch/csrc/jit/frontend/versioned_symbols.cpp b/torch/csrc/jit/frontend/versioned_symbols.cpp
index 8e39e6f..d7ee0d3 100644
--- a/torch/csrc/jit/frontend/versioned_symbols.cpp
+++ b/torch/csrc/jit/frontend/versioned_symbols.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/frontend/versioned_symbols.h>
+
 #include <torch/csrc/api/include/torch/jit.h>
 
 #include <unordered_map>
diff --git a/torch/csrc/jit/ir/attributes.h b/torch/csrc/jit/ir/attributes.h
index ef8daee..16789a1 100644
--- a/torch/csrc/jit/ir/attributes.h
+++ b/torch/csrc/jit/ir/attributes.h
@@ -21,19 +21,20 @@
 
 enum class AttributeKind { f, fs, i, is, s, ss, t, ts, g, gs, ty, tys, ival };
 static inline const char* toString(AttributeKind kind) {
-  static const char* names[] = {"f",
-                                "fs",
-                                "i",
-                                "is",
-                                "s",
-                                "ss",
-                                "t",
-                                "ts",
-                                "g",
-                                "gs",
-                                "ty",
-                                "tys",
-                                "ival"};
+  static const char* names[] = {
+      "f",
+      "fs",
+      "i",
+      "is",
+      "s",
+      "ss",
+      "t",
+      "ts",
+      "g",
+      "gs",
+      "ty",
+      "tys",
+      "ival"};
   AT_ASSERT(size_t(kind) < sizeof(names) / sizeof(*names));
   return names[int(kind)];
 }
diff --git a/torch/csrc/jit/ir/constants.cpp b/torch/csrc/jit/ir/constants.cpp
index df948a6..512533e 100644
--- a/torch/csrc/jit/ir/constants.cpp
+++ b/torch/csrc/jit/ir/constants.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/ir/constants.h>
+
 #include <ATen/core/functional.h>
 #include <torch/csrc/autograd/variable.h>
 #include <torch/csrc/jit/ir/ir.h>
diff --git a/torch/csrc/jit/ir/irparser.cpp b/torch/csrc/jit/ir/irparser.cpp
index 8c278cb..ce4307c 100644
--- a/torch/csrc/jit/ir/irparser.cpp
+++ b/torch/csrc/jit/ir/irparser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/ir/irparser.h>
+
 #include <torch/csrc/jit/frontend/lexer.h>
 #include <torch/csrc/jit/frontend/parse_string_literal.h>
 #include <torch/csrc/jit/frontend/schema_type_parser.h>
diff --git a/torch/csrc/jit/ir/scope.cpp b/torch/csrc/jit/ir/scope.cpp
index eba291b..ba1bced 100644
--- a/torch/csrc/jit/ir/scope.cpp
+++ b/torch/csrc/jit/ir/scope.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/ir/scope.h>
+
 #include <ATen/core/function.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/ir/type_hashing.cpp b/torch/csrc/jit/ir/type_hashing.cpp
index a03f650..c447ceb 100644
--- a/torch/csrc/jit/ir/type_hashing.cpp
+++ b/torch/csrc/jit/ir/type_hashing.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/ir/type_hashing.h>
+
 #include <ATen/core/functional.h>
 #include <ATen/core/jit_type.h>
 #include <ATen/core/qualified_name.h>
diff --git a/torch/csrc/jit/mobile/function.cpp b/torch/csrc/jit/mobile/function.cpp
index 8748bb6..47d492c 100644
--- a/torch/csrc/jit/mobile/function.cpp
+++ b/torch/csrc/jit/mobile/function.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/mobile/function.h>
+
 #include <caffe2/serialize/inline_container.h>
 #include <torch/csrc/jit/mobile/interpreter.h>
 #include <torch/csrc/jit/runtime/instruction.h>
diff --git a/torch/csrc/jit/mobile/import.cpp b/torch/csrc/jit/mobile/import.cpp
index a32eff6..3e7e005 100644
--- a/torch/csrc/jit/mobile/import.cpp
+++ b/torch/csrc/jit/mobile/import.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/mobile/import.h>
+
 #include <ATen/core/ivalue.h>
 #include <caffe2/serialize/inline_container.h>
 #include <torch/csrc/jit/api/compilation_unit.h>
diff --git a/torch/csrc/jit/mobile/interpreter.cpp b/torch/csrc/jit/mobile/interpreter.cpp
index 681eddf..008c42b 100644
--- a/torch/csrc/jit/mobile/interpreter.cpp
+++ b/torch/csrc/jit/mobile/interpreter.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/mobile/interpreter.h>
+
 #include <ATen/core/function.h>
 #include <ATen/core/jit_type.h>
 #include <ATen/core/operator_name.h>
diff --git a/torch/csrc/jit/mobile/module.cpp b/torch/csrc/jit/mobile/module.cpp
index 6e78d4f..1155a5f 100644
--- a/torch/csrc/jit/mobile/module.cpp
+++ b/torch/csrc/jit/mobile/module.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/mobile/module.h>
+
 #include <torch/csrc/jit/mobile/interpreter.h>
 #include <torch/csrc/jit/mobile/observer.h>
 #include <torch/csrc/jit/runtime/jit_exception.h>
diff --git a/torch/csrc/jit/mobile/type_parser.cpp b/torch/csrc/jit/mobile/type_parser.cpp
index 546aabd..42814e5 100644
--- a/torch/csrc/jit/mobile/type_parser.cpp
+++ b/torch/csrc/jit/mobile/type_parser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/mobile/type_parser.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/jit/frontend/parser_constants.h>
 #include <torch/custom_class.h>
diff --git a/torch/csrc/jit/passes/bailout_graph.cpp b/torch/csrc/jit/passes/bailout_graph.cpp
index cfe1850..7e721e1 100644
--- a/torch/csrc/jit/passes/bailout_graph.cpp
+++ b/torch/csrc/jit/passes/bailout_graph.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/bailout_graph.h>
+
 #include <ATen/core/function.h>
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/ir/ir_views.h>
diff --git a/torch/csrc/jit/passes/canonicalize.cpp b/torch/csrc/jit/passes/canonicalize.cpp
index 5180ce6..4a41e8f 100644
--- a/torch/csrc/jit/passes/canonicalize.cpp
+++ b/torch/csrc/jit/passes/canonicalize.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/canonicalize.h>
+
 #include <torch/csrc/jit/ir/ir_views.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/clear_profiling.cpp b/torch/csrc/jit/passes/clear_profiling.cpp
index 809d177..9acb9fb 100644
--- a/torch/csrc/jit/passes/clear_profiling.cpp
+++ b/torch/csrc/jit/passes/clear_profiling.cpp
@@ -1,5 +1,6 @@
 
 #include <torch/csrc/jit/passes/clear_profiling.h>
+
 #include <torch/csrc/jit/jit_log.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/clear_undefinedness.cpp b/torch/csrc/jit/passes/clear_undefinedness.cpp
index 591c824..b235cbb 100644
--- a/torch/csrc/jit/passes/clear_undefinedness.cpp
+++ b/torch/csrc/jit/passes/clear_undefinedness.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/clear_undefinedness.h>
+
 #include <torch/csrc/jit/jit_log.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/constant_pooling.cpp b/torch/csrc/jit/passes/constant_pooling.cpp
index 2ee9dd1..06a5d61 100644
--- a/torch/csrc/jit/passes/constant_pooling.cpp
+++ b/torch/csrc/jit/passes/constant_pooling.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/constant_pooling.h>
+
 #include <ATen/core/interned_strings.h>
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/ir/ir.h>
diff --git a/torch/csrc/jit/passes/constant_propagation.cpp b/torch/csrc/jit/passes/constant_propagation.cpp
index 75be7e86..3db17e5 100644
--- a/torch/csrc/jit/passes/constant_propagation.cpp
+++ b/torch/csrc/jit/passes/constant_propagation.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/constant_propagation.h>
+
 #include <ATen/core/functional.h>
 #include <ATen/core/ivalue.h>
 #include <c10/util/Exception.h>
diff --git a/torch/csrc/jit/passes/create_functional_graphs.cpp b/torch/csrc/jit/passes/create_functional_graphs.cpp
index a5c2c6f..d5d85f6 100644
--- a/torch/csrc/jit/passes/create_functional_graphs.cpp
+++ b/torch/csrc/jit/passes/create_functional_graphs.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/create_functional_graphs.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/passes/constant_pooling.h>
diff --git a/torch/csrc/jit/passes/decompose_ops.cpp b/torch/csrc/jit/passes/decompose_ops.cpp
index 2d0f70a..ad8dcf3 100644
--- a/torch/csrc/jit/passes/decompose_ops.cpp
+++ b/torch/csrc/jit/passes/decompose_ops.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/decompose_ops.h>
+
 #include <torch/csrc/jit/frontend/ir_emitter.h>
 #include <torch/csrc/jit/passes/constant_propagation.h>
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
@@ -125,12 +126,13 @@
       Graph* graph = it->owningGraph();
       Value* input = it->namedInput(attr::input);
       Value* input_dim = graph->insert(aten::dim, {input});
-      std::vector<Value*> inputs{input,
-                                 it->namedInput(attr::running_mean),
-                                 it->namedInput(attr::running_var),
-                                 it->namedInput(attr::training),
-                                 it->namedInput(attr::momentum),
-                                 it->namedInput(attr::eps)};
+      std::vector<Value*> inputs{
+          input,
+          it->namedInput(attr::running_mean),
+          it->namedInput(attr::running_var),
+          it->namedInput(attr::training),
+          it->namedInput(attr::momentum),
+          it->namedInput(attr::eps)};
 
       // inline the compiled decomposed batchnorm
       std::shared_ptr<Graph> d_graph =
@@ -161,10 +163,11 @@
       decomposed = true;
       WithInsertPoint insert_guard{*it};
       Graph* graph = it->owningGraph();
-      std::vector<Value*> inputs{it->namedInput(attr::input),
-                                 it->namedInput(attr::normalized_shape),
-                                 it->namedInput(attr::eps),
-                                 it->namedInput(attr::cudnn_enable)};
+      std::vector<Value*> inputs{
+          it->namedInput(attr::input),
+          it->namedInput(attr::normalized_shape),
+          it->namedInput(attr::eps),
+          it->namedInput(attr::cudnn_enable)};
 
       // inline the compiled decomposed layernorm
       std::shared_ptr<Graph> d_graph =
diff --git a/torch/csrc/jit/passes/erase_number_types.cpp b/torch/csrc/jit/passes/erase_number_types.cpp
index 7e6a453..738bb48 100644
--- a/torch/csrc/jit/passes/erase_number_types.cpp
+++ b/torch/csrc/jit/passes/erase_number_types.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/erase_number_types.h>
+
 #include <torch/csrc/jit/ir/constants.h>
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
 
diff --git a/torch/csrc/jit/passes/fold_conv_bn.cpp b/torch/csrc/jit/passes/fold_conv_bn.cpp
index 7d34463..8ebde97 100644
--- a/torch/csrc/jit/passes/fold_conv_bn.cpp
+++ b/torch/csrc/jit/passes/fold_conv_bn.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/fold_conv_bn.h>
+
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/graph_rewrite_helper.h>
diff --git a/torch/csrc/jit/passes/freeze_module.cpp b/torch/csrc/jit/passes/freeze_module.cpp
index f66f54e..38684d2 100644
--- a/torch/csrc/jit/passes/freeze_module.cpp
+++ b/torch/csrc/jit/passes/freeze_module.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/freeze_module.h>
+
 #include <torch/csrc/jit/jit_log.h>
 
 #include <torch/csrc/jit/ir/alias_analysis.h>
diff --git a/torch/csrc/jit/passes/fuse_relu.cpp b/torch/csrc/jit/passes/fuse_relu.cpp
index 02e8f9f..479b4b2 100644
--- a/torch/csrc/jit/passes/fuse_relu.cpp
+++ b/torch/csrc/jit/passes/fuse_relu.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/fuse_relu.h>
+
 #include <torch/csrc/jit/ir/ir.h>
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 #include <torch/csrc/jit/passes/subgraph_rewrite.h>
diff --git a/torch/csrc/jit/passes/graph_rewrite_helper.cpp b/torch/csrc/jit/passes/graph_rewrite_helper.cpp
index 8b612a9..8072174 100644
--- a/torch/csrc/jit/passes/graph_rewrite_helper.cpp
+++ b/torch/csrc/jit/passes/graph_rewrite_helper.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/graph_rewrite_helper.h>
+
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 #include <torch/csrc/jit/passes/constant_propagation.h>
 #include <torch/csrc/jit/passes/subgraph_rewrite.h>
diff --git a/torch/csrc/jit/passes/graph_rewrite_helper.h b/torch/csrc/jit/passes/graph_rewrite_helper.h
index f21859d..57b91be 100644
--- a/torch/csrc/jit/passes/graph_rewrite_helper.h
+++ b/torch/csrc/jit/passes/graph_rewrite_helper.h
@@ -39,10 +39,11 @@
   static PatternInfo parse_from_str(
       std::string pattern_string,
       const std::vector<MatchFilter>& filters = {}) {
-    PatternInfo rv{std::move(pattern_string),
-                   std::make_unique<Graph>(),
-                   decltype(vmap){},
-                   filters};
+    PatternInfo rv{
+        std::move(pattern_string),
+        std::make_unique<Graph>(),
+        decltype(vmap){},
+        filters};
     parseIR(rv.pattern_string, rv.pattern_graph.get(), rv.vmap);
     return rv;
   }
diff --git a/torch/csrc/jit/passes/guard_elimination.cpp b/torch/csrc/jit/passes/guard_elimination.cpp
index 9ae808e4..c26404d 100644
--- a/torch/csrc/jit/passes/guard_elimination.cpp
+++ b/torch/csrc/jit/passes/guard_elimination.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/guard_elimination.h>
+
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/constant_propagation.h>
diff --git a/torch/csrc/jit/passes/inline_forked_closures.cpp b/torch/csrc/jit/passes/inline_forked_closures.cpp
index e97d71e..7710500 100644
--- a/torch/csrc/jit/passes/inline_forked_closures.cpp
+++ b/torch/csrc/jit/passes/inline_forked_closures.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/inline_forked_closures.h>
+
 #include <torch/csrc/jit/frontend/ir_emitter.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/inliner.cpp b/torch/csrc/jit/passes/inliner.cpp
index 02edaa7..2d510bd 100644
--- a/torch/csrc/jit/passes/inliner.cpp
+++ b/torch/csrc/jit/passes/inliner.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/inliner.h>
+
 #include <torch/csrc/jit/api/module.h>
 #include <torch/csrc/jit/frontend/error_report.h>
 #include <torch/csrc/jit/jit_log.h>
diff --git a/torch/csrc/jit/passes/lift_closures.cpp b/torch/csrc/jit/passes/lift_closures.cpp
index 4f5941c..09ac524 100644
--- a/torch/csrc/jit/passes/lift_closures.cpp
+++ b/torch/csrc/jit/passes/lift_closures.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/lift_closures.h>
+
 #include <torch/csrc/jit/frontend/ir_emitter.h>
 #include <torch/csrc/jit/ir/ir.h>
 
diff --git a/torch/csrc/jit/passes/liveness.cpp b/torch/csrc/jit/passes/liveness.cpp
index e580eb4..3b2cf54 100644
--- a/torch/csrc/jit/passes/liveness.cpp
+++ b/torch/csrc/jit/passes/liveness.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/liveness.h>
+
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/ir/ir_views.h>
 #include <torch/csrc/jit/passes/constant_pooling.h>
diff --git a/torch/csrc/jit/passes/lower_grad_of.cpp b/torch/csrc/jit/passes/lower_grad_of.cpp
index d9bdd91..3f3de5f 100644
--- a/torch/csrc/jit/passes/lower_grad_of.cpp
+++ b/torch/csrc/jit/passes/lower_grad_of.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/lower_grad_of.h>
+
 #include <torch/csrc/jit/jit_log.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/lower_graph.cpp b/torch/csrc/jit/passes/lower_graph.cpp
index 581f4d1..b4da821 100644
--- a/torch/csrc/jit/passes/lower_graph.cpp
+++ b/torch/csrc/jit/passes/lower_graph.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/lower_graph.h>
+
 #include <torch/csrc/jit/api/object.h>
 #include <torch/csrc/jit/frontend/error_report.h>
 #include <torch/csrc/jit/passes/inliner.h>
diff --git a/torch/csrc/jit/passes/lower_tuples.cpp b/torch/csrc/jit/passes/lower_tuples.cpp
index d241e0d..2fc61d3 100644
--- a/torch/csrc/jit/passes/lower_tuples.cpp
+++ b/torch/csrc/jit/passes/lower_tuples.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/lower_tuples.h>
+
 #include <ATen/core/functional.h>
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/ir/constants.h>
diff --git a/torch/csrc/jit/passes/normalize_ops.cpp b/torch/csrc/jit/passes/normalize_ops.cpp
index e032eeb..a2ea0a7 100644
--- a/torch/csrc/jit/passes/normalize_ops.cpp
+++ b/torch/csrc/jit/passes/normalize_ops.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/normalize_ops.h>
+
 #include <c10/util/Exception.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/onnx.cpp b/torch/csrc/jit/passes/onnx.cpp
index 5bc0a2e..51476dc 100644
--- a/torch/csrc/jit/passes/onnx.cpp
+++ b/torch/csrc/jit/passes/onnx.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx.h>
+
 #include <ATen/core/functional.h>
 #include <c10/util/Exception.h>
 #include <torch/csrc/autograd/function.h>
diff --git a/torch/csrc/jit/passes/onnx/constant_fold.cpp b/torch/csrc/jit/passes/onnx/constant_fold.cpp
index ffb98d9..66c8398 100644
--- a/torch/csrc/jit/passes/onnx/constant_fold.cpp
+++ b/torch/csrc/jit/passes/onnx/constant_fold.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/constant_fold.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/passes/onnx/helper.h>
 
diff --git a/torch/csrc/jit/passes/onnx/fixup_onnx_controlflow.cpp b/torch/csrc/jit/passes/onnx/fixup_onnx_controlflow.cpp
index 19523f5..b826db9 100644
--- a/torch/csrc/jit/passes/onnx/fixup_onnx_controlflow.cpp
+++ b/torch/csrc/jit/passes/onnx/fixup_onnx_controlflow.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/fixup_onnx_controlflow.h>
+
 #include <aten/src/ATen/InitialTensorOptions.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
diff --git a/torch/csrc/jit/passes/onnx/function_substitution.cpp b/torch/csrc/jit/passes/onnx/function_substitution.cpp
index 6c51797..460deae 100644
--- a/torch/csrc/jit/passes/onnx/function_substitution.cpp
+++ b/torch/csrc/jit/passes/onnx/function_substitution.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/function_substitution.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/onnx/helper.h>
 
diff --git a/torch/csrc/jit/passes/onnx/helper.cpp b/torch/csrc/jit/passes/onnx/helper.cpp
index 07f71fc..a14dcd6 100644
--- a/torch/csrc/jit/passes/onnx/helper.cpp
+++ b/torch/csrc/jit/passes/onnx/helper.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/helper.h>
+
 #include <onnx/onnx_pb.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/onnx/list_model_parameters.cpp b/torch/csrc/jit/passes/onnx/list_model_parameters.cpp
index 4f0b4e2..78cc7ae 100644
--- a/torch/csrc/jit/passes/onnx/list_model_parameters.cpp
+++ b/torch/csrc/jit/passes/onnx/list_model_parameters.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/list_model_parameters.h>
+
 #include <torch/csrc/jit/frontend/error_report.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
diff --git a/torch/csrc/jit/passes/onnx/peephole.cpp b/torch/csrc/jit/passes/onnx/peephole.cpp
index e04a74a..3a0bc9a 100644
--- a/torch/csrc/jit/passes/onnx/peephole.cpp
+++ b/torch/csrc/jit/passes/onnx/peephole.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/peephole.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/passes/onnx/helper.h>
 
diff --git a/torch/csrc/jit/passes/onnx/prepare_division_for_onnx.cpp b/torch/csrc/jit/passes/onnx/prepare_division_for_onnx.cpp
index edbb159..709b174 100644
--- a/torch/csrc/jit/passes/onnx/prepare_division_for_onnx.cpp
+++ b/torch/csrc/jit/passes/onnx/prepare_division_for_onnx.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/prepare_division_for_onnx.h>
+
 #include <torch/csrc/jit/ir/constants.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/onnx/preprocess_for_onnx.cpp b/torch/csrc/jit/passes/onnx/preprocess_for_onnx.cpp
index 8eb0746..ec98c01 100644
--- a/torch/csrc/jit/passes/onnx/preprocess_for_onnx.cpp
+++ b/torch/csrc/jit/passes/onnx/preprocess_for_onnx.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/preprocess_for_onnx.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/onnx/helper.h>
 
diff --git a/torch/csrc/jit/passes/onnx/scalar_type_analysis.cpp b/torch/csrc/jit/passes/onnx/scalar_type_analysis.cpp
index 1a63edf..e637169 100644
--- a/torch/csrc/jit/passes/onnx/scalar_type_analysis.cpp
+++ b/torch/csrc/jit/passes/onnx/scalar_type_analysis.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/scalar_type_analysis.h>
+
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/passes/onnx/shape_type_inference.cpp b/torch/csrc/jit/passes/onnx/shape_type_inference.cpp
index e84901a..4c2bf32 100644
--- a/torch/csrc/jit/passes/onnx/shape_type_inference.cpp
+++ b/torch/csrc/jit/passes/onnx/shape_type_inference.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/shape_type_inference.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/onnx/helper.h>
 #include <torch/csrc/jit/passes/onnx/scalar_type_analysis.h>
diff --git a/torch/csrc/jit/passes/onnx/unpack_quantized_weights.cpp b/torch/csrc/jit/passes/onnx/unpack_quantized_weights.cpp
index 0de7f36..b9b9fe1 100644
--- a/torch/csrc/jit/passes/onnx/unpack_quantized_weights.cpp
+++ b/torch/csrc/jit/passes/onnx/unpack_quantized_weights.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/onnx/unpack_quantized_weights.h>
+
 #include <ATen/native/quantized/cpu/packed_params.h>
 #include <torch/csrc/jit/ir/constants.h>
 #include <torch/csrc/jit/ir/irparser.h>
@@ -25,20 +26,21 @@
 double getScaleFromInput(Node* input_node) {
   c10::optional<IValue> scale;
   std::string input_name = input_node->kind().toQualString();
-  std::unordered_set<std::string> noscale_ops = {"quantized::max_pool2d",
-                                                 "aten::max_pool2d",
-                                                 "aten::relu",
-                                                 "prim::ListUnpack",
-                                                 "aten::split_with_sizes",
-                                                 "quantized::nchw2nhwc",
-                                                 "quantized::nhwc2nchw",
-                                                 "aten::slice",
-                                                 "aten::avg_pool2d",
-                                                 "quantized::cat",
-                                                 "prim::ListConstruct",
-                                                 "aten::upsample_nearest2d",
-                                                 "aten::sigmoid",
-                                                 "aten::reshape"};
+  std::unordered_set<std::string> noscale_ops = {
+      "quantized::max_pool2d",
+      "aten::max_pool2d",
+      "aten::relu",
+      "prim::ListUnpack",
+      "aten::split_with_sizes",
+      "quantized::nchw2nhwc",
+      "quantized::nhwc2nchw",
+      "aten::slice",
+      "aten::avg_pool2d",
+      "quantized::cat",
+      "prim::ListConstruct",
+      "aten::upsample_nearest2d",
+      "aten::sigmoid",
+      "aten::reshape"};
   if (input_name == "aten::quantize_per_tensor") {
     TORCH_CHECK(
         input_node->inputs().size() > 1,
@@ -273,10 +275,11 @@
     std::vector<int64_t> wt_sizes = unpacked_weight.sizes().vec();
     if (unpacked_weight.ndimension() == 4) {
       unpacked_weight.permute({0, 2, 3, 1});
-      wt_sizes = {unpacked_weight.size(0),
-                  unpacked_weight.size(2),
-                  unpacked_weight.size(3),
-                  unpacked_weight.size(1)};
+      wt_sizes = {
+          unpacked_weight.size(0),
+          unpacked_weight.size(2),
+          unpacked_weight.size(3),
+          unpacked_weight.size(1)};
     }
 
     // Remove packed_params
diff --git a/torch/csrc/jit/passes/peephole.cpp b/torch/csrc/jit/passes/peephole.cpp
index 167173d..7c659f5 100644
--- a/torch/csrc/jit/passes/peephole.cpp
+++ b/torch/csrc/jit/passes/peephole.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/peephole.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/jit/ir/alias_analysis.h>
 #include <torch/csrc/jit/ir/ir_views.h>
diff --git a/torch/csrc/jit/passes/quantization/dedup_module_uses.cpp b/torch/csrc/jit/passes/quantization/dedup_module_uses.cpp
index c3b78a0..2f20dc8 100644
--- a/torch/csrc/jit/passes/quantization/dedup_module_uses.cpp
+++ b/torch/csrc/jit/passes/quantization/dedup_module_uses.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/quantization/dedup_module_uses.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/quantization/helper.h>
 
diff --git a/torch/csrc/jit/passes/quantization/finalize.cpp b/torch/csrc/jit/passes/quantization/finalize.cpp
index 0f12e94..af4a7bb 100644
--- a/torch/csrc/jit/passes/quantization/finalize.cpp
+++ b/torch/csrc/jit/passes/quantization/finalize.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/quantization/finalize.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/clear_profiling.h>
 #include <torch/csrc/jit/passes/freeze_module.h>
diff --git a/torch/csrc/jit/passes/quantization/helper.cpp b/torch/csrc/jit/passes/quantization/helper.cpp
index 10e46ae..1549b4f 100644
--- a/torch/csrc/jit/passes/quantization/helper.cpp
+++ b/torch/csrc/jit/passes/quantization/helper.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/quantization/helper.h>
+
 #include <torch/csrc/jit/passes/graph_rewrite_helper.h>
 
 namespace torch {
@@ -176,17 +177,19 @@
 std::tuple<c10::QScheme, QParamVector> _per_tensor_asym_qparam =
     std::make_tuple(
         c10::kPerTensorAffine,
-        QParamVector({std::make_pair(".scale", IValue(_asym_scale)),
-                      std::make_pair(".zero_point", IValue(_asym_zero_point)),
-                      std::make_pair(".scalar_type", IValue(c10::kQUInt8))}));
+        QParamVector(
+            {std::make_pair(".scale", IValue(_asym_scale)),
+             std::make_pair(".zero_point", IValue(_asym_zero_point)),
+             std::make_pair(".scalar_type", IValue(c10::kQUInt8))}));
 
 // quantization parrameters for ops with range -1 to 1
 // for example: aten/src/ATen/native/quantized/cpu/qtanh.cpp
 std::tuple<c10::QScheme, QParamVector> _per_tensor_sym_qparam = std::make_tuple(
     c10::kPerTensorAffine,
-    QParamVector({std::make_pair(".scale", IValue(_sym_scale)),
-                  std::make_pair(".zero_point", IValue(_sym_zero_point)),
-                  std::make_pair(".scalar_type", IValue(c10::kQUInt8))}));
+    QParamVector(
+        {std::make_pair(".scale", IValue(_sym_scale)),
+         std::make_pair(".zero_point", IValue(_sym_zero_point)),
+         std::make_pair(".scalar_type", IValue(c10::kQUInt8))}));
 
 // Map from aten op symbol to the quantization parameters
 // for the ops with fixed quantization parameters
@@ -218,10 +221,11 @@
 // the inputs are quantized
 // if the second input is a Scalar, we'll only look at the first input to decide
 // if we need to quantize the output
-std::vector<std::string> _propagate_quant_binary_ops = {"add",
-                                                        "add_",
-                                                        "mul",
-                                                        "mul_"};
+std::vector<std::string> _propagate_quant_binary_ops = {
+    "add",
+    "add_",
+    "mul",
+    "mul_"};
 
 // Check if `use` is an aten function of name `func_name` and if value
 // `v` is the nth argument (if provided) of the function.
@@ -272,13 +276,14 @@
       v,
       // ate::embedding_bag(%weight, %input, %offsets, %scale_grad_by_freq,
       // %mode_enum, %sparse, %per_sample_weights, %include_last_offset)
-      AtenFuncArgs({{"conv1d", 1},
-                    {"conv2d", 1},
-                    {"conv3d", 1},
-                    {"conv_transpose1d", 1},
-                    {"conv_transpose2d", 1},
-                    {"linear", 1},
-                    {"embedding_bag", 0}}),
+      AtenFuncArgs(
+          {{"conv1d", 1},
+           {"conv2d", 1},
+           {"conv3d", 1},
+           {"conv_transpose1d", 1},
+           {"conv_transpose2d", 1},
+           {"linear", 1},
+           {"embedding_bag", 0}}),
       // embedding_bag - prim::CallFunction(%func, %input.1, %weight,
       // %offsets.1, %max_norm, %norm_type, %scale_grad_by_freq, %mode, %sparse,
       // %per_sample_weights.1, %include_last_offset)
@@ -289,12 +294,13 @@
 bool isBiasOfConvOrLinear(Value* v) {
   bool result = matchArgPattern(
       v,
-      AtenFuncArgs({{"conv1d", 2},
-                    {"conv2d", 2},
-                    {"conv3d", 2},
-                    {"conv_transpose1d", 2},
-                    {"conv_transpose2d", 2},
-                    {"linear", 2}}),
+      AtenFuncArgs(
+          {{"conv1d", 2},
+           {"conv2d", 2},
+           {"conv3d", 2},
+           {"conv_transpose1d", 2},
+           {"conv_transpose2d", 2},
+           {"linear", 2}}),
       CallFuncArgs({{"linear", 3}}));
   return result;
 }
diff --git a/torch/csrc/jit/passes/quantization/insert_observers.cpp b/torch/csrc/jit/passes/quantization/insert_observers.cpp
index bacd8cf..649869c 100644
--- a/torch/csrc/jit/passes/quantization/insert_observers.cpp
+++ b/torch/csrc/jit/passes/quantization/insert_observers.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/quantization/insert_observers.h>
+
 #include <torch/csrc/jit/frontend/schema_matching.h>
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 #include <torch/csrc/jit/jit_log.h>
diff --git a/torch/csrc/jit/passes/quantization/insert_quant_dequant.cpp b/torch/csrc/jit/passes/quantization/insert_quant_dequant.cpp
index 0d2c1c2..53a13b6 100644
--- a/torch/csrc/jit/passes/quantization/insert_quant_dequant.cpp
+++ b/torch/csrc/jit/passes/quantization/insert_quant_dequant.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/quantization/insert_quant_dequant.h>
+
 #include <c10/core/QScheme.h>
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 #include <torch/csrc/jit/jit_log.h>
@@ -1122,10 +1123,11 @@
         "q_zero_point");
     Node* dtype = insertQParam(
         graph, quantized_input, prim::dtype, IntType::get(), "dtype");
-    quant_inputs = {original_output,
-                    scale->output(),
-                    zero_point->output(),
-                    dtype->output()};
+    quant_inputs = {
+        original_output,
+        scale->output(),
+        zero_point->output(),
+        dtype->output()};
   }
   Node* quant = insertQuant(
       graph, quant_inputs, quant_kind, original_output->debugName() + ".quant");
diff --git a/torch/csrc/jit/passes/requires_grad_analysis.cpp b/torch/csrc/jit/passes/requires_grad_analysis.cpp
index deb7837..12c1b7b 100644
--- a/torch/csrc/jit/passes/requires_grad_analysis.cpp
+++ b/torch/csrc/jit/passes/requires_grad_analysis.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/requires_grad_analysis.h>
+
 #include <ATen/core/jit_type.h>
 #include <torch/csrc/autograd/autograd.h>
 #include <torch/csrc/jit/ir/constants.h>
diff --git a/torch/csrc/jit/passes/shape_analysis.cpp b/torch/csrc/jit/passes/shape_analysis.cpp
index fd076ca..a720f93 100644
--- a/torch/csrc/jit/passes/shape_analysis.cpp
+++ b/torch/csrc/jit/passes/shape_analysis.cpp
@@ -1270,9 +1270,10 @@
                       type->withScalarType(maybe_dtype_option->toScalarType())};
                 }
                 if (type->scalarType()) {
-                  return {at::isFloatingType(*type->scalarType())
-                              ? type
-                              : type->withScalarType(at::kLong)};
+                  return {
+                      at::isFloatingType(*type->scalarType())
+                          ? type
+                          : type->withScalarType(at::kLong)};
                 } else {
                   return {type};
                 }
diff --git a/torch/csrc/jit/passes/specialize_autogradzero.cpp b/torch/csrc/jit/passes/specialize_autogradzero.cpp
index 38d0913..6e660a3 100644
--- a/torch/csrc/jit/passes/specialize_autogradzero.cpp
+++ b/torch/csrc/jit/passes/specialize_autogradzero.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/specialize_autogradzero.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/ir/ir.h>
 #include <torch/csrc/jit/jit_log.h>
diff --git a/torch/csrc/jit/passes/subgraph_rewrite.cpp b/torch/csrc/jit/passes/subgraph_rewrite.cpp
index ff066a8..45d07fe 100644
--- a/torch/csrc/jit/passes/subgraph_rewrite.cpp
+++ b/torch/csrc/jit/passes/subgraph_rewrite.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/subgraph_rewrite.h>
+
 #include <torch/csrc/jit/ir/irparser.h>
 #include <torch/csrc/jit/ir/subgraph_matcher.h>
 
diff --git a/torch/csrc/jit/passes/tensorexpr_fuser.cpp b/torch/csrc/jit/passes/tensorexpr_fuser.cpp
index b8b6f72..7b7de1e 100644
--- a/torch/csrc/jit/passes/tensorexpr_fuser.cpp
+++ b/torch/csrc/jit/passes/tensorexpr_fuser.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/tensorexpr_fuser.h>
+
 #include <ATen/record_function.h>
 #include <c10/util/FunctionRef.h>
 #include <torch/csrc/jit/codegen/fuser/interface.h>
diff --git a/torch/csrc/jit/passes/update_differentiable_graph_requires_grad.cpp b/torch/csrc/jit/passes/update_differentiable_graph_requires_grad.cpp
index 805f9e2..5013dcf 100644
--- a/torch/csrc/jit/passes/update_differentiable_graph_requires_grad.cpp
+++ b/torch/csrc/jit/passes/update_differentiable_graph_requires_grad.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/update_differentiable_graph_requires_grad.h>
+
 #include <torch/csrc/jit/ir/ir.h>
 #include <torch/csrc/jit/passes/utils/subgraph_utils.h>
 
diff --git a/torch/csrc/jit/passes/utils/check_alias_annotation.cpp b/torch/csrc/jit/passes/utils/check_alias_annotation.cpp
index 40aa7db..78ab4a6 100644
--- a/torch/csrc/jit/passes/utils/check_alias_annotation.cpp
+++ b/torch/csrc/jit/passes/utils/check_alias_annotation.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/utils/check_alias_annotation.h>
+
 #include <torch/csrc/jit/passes/constant_propagation.h>
 #include <torch/csrc/jit/passes/normalize_ops.h>
 #include <torch/csrc/jit/runtime/operator.h>
diff --git a/torch/csrc/jit/passes/utils/subgraph_utils.cpp b/torch/csrc/jit/passes/utils/subgraph_utils.cpp
index d104bb7..bee1d50 100644
--- a/torch/csrc/jit/passes/utils/subgraph_utils.cpp
+++ b/torch/csrc/jit/passes/utils/subgraph_utils.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/passes/utils/subgraph_utils.h>
+
 #include <torch/csrc/jit/passes/canonicalize.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/python/python_custom_class.cpp b/torch/csrc/jit/python/python_custom_class.cpp
index 9809b85..a4120d4 100644
--- a/torch/csrc/jit/python/python_custom_class.cpp
+++ b/torch/csrc/jit/python/python_custom_class.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/python/python_custom_class.h>
+
 #include <torch/csrc/jit/frontend/sugared_value.h>
 
 #include <fmt/format.h>
diff --git a/torch/csrc/jit/python/python_ir.cpp b/torch/csrc/jit/python/python_ir.cpp
index b6f6169..495de26 100644
--- a/torch/csrc/jit/python/python_ir.cpp
+++ b/torch/csrc/jit/python/python_ir.cpp
@@ -568,14 +568,12 @@
       .AS(removeAttribute)
       .AS(attributeNames)
 #undef AS
-#define CREATE_ACCESSOR(Kind, method)                          \
-  def(#method "_",                                             \
-      [](Node& n, const char* name, Kind##Attr::ValueType v) { \
-        return n.method##_(Symbol::attr(name), std::move(v));  \
-      })                                                       \
-      .def(#method, [](Node& n, const char* name) {            \
-        return n.method(Symbol::attr(name));                   \
-      })
+#define CREATE_ACCESSOR(Kind, method)                                       \
+  def(#method "_", [](Node& n, const char* name, Kind##Attr::ValueType v) { \
+    return n.method##_(Symbol::attr(name), std::move(v));                   \
+  }).def(#method, [](Node& n, const char* name) {                           \
+    return n.method(Symbol::attr(name));                                    \
+  })
       .CREATE_ACCESSOR(Float, f)
       .CREATE_ACCESSOR(Floats, fs)
       .CREATE_ACCESSOR(String, s)
diff --git a/torch/csrc/jit/python/python_sugared_value.cpp b/torch/csrc/jit/python/python_sugared_value.cpp
index 056e23d..f6f6fbf 100644
--- a/torch/csrc/jit/python/python_sugared_value.cpp
+++ b/torch/csrc/jit/python/python_sugared_value.cpp
@@ -223,13 +223,14 @@
     Function& m,
     const std::string& field) {
   // List of all the cuda operators which are supported in JIT
-  const std::unordered_set<std::string> cuda_ops = {"current_stream",
-                                                    "default_stream",
-                                                    "_current_device",
-                                                    "_set_device",
-                                                    "device_index",
-                                                    "device_count",
-                                                    "set_stream"};
+  const std::unordered_set<std::string> cuda_ops = {
+      "current_stream",
+      "default_stream",
+      "_current_device",
+      "_set_device",
+      "device_index",
+      "device_count",
+      "set_stream"};
 
   if (cuda_ops.find(field) != cuda_ops.end()) {
     return std::make_shared<BuiltinFunction>(Symbol::cuda(field), c10::nullopt);
diff --git a/torch/csrc/jit/runtime/autodiff.cpp b/torch/csrc/jit/runtime/autodiff.cpp
index 9e6ac44..c3eebdf 100644
--- a/torch/csrc/jit/runtime/autodiff.cpp
+++ b/torch/csrc/jit/runtime/autodiff.cpp
@@ -253,12 +253,13 @@
           graph->insertNode(graph->createTupleUnpack(backward_value));
       auto tuple_outputs = tuple_unpack_node->outputs();
       AT_ASSERT(tuple_outputs.size() == size_t(3));
-      return {tuple_outputs[0],
-              tuple_outputs[1],
-              nullptr,
-              tuple_outputs[2],
-              nullptr,
-              nullptr};
+      return {
+          tuple_outputs[0],
+          tuple_outputs[1],
+          nullptr,
+          tuple_outputs[2],
+          nullptr,
+          nullptr};
 
     } else if (
         node->matches(
@@ -282,14 +283,15 @@
           graph->insertNode(graph->createTupleUnpack(backward_value));
       auto tuple_outputs = tuple_unpack_node->outputs();
       AT_ASSERT(tuple_outputs.size() == size_t(3));
-      return {tuple_outputs[0],
-              tuple_outputs[1],
-              tuple_outputs[2],
-              nullptr,
-              nullptr,
-              nullptr,
-              nullptr,
-              nullptr};
+      return {
+          tuple_outputs[0],
+          tuple_outputs[1],
+          tuple_outputs[2],
+          nullptr,
+          nullptr,
+          nullptr,
+          nullptr,
+          nullptr};
     }
 
     throw std::runtime_error(
diff --git a/torch/csrc/jit/runtime/logging.h b/torch/csrc/jit/runtime/logging.h
index ce03394..a59e809 100644
--- a/torch/csrc/jit/runtime/logging.h
+++ b/torch/csrc/jit/runtime/logging.h
@@ -74,10 +74,11 @@
     "pytorch_runtime.execution_plan_cache_miss";
 
 inline std::vector<const char*> allRuntimeCounters() {
-  return {GRAPH_EXECUTORS_CONSTRUCTED,
-          GRAPH_EXECUTOR_INVOCATIONS,
-          EXECUTION_PLAN_CACHE_HIT,
-          EXECUTION_PLAN_CACHE_MISS};
+  return {
+      GRAPH_EXECUTORS_CONSTRUCTED,
+      GRAPH_EXECUTOR_INVOCATIONS,
+      EXECUTION_PLAN_CACHE_HIT,
+      EXECUTION_PLAN_CACHE_MISS};
 }
 
 } // namespace runtime_counters
diff --git a/torch/csrc/jit/runtime/operator.cpp b/torch/csrc/jit/runtime/operator.cpp
index 1964679..49db701 100644
--- a/torch/csrc/jit/runtime/operator.cpp
+++ b/torch/csrc/jit/runtime/operator.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/operator.h>
+
 #include <ATen/ATen.h>
 #include <ATen/core/alias_info.h>
 #include <torch/csrc/jit/frontend/edit_distance.h>
diff --git a/torch/csrc/jit/runtime/profiling_graph_executor_impl.cpp b/torch/csrc/jit/runtime/profiling_graph_executor_impl.cpp
index dec64c4..aed5d9c 100644
--- a/torch/csrc/jit/runtime/profiling_graph_executor_impl.cpp
+++ b/torch/csrc/jit/runtime/profiling_graph_executor_impl.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/profiling_graph_executor_impl.h>
+
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/bailout_graph.h>
 #include <torch/csrc/jit/passes/batch_mm.h>
diff --git a/torch/csrc/jit/runtime/profiling_record.cpp b/torch/csrc/jit/runtime/profiling_record.cpp
index 12e805c..273b242 100644
--- a/torch/csrc/jit/runtime/profiling_record.cpp
+++ b/torch/csrc/jit/runtime/profiling_record.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/profiling_record.h>
+
 #include <ATen/core/interned_strings.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/clear_profiling.h>
diff --git a/torch/csrc/jit/runtime/register_prim_ops.cpp b/torch/csrc/jit/runtime/register_prim_ops.cpp
index fe75ec5..07c7445 100644
--- a/torch/csrc/jit/runtime/register_prim_ops.cpp
+++ b/torch/csrc/jit/runtime/register_prim_ops.cpp
@@ -112,7 +112,7 @@
          // depends on the type hint and input. The implementation of this
          // operator below is intended to be as close to the Python
          // implementation in torch/csrc/utils/tensor_list.cpp as possible.
-         [](const Node * /*node*/) -> Operation {
+         [](const Node* /*node*/) -> Operation {
            return [](Stack* stack) {
              int elem_ty_val;
              int dim_val;
diff --git a/torch/csrc/jit/runtime/register_prim_ops_fulljit.cpp b/torch/csrc/jit/runtime/register_prim_ops_fulljit.cpp
index 8361fb3..2de6299 100644
--- a/torch/csrc/jit/runtime/register_prim_ops_fulljit.cpp
+++ b/torch/csrc/jit/runtime/register_prim_ops_fulljit.cpp
@@ -66,7 +66,7 @@
          aliasAnalysisSpecialCase()),
      Operator(
          prim::TypeCheck /* (...)  -> (..., bool) */,
-         [](const Node * /* node */) -> Operation {
+         [](const Node* /* node */) -> Operation {
            return [](Stack* /* stack */) {
              AT_ERROR("prim::TypeCheck not yet implemented"); // NOLINT
            };
diff --git a/torch/csrc/jit/runtime/slice_indices_adjust.cpp b/torch/csrc/jit/runtime/slice_indices_adjust.cpp
index e71d6ba..8c0d391 100644
--- a/torch/csrc/jit/runtime/slice_indices_adjust.cpp
+++ b/torch/csrc/jit/runtime/slice_indices_adjust.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/slice_indices_adjust.h>
+
 #include <c10/util/Exception.h>
 #include <cstdint>
 
diff --git a/torch/csrc/jit/runtime/static/fusion.cpp b/torch/csrc/jit/runtime/static/fusion.cpp
index 5050b8d..b4ef21e 100644
--- a/torch/csrc/jit/runtime/static/fusion.cpp
+++ b/torch/csrc/jit/runtime/static/fusion.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/static/fusion.h>
+
 #include <ATen/core/interned_strings.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/passes/dead_code_elimination.h>
diff --git a/torch/csrc/jit/runtime/static/impl.cpp b/torch/csrc/jit/runtime/static/impl.cpp
index a398f8e..8160e3a 100644
--- a/torch/csrc/jit/runtime/static/impl.cpp
+++ b/torch/csrc/jit/runtime/static/impl.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/static/impl.h>
+
 #include <ATen/core/LegacyTypeDispatch.h>
 #include <ATen/core/interned_strings.h>
 #include <c10/core/CPUAllocator.h>
@@ -306,8 +307,8 @@
     const std::shared_ptr<torch::jit::Graph>& graph,
     const std::unordered_map<Value*, size_t>& value_to_reg,
     std::vector<size_t>& internals) {
-  std::unordered_set<Value*> outputs{graph->outputs().begin(),
-                                     graph->outputs().end()};
+  std::unordered_set<Value*> outputs{
+      graph->outputs().begin(), graph->outputs().end()};
   for (Node* node : graph->nodes()) {
     if (node->kind() != prim::Constant) {
       for (Value* output : node->outputs()) {
diff --git a/torch/csrc/jit/runtime/static/init.cpp b/torch/csrc/jit/runtime/static/init.cpp
index 4799b5b..9027549 100644
--- a/torch/csrc/jit/runtime/static/init.cpp
+++ b/torch/csrc/jit/runtime/static/init.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/static/init.h>
+
 #include <torch/csrc/jit/passes/freeze_module.h>
 #include <torch/csrc/jit/runtime/static/fusion.h>
 #include <torch/csrc/jit/runtime/static/impl.h>
diff --git a/torch/csrc/jit/runtime/static/ops.cpp b/torch/csrc/jit/runtime/static/ops.cpp
index 3e21224..2d4dc67 100644
--- a/torch/csrc/jit/runtime/static/ops.cpp
+++ b/torch/csrc/jit/runtime/static/ops.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/static/ops.h>
+
 #include <ATen/NativeFunctions.h>
 #include <torch/csrc/jit/ir/ir.h>
 #include <torch/csrc/jit/runtime/vararg_functions.h>
diff --git a/torch/csrc/jit/runtime/static/passes.cpp b/torch/csrc/jit/runtime/static/passes.cpp
index bbaaa66..0ef2e3a 100644
--- a/torch/csrc/jit/runtime/static/passes.cpp
+++ b/torch/csrc/jit/runtime/static/passes.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/static/passes.h>
+
 #include <torch/csrc/jit/passes/subgraph_rewrite.h>
 
 namespace torch {
diff --git a/torch/csrc/jit/runtime/symbolic_script.cpp b/torch/csrc/jit/runtime/symbolic_script.cpp
index 6887be5..1113d45 100644
--- a/torch/csrc/jit/runtime/symbolic_script.cpp
+++ b/torch/csrc/jit/runtime/symbolic_script.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/symbolic_script.h>
+
 #include <torch/csrc/jit/frontend/ir_emitter.h>
 #include <torch/csrc/jit/runtime/operator.h>
 
diff --git a/torch/csrc/jit/runtime/vararg_functions.cpp b/torch/csrc/jit/runtime/vararg_functions.cpp
index 930ea5e..2bb8840 100644
--- a/torch/csrc/jit/runtime/vararg_functions.cpp
+++ b/torch/csrc/jit/runtime/vararg_functions.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/runtime/vararg_functions.h>
+
 #include <ATen/ATen.h>
 
 namespace torch {
@@ -186,8 +187,9 @@
 }
 
 void tupleConstruct(Stack& stack, size_t num_inputs) {
-  std::vector<IValue> elems{std::make_move_iterator(stack.end() - num_inputs),
-                            std::make_move_iterator(stack.end())};
+  std::vector<IValue> elems{
+      std::make_move_iterator(stack.end() - num_inputs),
+      std::make_move_iterator(stack.end())};
   drop(stack, num_inputs);
   push(stack, c10::ivalue::Tuple::create(std::move(elems)));
 }
@@ -196,8 +198,9 @@
     Stack& stack,
     at::TupleTypePtr type,
     size_t num_inputs) {
-  std::vector<IValue> elems{std::make_move_iterator(stack.end() - num_inputs),
-                            std::make_move_iterator(stack.end())};
+  std::vector<IValue> elems{
+      std::make_move_iterator(stack.end() - num_inputs),
+      std::make_move_iterator(stack.end())};
   drop(stack, num_inputs);
   push(
       stack,
diff --git a/torch/csrc/jit/serialization/export.cpp b/torch/csrc/jit/serialization/export.cpp
index b69a5b7..2afccb3 100644
--- a/torch/csrc/jit/serialization/export.cpp
+++ b/torch/csrc/jit/serialization/export.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/serialization/export.h>
+
 #include <torch/csrc/autograd/symbolic.h>
 #include <torch/csrc/jit/jit_log.h>
 #include <torch/csrc/jit/serialization/import_export_constants.h>
diff --git a/torch/csrc/jit/serialization/export_module.cpp b/torch/csrc/jit/serialization/export_module.cpp
index 493f3e5..0dd3ba2 100644
--- a/torch/csrc/jit/serialization/export_module.cpp
+++ b/torch/csrc/jit/serialization/export_module.cpp
@@ -169,9 +169,10 @@
         auto method_name_idx =
             code.constant_table().size() + method_names.size();
         method_names.emplace_back(node->s(attr::name));
-        Instruction new_instr{INTERFACE_CALL,
-                              static_cast<int32_t>(method_name_idx),
-                              static_cast<uint16_t>(node->inputs().size())};
+        Instruction new_instr{
+            INTERFACE_CALL,
+            static_cast<int32_t>(method_name_idx),
+            static_cast<uint16_t>(node->inputs().size())};
         instructions_copy[i] = new_instr;
       } else {
         TORCH_INTERNAL_ASSERT(
@@ -271,11 +272,12 @@
   // register size
   auto register_size = static_cast<int>(code.register_size());
 
-  auto table = Table({{"instructions", Tup(instructions)},
-                      {"operators", Tup(operators)},
-                      {"constants", Tup(deduplicated_constants)},
-                      {"types", Tup(types)},
-                      {"register_size", register_size}});
+  auto table = Table(
+      {{"instructions", Tup(instructions)},
+       {"operators", Tup(operators)},
+       {"constants", Tup(deduplicated_constants)},
+       {"types", Tup(types)},
+       {"register_size", register_size}});
   auto bytecode_vals = Tup({func.qualname().qualifiedName(), table});
 
   c10::optional<IValue> debug_info_vals;
diff --git a/torch/csrc/jit/serialization/import.cpp b/torch/csrc/jit/serialization/import.cpp
index ed5dde0..159ded3 100644
--- a/torch/csrc/jit/serialization/import.cpp
+++ b/torch/csrc/jit/serialization/import.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/serialization/import.h>
+
 #include <ATen/core/functional.h>
 #include <ATen/core/ivalue_inl.h>
 #include <c10/util/Exception.h>
diff --git a/torch/csrc/jit/serialization/import_export_helpers.cpp b/torch/csrc/jit/serialization/import_export_helpers.cpp
index 3f09d1f..e12e937 100644
--- a/torch/csrc/jit/serialization/import_export_helpers.cpp
+++ b/torch/csrc/jit/serialization/import_export_helpers.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/serialization/import_export_helpers.h>
+
 #include <caffe2/serialize/inline_container.h>
 #include <torch/csrc/jit/frontend/source_range.h>
 #include <torch/csrc/jit/serialization/source_range_serialization_impl.h>
diff --git a/torch/csrc/jit/serialization/pickle.cpp b/torch/csrc/jit/serialization/pickle.cpp
index 0e3284b..27b5fe2 100644
--- a/torch/csrc/jit/serialization/pickle.cpp
+++ b/torch/csrc/jit/serialization/pickle.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/serialization/pickle.h>
+
 #include <ATen/core/ivalue.h>
 #include <caffe2/serialize/inline_container.h>
 #include <torch/csrc/WindowsTorchApiMacro.h>
diff --git a/torch/csrc/jit/serialization/python_print.cpp b/torch/csrc/jit/serialization/python_print.cpp
index 18d656c..8b348de 100644
--- a/torch/csrc/jit/serialization/python_print.cpp
+++ b/torch/csrc/jit/serialization/python_print.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/serialization/python_print.h>
+
 #include <ATen/core/qualified_name.h>
 #include <c10/util/Exception.h>
 #include <c10/util/StringUtil.h>
diff --git a/torch/csrc/jit/serialization/source_range_serialization.cpp b/torch/csrc/jit/serialization/source_range_serialization.cpp
index 16d3de4..9f158e4 100644
--- a/torch/csrc/jit/serialization/source_range_serialization.cpp
+++ b/torch/csrc/jit/serialization/source_range_serialization.cpp
@@ -87,8 +87,8 @@
 std::vector<char> SourceRangePickler::pickle(const SourceRangeRecords& ranges) {
   std::vector<c10::IValue> ivalues;
   for (const auto& range : ranges) {
-    std::vector<c10::IValue> row_elems{(int64_t)range.bytes,
-                                       srs->serialize(range.range)};
+    std::vector<c10::IValue> row_elems{
+        (int64_t)range.bytes, srs->serialize(range.range)};
     ivalues.emplace_back(c10::ivalue::Tuple::create(std::move(row_elems)));
   }
   std::vector<at::Tensor> table;
diff --git a/torch/csrc/jit/tensorexpr/cuda_codegen.h b/torch/csrc/jit/tensorexpr/cuda_codegen.h
index b070dd3..acaa5dd 100644
--- a/torch/csrc/jit/tensorexpr/cuda_codegen.h
+++ b/torch/csrc/jit/tensorexpr/cuda_codegen.h
@@ -73,12 +73,14 @@
  public:
   explicit GPUMetaVarRewriter(const CudaAnalysis* cuda_analysis)
       : cuda_analysis_(cuda_analysis) {
-    gpu_block_vars_ = {new Var("blockIdx.x", kInt),
-                       new Var("blockIdx.y", kInt),
-                       new Var("blockIdx.z", kInt)};
-    gpu_thread_vars_ = {new Var("threadIdx.x", kInt),
-                        new Var("threadIdx.y", kInt),
-                        new Var("threadIdx.z", kInt)};
+    gpu_block_vars_ = {
+        new Var("blockIdx.x", kInt),
+        new Var("blockIdx.y", kInt),
+        new Var("blockIdx.z", kInt)};
+    gpu_thread_vars_ = {
+        new Var("threadIdx.x", kInt),
+        new Var("threadIdx.y", kInt),
+        new Var("threadIdx.z", kInt)};
 
     current_block_reach_ = {new IntImm(1), new IntImm(1), new IntImm(1)};
     current_thread_reach_ = {new IntImm(1), new IntImm(1), new IntImm(1)};
diff --git a/torch/csrc/jit/tensorexpr/llvm_codegen.cpp b/torch/csrc/jit/tensorexpr/llvm_codegen.cpp
index adc3be9..c1e7508 100644
--- a/torch/csrc/jit/tensorexpr/llvm_codegen.cpp
+++ b/torch/csrc/jit/tensorexpr/llvm_codegen.cpp
@@ -1,6 +1,7 @@
 #ifdef TORCH_ENABLE_LLVM
 
 #include <torch/csrc/jit/tensorexpr/llvm_codegen.h>
+
 #include <c10/util/Exception.h>
 #include <torch/csrc/jit/tensorexpr/llvm_jit.h>
 
diff --git a/torch/csrc/jit/tensorexpr/loopnest.cpp b/torch/csrc/jit/tensorexpr/loopnest.cpp
index c2b274a..f4f346f 100644
--- a/torch/csrc/jit/tensorexpr/loopnest.cpp
+++ b/torch/csrc/jit/tensorexpr/loopnest.cpp
@@ -2296,8 +2296,8 @@
 
   For* root_for = nullptr;
   For* target_for = nullptr;
-  std::set<const Var*> reduce_args = {reduce_op->reduce_args().begin(),
-                                      reduce_op->reduce_args().end()};
+  std::set<const Var*> reduce_args = {
+      reduce_op->reduce_args().begin(), reduce_op->reduce_args().end()};
 
   // Store loops below the target point.
   std::vector<const For*> output_loops;
diff --git a/torch/csrc/jit/testing/hooks_for_testing.cpp b/torch/csrc/jit/testing/hooks_for_testing.cpp
index 29dbc69..553938a 100644
--- a/torch/csrc/jit/testing/hooks_for_testing.cpp
+++ b/torch/csrc/jit/testing/hooks_for_testing.cpp
@@ -1,4 +1,5 @@
 #include <torch/csrc/jit/testing/hooks_for_testing.h>
+
 #include <torch/csrc/jit/api/module.h>
 
 namespace torch {