used clang format to clean up

src/targets/cpu/cpu_lowering.cpp

@@ -44,18 +44,19 @@ struct cpu_batch_norm_inference
     {
         argument output{output_shape};
 
-        double epsilon = op.epsilon;
-        auto input = args[0];
-        auto mini_batch_mean = args[1].at<float>();
+        double epsilon           = op.epsilon;
+        auto input               = args[0];
+        auto mini_batch_mean     = args[1].at<float>();
         auto mini_batch_variance = args[2].at<float>();
-        auto gamma = args[3].at<float>();
-        auto beta = args[4].at<float>();
+        auto gamma               = args[3].at<float>();
+        auto beta                = args[4].at<float>();
 
-	    visit_all(output, input) ([&](auto result, auto buffer) {
+        visit_all(output, input)([&](auto result, auto buffer) {
             std::transform(buffer.begin(), buffer.end(), result.begin(), [&](auto x) {
-                return gamma * (x - mini_batch_mean) / std::sqrt(mini_batch_variance + epsilon) + beta;
+                return gamma * (x - mini_batch_mean) / std::sqrt(mini_batch_variance + epsilon) +
+                       beta;
             });
-	    });
+        });
 
         return output;
     }
@@ -515,10 +516,11 @@ struct cpu_apply
     {
         apply_map["convolution"] = extend_op<cpu_convolution, convolution>();
         apply_map["gemm"]        = extend_op<cpu_gemm, gemm>();
-        apply_map["batch_norm_inference"]        = extend_op<cpu_batch_norm_inference, batch_norm_inference>();
-        apply_map["reshape"]     = extend_op<cpu_reshape, reshape>();
-        apply_map["contiguous"]  = extend_op<cpu_contiguous, contiguous>();
-        apply_map["transpose"]   = extend_op<cpu_transpose, transpose>();
+        apply_map["batch_norm_inference"] =
+            extend_op<cpu_batch_norm_inference, batch_norm_inference>();
+        apply_map["reshape"]    = extend_op<cpu_reshape, reshape>();
+        apply_map["contiguous"] = extend_op<cpu_contiguous, contiguous>();
+        apply_map["transpose"]  = extend_op<cpu_transpose, transpose>();
 
         apply_map["identity"] = simple_op<cpu_unary<identity_op>>();
         apply_map["tanh"]     = simple_op<cpu_unary<tanh_op>>();

test/cpu_ops_test.cpp

@@ -10,17 +10,18 @@ void batch_norm_inference_test()
 {
     migraph::program p;
     migraph::shape s{migraph::shape::float_type, {4}};
-    auto x = p.add_literal(migraph::literal{s, {1, 2, 3, 4}});
-    auto gamma = p.add_literal(migraph::literal{s, {1}});
-    auto beta = p.add_literal(migraph::literal{s, {0}});
-    auto mean = p.add_literal(migraph::literal{s, {0}});
+    auto x        = p.add_literal(migraph::literal{s, {1, 2, 3, 4}});
+    auto gamma    = p.add_literal(migraph::literal{s, {1}});
+    auto beta     = p.add_literal(migraph::literal{s, {0}});
+    auto mean     = p.add_literal(migraph::literal{s, {0}});
     auto variance = p.add_literal(migraph::literal{s, {1}});
     p.add_instruction(migraph::batch_norm_inference{}, x, mean, variance, gamma, beta);
     p.compile(migraph::cpu::cpu_target{});
     auto result = p.eval({});
     std::vector<float> result_vector(4);
-    result.visit([&](auto output) {result_vector.assign(output.begin(), output.end()); });
-    std::vector<float> gold = { 1 / (1 + 1.0e-6), 2 / (1 + 1.0e-6), 3 / (1 + 1.0e-6), 4 / (1 + 1.0e-6)};
+    result.visit([&](auto output) { result_vector.assign(output.begin(), output.end()); });
+    std::vector<float> gold = {
+        1 / (1 + 1.0e-6), 2 / (1 + 1.0e-6), 3 / (1 + 1.0e-6), 4 / (1 + 1.0e-6)};
     EXPECT(test::verify_range(result_vector, gold));
 }