diff --git a/backends/cadence/fused_quant/op_add.cpp b/backends/cadence/fused_quant/op_add.cpp
index 62e58c71c83..1aea2ccfb6c 100644
--- a/backends/cadence/fused_quant/op_add.cpp
+++ b/backends/cadence/fused_quant/op_add.cpp
@@ -14,10 +14,10 @@ namespace cadence {
 namespace fused_quant {
 namespace native {
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/op_add.h b/backends/cadence/fused_quant/op_add.h
index 9db1e907294..b32710f41de 100644
--- a/backends/cadence/fused_quant/op_add.h
+++ b/backends/cadence/fused_quant/op_add.h
@@ -19,19 +19,18 @@ executorch::aten::Tensor& add_out(
     executorch::runtime::KernelRuntimeContext& ctx,
     const executorch::aten::Tensor& inp,
     const executorch::aten::Tensor& other,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_zero_point,
+    const std::optional<executorch::aten::Tensor>& inp_scale,
+    const std::optional<executorch::aten::Tensor>& inp_zero_point,
     executorch::aten::ScalarType inp_dtype,
     int64_t inp_quant_min,
     int64_t inp_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& other_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>&
-        other_zero_point,
+    const std::optional<executorch::aten::Tensor>& other_scale,
+    const std::optional<executorch::aten::Tensor>& other_zero_point,
     executorch::aten::ScalarType other_dtype,
     int64_t other_quant_min,
     int64_t other_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_zero_point,
+    const std::optional<executorch::aten::Tensor>& out_scale,
+    const std::optional<executorch::aten::Tensor>& out_zero_point,
     executorch::aten::ScalarType out_dtype,
     int64_t out_quant_min,
     int64_t out_quant_max,
diff --git a/backends/cadence/fused_quant/op_bmm.cpp b/backends/cadence/fused_quant/op_bmm.cpp
index 7204ab6c88f..8d071b48a33 100644
--- a/backends/cadence/fused_quant/op_bmm.cpp
+++ b/backends/cadence/fused_quant/op_bmm.cpp
@@ -14,10 +14,10 @@ namespace cadence {
 namespace fused_quant {
 namespace native {
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/op_bmm.h b/backends/cadence/fused_quant/op_bmm.h
index ef9598eac98..c6a4502983f 100644
--- a/backends/cadence/fused_quant/op_bmm.h
+++ b/backends/cadence/fused_quant/op_bmm.h
@@ -19,19 +19,18 @@ executorch::aten::Tensor& bmm_out(
     executorch::runtime::KernelRuntimeContext& ctx,
     const executorch::aten::Tensor& inp,
     const executorch::aten::Tensor& other,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_zero_point,
+    const std::optional<executorch::aten::Tensor>& inp_scale,
+    const std::optional<executorch::aten::Tensor>& inp_zero_point,
     executorch::aten::ScalarType inp_dtype,
     int64_t inp_quant_min,
     int64_t inp_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& other_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>&
-        other_zero_point,
+    const std::optional<executorch::aten::Tensor>& other_scale,
+    const std::optional<executorch::aten::Tensor>& other_zero_point,
     executorch::aten::ScalarType other_dtype,
     int64_t other_quant_min,
     int64_t other_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_zero_point,
+    const std::optional<executorch::aten::Tensor>& out_scale,
+    const std::optional<executorch::aten::Tensor>& out_zero_point,
     executorch::aten::ScalarType out_dtype,
     int64_t out_quant_min,
     int64_t out_quant_max,
diff --git a/backends/cadence/fused_quant/op_hardswish.cpp b/backends/cadence/fused_quant/op_hardswish.cpp
index 452ea90a405..4b968cebe6c 100644
--- a/backends/cadence/fused_quant/op_hardswish.cpp
+++ b/backends/cadence/fused_quant/op_hardswish.cpp
@@ -16,10 +16,10 @@ namespace cadence {
 namespace fused_quant {
 namespace native {
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/op_hardswish.h b/backends/cadence/fused_quant/op_hardswish.h
index ba9e09da23c..de7d88b427b 100644
--- a/backends/cadence/fused_quant/op_hardswish.h
+++ b/backends/cadence/fused_quant/op_hardswish.h
@@ -18,13 +18,13 @@ namespace native {
 executorch::aten::Tensor& hardswish_out(
     executorch::runtime::KernelRuntimeContext& ctx,
     const executorch::aten::Tensor& inp,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_zero_point,
+    const std::optional<executorch::aten::Tensor>& inp_scale,
+    const std::optional<executorch::aten::Tensor>& inp_zero_point,
     executorch::aten::ScalarType inp_dtype,
     int64_t inp_quant_min,
     int64_t inp_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_zero_point,
+    const std::optional<executorch::aten::Tensor>& out_scale,
+    const std::optional<executorch::aten::Tensor>& out_zero_point,
     executorch::aten::ScalarType out_dtype,
     int64_t out_quant_min,
     int64_t out_quant_max,
diff --git a/backends/cadence/fused_quant/op_mul.cpp b/backends/cadence/fused_quant/op_mul.cpp
index 3d071f7c2da..a2595104ae8 100644
--- a/backends/cadence/fused_quant/op_mul.cpp
+++ b/backends/cadence/fused_quant/op_mul.cpp
@@ -14,10 +14,10 @@ namespace cadence {
 namespace fused_quant {
 namespace native {
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/op_mul.h b/backends/cadence/fused_quant/op_mul.h
index f7afa016b79..62314c98003 100644
--- a/backends/cadence/fused_quant/op_mul.h
+++ b/backends/cadence/fused_quant/op_mul.h
@@ -19,19 +19,18 @@ executorch::aten::Tensor& mul_out(
     executorch::runtime::KernelRuntimeContext& ctx,
     const executorch::aten::Tensor& inp,
     const executorch::aten::Tensor& other,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_zero_point,
+    const std::optional<executorch::aten::Tensor>& inp_scale,
+    const std::optional<executorch::aten::Tensor>& inp_zero_point,
     executorch::aten::ScalarType inp_dtype,
     int64_t inp_quant_min,
     int64_t inp_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& other_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>&
-        other_zero_point,
+    const std::optional<executorch::aten::Tensor>& other_scale,
+    const std::optional<executorch::aten::Tensor>& other_zero_point,
     executorch::aten::ScalarType other_dtype,
     int64_t other_quant_min,
     int64_t other_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_zero_point,
+    const std::optional<executorch::aten::Tensor>& out_scale,
+    const std::optional<executorch::aten::Tensor>& out_zero_point,
     executorch::aten::ScalarType out_dtype,
     int64_t out_quant_min,
     int64_t out_quant_max,
diff --git a/backends/cadence/fused_quant/op_relu.cpp b/backends/cadence/fused_quant/op_relu.cpp
index ebe7933a7b9..e8e58522d2e 100644
--- a/backends/cadence/fused_quant/op_relu.cpp
+++ b/backends/cadence/fused_quant/op_relu.cpp
@@ -16,10 +16,10 @@ namespace cadence {
 namespace fused_quant {
 namespace native {
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/op_relu.h b/backends/cadence/fused_quant/op_relu.h
index e8527c7633f..522144eacd0 100644
--- a/backends/cadence/fused_quant/op_relu.h
+++ b/backends/cadence/fused_quant/op_relu.h
@@ -18,13 +18,13 @@ namespace native {
 executorch::aten::Tensor& relu_out(
     executorch::runtime::KernelRuntimeContext& ctx,
     const executorch::aten::Tensor& inp,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& inp_zero_point,
+    const std::optional<executorch::aten::Tensor>& inp_scale,
+    const std::optional<executorch::aten::Tensor>& inp_zero_point,
     executorch::aten::ScalarType inp_dtype,
     int64_t inp_quant_min,
     int64_t inp_quant_max,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_scale,
-    const executorch::aten::optional<executorch::aten::Tensor>& out_zero_point,
+    const std::optional<executorch::aten::Tensor>& out_scale,
+    const std::optional<executorch::aten::Tensor>& out_zero_point,
     executorch::aten::ScalarType out_dtype,
     int64_t out_quant_min,
     int64_t out_quant_max,
diff --git a/backends/cadence/fused_quant/quant_utils.h b/backends/cadence/fused_quant/quant_utils.h
index fff669a9e0e..78884bfcceb 100644
--- a/backends/cadence/fused_quant/quant_utils.h
+++ b/backends/cadence/fused_quant/quant_utils.h
@@ -64,8 +64,8 @@ struct QParams {
 };
 
 inline QParams extract_qparams(
-    const executorch::aten::optional<executorch::aten::Tensor>& scale_tensor,
-    const executorch::aten::optional<executorch::aten::Tensor>& zp_tensor,
+    const std::optional<executorch::aten::Tensor>& scale_tensor,
+    const std::optional<executorch::aten::Tensor>& zp_tensor,
     int64_t quant_min,
     int64_t quant_max,
     const executorch::aten::Tensor& data_tensor) {
diff --git a/backends/cadence/fused_quant/tests/test_op_add.cpp b/backends/cadence/fused_quant/tests/test_op_add.cpp
index dca110cf0e1..61124f0b9b2 100644
--- a/backends/cadence/fused_quant/tests/test_op_add.cpp
+++ b/backends/cadence/fused_quant/tests/test_op_add.cpp
@@ -14,10 +14,10 @@
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_factory.h>
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_util.h>
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::testing::TensorFactory;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/tests/test_op_bmm.cpp b/backends/cadence/fused_quant/tests/test_op_bmm.cpp
index 5ede47ea8a9..bae04993a7a 100644
--- a/backends/cadence/fused_quant/tests/test_op_bmm.cpp
+++ b/backends/cadence/fused_quant/tests/test_op_bmm.cpp
@@ -14,10 +14,10 @@
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_factory.h>
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_util.h>
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::testing::TensorFactory;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/tests/test_op_hardswish.cpp b/backends/cadence/fused_quant/tests/test_op_hardswish.cpp
index 502d680d2e3..eb6231161f2 100644
--- a/backends/cadence/fused_quant/tests/test_op_hardswish.cpp
+++ b/backends/cadence/fused_quant/tests/test_op_hardswish.cpp
@@ -14,10 +14,10 @@
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_factory.h>
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_util.h>
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::testing::TensorFactory;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/tests/test_op_mul.cpp b/backends/cadence/fused_quant/tests/test_op_mul.cpp
index 0b9addabc5e..da27c7287c9 100644
--- a/backends/cadence/fused_quant/tests/test_op_mul.cpp
+++ b/backends/cadence/fused_quant/tests/test_op_mul.cpp
@@ -14,10 +14,10 @@
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_factory.h>
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_util.h>
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::testing::TensorFactory;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/fused_quant/tests/test_op_relu.cpp b/backends/cadence/fused_quant/tests/test_op_relu.cpp
index 6b83551fd2b..1096daae202 100644
--- a/backends/cadence/fused_quant/tests/test_op_relu.cpp
+++ b/backends/cadence/fused_quant/tests/test_op_relu.cpp
@@ -14,10 +14,10 @@
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_factory.h>
 #include <executorch/runtime/core/exec_aten/testing_util/tensor_util.h>
 
-using executorch::aten::optional;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::testing::TensorFactory;
+using std::optional;
 
 namespace {
 
diff --git a/backends/cadence/generic/operators/op_avg_pool2d.cpp b/backends/cadence/generic/operators/op_avg_pool2d.cpp
index b04187db62e..c33f91151fb 100644
--- a/backends/cadence/generic/operators/op_avg_pool2d.cpp
+++ b/backends/cadence/generic/operators/op_avg_pool2d.cpp
@@ -19,11 +19,11 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::getLeadingDims;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 // Compute the avg_pool2d for in_data in NCHW layout. IT is the input datatype,
 // and AT is the accumulation datatype. 'quantized' is true when the input is
diff --git a/backends/cadence/generic/operators/op_avg_pool2d.h b/backends/cadence/generic/operators/op_avg_pool2d.h
index 05f1810bb61..85b5d55a84b 100644
--- a/backends/cadence/generic/operators/op_avg_pool2d.h
+++ b/backends/cadence/generic/operators/op_avg_pool2d.h
@@ -23,9 +23,8 @@ ::executorch::aten::Tensor& avg_pool2d_out(
     ::executorch::aten::IntArrayRef padding,
     bool ceil_mode,
     bool count_include_pad,
-    ::executorch::aten::optional<int64_t> divisor_override,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>&
-        in_zero_point_t,
+    std::optional<int64_t> divisor_override,
+    const std::optional<::executorch::aten::Tensor>& in_zero_point_t,
     bool channel_last,
     ::executorch::aten::Tensor& out);
 
diff --git a/backends/cadence/generic/operators/op_fully_connected.cpp b/backends/cadence/generic/operators/op_fully_connected.cpp
index 36befc52102..b65f8016880 100644
--- a/backends/cadence/generic/operators/op_fully_connected.cpp
+++ b/backends/cadence/generic/operators/op_fully_connected.cpp
@@ -15,10 +15,10 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::getLeadingDims;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 void linear(
     const Tensor& input,
diff --git a/backends/cadence/generic/operators/op_fully_connected.h b/backends/cadence/generic/operators/op_fully_connected.h
index d23bcbeb70c..7e03f5ef664 100644
--- a/backends/cadence/generic/operators/op_fully_connected.h
+++ b/backends/cadence/generic/operators/op_fully_connected.h
@@ -15,9 +15,9 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 Tensor& fully_connected_out(
     KernelRuntimeContext& ctx,
diff --git a/backends/cadence/generic/operators/op_linalg_svd.cpp b/backends/cadence/generic/operators/op_linalg_svd.cpp
index 4974b617418..4cb4f6397ea 100644
--- a/backends/cadence/generic/operators/op_linalg_svd.cpp
+++ b/backends/cadence/generic/operators/op_linalg_svd.cpp
@@ -261,7 +261,7 @@ std::tuple<Tensor&, Tensor&, Tensor&> linalg_svd_out(
     const Tensor& A,
     bool full_matrices,
     bool compute_uv,
-    ::executorch::aten::optional<::executorch::aten::string_view> driver,
+    std::optional<std::string_view> driver,
     Tensor& U,
     Tensor& S,
     Tensor& Vh) {
diff --git a/backends/cadence/generic/operators/op_linalg_svd.h b/backends/cadence/generic/operators/op_linalg_svd.h
index 7635276c4f5..e8335b7fa0e 100644
--- a/backends/cadence/generic/operators/op_linalg_svd.h
+++ b/backends/cadence/generic/operators/op_linalg_svd.h
@@ -26,7 +26,7 @@ linalg_svd_out(
     const ::executorch::aten::Tensor& A,
     bool full_matrices,
     bool compute_uv,
-    ::executorch::aten::optional<::executorch::aten::string_view> driver,
+    std::optional<std::string_view> driver,
     ::executorch::aten::Tensor& U,
     ::executorch::aten::Tensor& S,
     ::executorch::aten::Tensor& Vh);
diff --git a/backends/cadence/generic/operators/op_quantized_conv1d_nlc.cpp b/backends/cadence/generic/operators/op_quantized_conv1d_nlc.cpp
index 6f42543cfc1..8a427045a83 100644
--- a/backends/cadence/generic/operators/op_quantized_conv1d_nlc.cpp
+++ b/backends/cadence/generic/operators/op_quantized_conv1d_nlc.cpp
@@ -256,7 +256,7 @@ ::executorch::aten::Tensor& quantized_conv1d_nlc_per_tensor_out(
     int64_t output_zero_point,
     __ET_UNUSED int64_t out_multiplier,
     __ET_UNUSED int64_t out_shift,
-    __ET_UNUSED const ::executorch::aten::optional<Tensor>& offset,
+    __ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
   (void)ctx;
   quantized_conv1d_nlc(
diff --git a/backends/cadence/generic/operators/op_quantized_conv1d_nlc.h b/backends/cadence/generic/operators/op_quantized_conv1d_nlc.h
index 4f4d2877b27..f1780497f73 100644
--- a/backends/cadence/generic/operators/op_quantized_conv1d_nlc.h
+++ b/backends/cadence/generic/operators/op_quantized_conv1d_nlc.h
@@ -54,7 +54,7 @@ ::executorch::aten::Tensor& quantized_conv1d_nlc_per_tensor_out(
     int64_t output_zero_point,
     int64_t out_multiplier,
     int64_t out_shift,
-    const ::executorch::aten::optional<Tensor>& offset,
+    const std::optional<Tensor>& offset,
     Tensor& out);
 
 } // namespace native
diff --git a/backends/cadence/generic/operators/op_quantized_conv2d.cpp b/backends/cadence/generic/operators/op_quantized_conv2d.cpp
index 0811267a3b8..f6755f9dda8 100644
--- a/backends/cadence/generic/operators/op_quantized_conv2d.cpp
+++ b/backends/cadence/generic/operators/op_quantized_conv2d.cpp
@@ -16,11 +16,11 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 /* This implements a generic 2d conv kernel that operates on raw pointers.
  * The quantized version handles quantized convolutions for 2D inputs.
@@ -936,7 +936,7 @@ Tensor& quantized_conv2d_nhwc_per_tensor_out(
     int64_t output_zero_point,
     ET_UNUSED int64_t out_multiplier,
     ET_UNUSED int64_t out_shift,
-    ET_UNUSED const ::executorch::aten::optional<Tensor>& offset,
+    ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
   quantized_conv2d_nhwc(
       input,
diff --git a/backends/cadence/generic/operators/op_quantized_conv2d.h b/backends/cadence/generic/operators/op_quantized_conv2d.h
index bb9476e2644..02740d3afec 100644
--- a/backends/cadence/generic/operators/op_quantized_conv2d.h
+++ b/backends/cadence/generic/operators/op_quantized_conv2d.h
@@ -205,7 +205,7 @@ ::executorch::aten::Tensor& quantized_conv2d_nhwc_per_tensor_out(
     int64_t output_zero_point,
     int64_t out_multiplier,
     int64_t out_shift,
-    const ::executorch::aten::optional<Tensor>& offset,
+    const std::optional<Tensor>& offset,
     Tensor& out);
 
 ::executorch::aten::Tensor& quantized_conv2d_depthwise_nhwc_out(
diff --git a/backends/cadence/generic/operators/op_quantized_depthwise_conv1d_nlc.cpp b/backends/cadence/generic/operators/op_quantized_depthwise_conv1d_nlc.cpp
index a8f98a76ffc..05fb809cd51 100644
--- a/backends/cadence/generic/operators/op_quantized_depthwise_conv1d_nlc.cpp
+++ b/backends/cadence/generic/operators/op_quantized_depthwise_conv1d_nlc.cpp
@@ -57,7 +57,7 @@ ::executorch::aten::Tensor& quantized_depthwise_conv1d_nlc_per_tensor_out(
       output_zero_point,
       out_multiplier,
       out_shift,
-      ::executorch::aten::optional<Tensor>(),
+      std::optional<Tensor>(),
       out);
 }
 
diff --git a/backends/cadence/generic/operators/op_quantized_embedding_byte.cpp b/backends/cadence/generic/operators/op_quantized_embedding_byte.cpp
index 55ca67648ca..d2e0d6a8bd9 100644
--- a/backends/cadence/generic/operators/op_quantized_embedding_byte.cpp
+++ b/backends/cadence/generic/operators/op_quantized_embedding_byte.cpp
@@ -19,11 +19,11 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 #define ET_FORALL_CADENCE_QUANTIZED_TYPES(_) \
   _(uint8_t, Byte)                           \
diff --git a/backends/cadence/generic/operators/op_quantized_embedding_byte.h b/backends/cadence/generic/operators/op_quantized_embedding_byte.h
index a46bebe09df..84fc53620a0 100644
--- a/backends/cadence/generic/operators/op_quantized_embedding_byte.h
+++ b/backends/cadence/generic/operators/op_quantized_embedding_byte.h
@@ -19,8 +19,7 @@ ::executorch::aten::Tensor& quantized_embedding_byte_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& weight,
     const ::executorch::aten::Tensor& weight_scales,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>&
-        weight_zero_points,
+    const std::optional<::executorch::aten::Tensor>& weight_zero_points,
     const ::executorch::aten::Tensor& indices,
     bool pruned_weights,
     ::executorch::aten::Tensor& out);
diff --git a/backends/cadence/generic/operators/op_quantized_fully_connected.cpp b/backends/cadence/generic/operators/op_quantized_fully_connected.cpp
index 55e29cb7f52..ce74b5b8b7f 100644
--- a/backends/cadence/generic/operators/op_quantized_fully_connected.cpp
+++ b/backends/cadence/generic/operators/op_quantized_fully_connected.cpp
@@ -16,10 +16,10 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 Tensor& quantized_fully_connected_out(
     ET_UNUSED KernelRuntimeContext& ctx,
diff --git a/backends/cadence/generic/operators/op_quantized_fully_connected.h b/backends/cadence/generic/operators/op_quantized_fully_connected.h
index a7510fba95f..408fbabe726 100644
--- a/backends/cadence/generic/operators/op_quantized_fully_connected.h
+++ b/backends/cadence/generic/operators/op_quantized_fully_connected.h
@@ -25,7 +25,7 @@ ::executorch::aten::Tensor& quantized_fully_connected_out(
     const ::executorch::aten::Tensor& out_multiplier,
     const ::executorch::aten::Tensor& out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& quantized_fully_connected_per_tensor_out(
@@ -38,7 +38,7 @@ ::executorch::aten::Tensor& quantized_fully_connected_per_tensor_out(
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor&
@@ -52,7 +52,7 @@ quantized_fully_connected_asym8sxasym8s_asym8s_per_tensor_out(
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor&
@@ -66,7 +66,7 @@ quantized_fully_connected_asym8uxasym8u_asym8u_per_tensor_out(
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 } // namespace native
diff --git a/backends/cadence/generic/operators/op_quantized_layer_norm.cpp b/backends/cadence/generic/operators/op_quantized_layer_norm.cpp
index e34ed342d22..85825cff94d 100644
--- a/backends/cadence/generic/operators/op_quantized_layer_norm.cpp
+++ b/backends/cadence/generic/operators/op_quantized_layer_norm.cpp
@@ -24,7 +24,6 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
@@ -32,6 +31,7 @@ using ::executorch::runtime::getLeadingDims;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::dequantize;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 // Compute quantized layer_norm. The current implementation assumes that the
 // input is per-tensor quantized.
diff --git a/backends/cadence/generic/operators/op_quantized_linear.cpp b/backends/cadence/generic/operators/op_quantized_linear.cpp
index 87f990a855b..02ff97de74d 100644
--- a/backends/cadence/generic/operators/op_quantized_linear.cpp
+++ b/backends/cadence/generic/operators/op_quantized_linear.cpp
@@ -18,11 +18,11 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::executorch::runtime::toString;
+using std::optional;
 
 Tensor& quantized_linear_out(
     ET_UNUSED KernelRuntimeContext& ctx,
diff --git a/backends/cadence/generic/operators/op_quantized_linear.h b/backends/cadence/generic/operators/op_quantized_linear.h
index b5396cb9701..517357d5bf9 100644
--- a/backends/cadence/generic/operators/op_quantized_linear.h
+++ b/backends/cadence/generic/operators/op_quantized_linear.h
@@ -25,7 +25,7 @@ ::executorch::aten::Tensor& quantized_linear_out(
     const ::executorch::aten::Tensor& out_multiplier,
     const ::executorch::aten::Tensor& out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& quantized_linear_per_tensor_out(
@@ -38,7 +38,7 @@ ::executorch::aten::Tensor& quantized_linear_per_tensor_out(
     const int64_t out_multiplier,
     const int64_t out_shift,
     const int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor&
diff --git a/backends/cadence/generic/operators/op_quantized_matmul.cpp b/backends/cadence/generic/operators/op_quantized_matmul.cpp
index e3fb0f00fdc..b84c879e65d 100644
--- a/backends/cadence/generic/operators/op_quantized_matmul.cpp
+++ b/backends/cadence/generic/operators/op_quantized_matmul.cpp
@@ -21,12 +21,12 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 // The quantized matmul. The quantized matmul accumulates in a wider register,
 // whose type is TA.
diff --git a/backends/cadence/generic/operators/op_quantized_matmul.h b/backends/cadence/generic/operators/op_quantized_matmul.h
index 70775380aac..c28862aa11e 100644
--- a/backends/cadence/generic/operators/op_quantized_matmul.h
+++ b/backends/cadence/generic/operators/op_quantized_matmul.h
@@ -15,9 +15,9 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 Tensor& quantized_matmul_out(
     KernelRuntimeContext& ctx,
diff --git a/backends/cadence/generic/operators/op_quantized_mul.cpp b/backends/cadence/generic/operators/op_quantized_mul.cpp
index 30352ee9d52..359a305b020 100644
--- a/backends/cadence/generic/operators/op_quantized_mul.cpp
+++ b/backends/cadence/generic/operators/op_quantized_mul.cpp
@@ -21,13 +21,13 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::dequantize;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 DECLARE_POINTWISE_TENSOR_QUANTIZED_BINARY_OP(quantized_mul_, *);
 
diff --git a/backends/cadence/generic/operators/op_quantized_relu.cpp b/backends/cadence/generic/operators/op_quantized_relu.cpp
index 9430951f65b..ecb87bd1b90 100644
--- a/backends/cadence/generic/operators/op_quantized_relu.cpp
+++ b/backends/cadence/generic/operators/op_quantized_relu.cpp
@@ -21,12 +21,12 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 template <typename T>
 void quantized_relu_per_tensor_out_(
diff --git a/backends/cadence/generic/operators/op_requantize.cpp b/backends/cadence/generic/operators/op_requantize.cpp
index f846a1964a3..b9df6f1f355 100644
--- a/backends/cadence/generic/operators/op_requantize.cpp
+++ b/backends/cadence/generic/operators/op_requantize.cpp
@@ -19,13 +19,13 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::dequantize;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 // Requantize the int8_t/uint8_t input tensor to a uint8_t/int8_t out tensor.
 // The scale and zero_point for requantization are in the args.
diff --git a/backends/cadence/generic/operators/op_rope.cpp b/backends/cadence/generic/operators/op_rope.cpp
index 17ee6d2a684..fcc7d629cf7 100644
--- a/backends/cadence/generic/operators/op_rope.cpp
+++ b/backends/cadence/generic/operators/op_rope.cpp
@@ -12,8 +12,8 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Tensor;
+using std::optional;
 
 Tensor& rope_out(
     ET_UNUSED ::executorch::runtime::KernelRuntimeContext& ctx,
@@ -75,8 +75,8 @@ namespace impl {
 namespace generic {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Tensor;
+using std::optional;
 
 Tensor& rope_rotate_stacked_halves_out(
     ET_UNUSED ::executorch::runtime::KernelRuntimeContext& ctx,
diff --git a/backends/cadence/generic/operators/op_rope.h b/backends/cadence/generic/operators/op_rope.h
index 638677bf118..d738cfda6c1 100644
--- a/backends/cadence/generic/operators/op_rope.h
+++ b/backends/cadence/generic/operators/op_rope.h
@@ -20,7 +20,7 @@ ::executorch::aten::Tensor& rope_out(
     const ::executorch::aten::Tensor& input,
     const ::executorch::aten::Tensor& sin_tensor,
     const ::executorch::aten::Tensor& cos_tensor,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& pos,
+    const std::optional<::executorch::aten::Tensor>& pos,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& rope_rotate_stacked_halves_out(
@@ -28,7 +28,7 @@ ::executorch::aten::Tensor& rope_rotate_stacked_halves_out(
     const ::executorch::aten::Tensor& input,
     const ::executorch::aten::Tensor& sin_tensor,
     const ::executorch::aten::Tensor& cos_tensor,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& pos,
+    const std::optional<::executorch::aten::Tensor>& pos,
     ::executorch::aten::Tensor& out);
 
 } // namespace native
diff --git a/backends/cadence/generic/operators/op_softmax.cpp b/backends/cadence/generic/operators/op_softmax.cpp
index 97c64a22511..b680d1e2471 100644
--- a/backends/cadence/generic/operators/op_softmax.cpp
+++ b/backends/cadence/generic/operators/op_softmax.cpp
@@ -125,7 +125,7 @@ Tensor& _softmax_f32_f32_out(
     __ET_UNUSED KernelRuntimeContext& ctx,
     const Tensor& X,
     int64_t dim,
-    __ET_UNUSED ::executorch::aten::optional<bool> half_to_float,
+    __ET_UNUSED std::optional<bool> half_to_float,
     Tensor& Y) {
   _softmax_out(ctx, X, dim, false, Y);
 
diff --git a/backends/cadence/generic/operators/op_softmax.h b/backends/cadence/generic/operators/op_softmax.h
index ec51b1d00c0..d83703117b0 100644
--- a/backends/cadence/generic/operators/op_softmax.h
+++ b/backends/cadence/generic/operators/op_softmax.h
@@ -26,7 +26,7 @@ ::executorch::aten::Tensor& _softmax_f32_f32_out(
     __ET_UNUSED ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& X,
     int64_t dim,
-    __ET_UNUSED ::executorch::aten::optional<bool> half_to_float,
+    __ET_UNUSED std::optional<bool> half_to_float,
     ::executorch::aten::Tensor& Y);
 
 } // namespace native
diff --git a/backends/cadence/generic/operators/op_transposed_convolution.cpp b/backends/cadence/generic/operators/op_transposed_convolution.cpp
index 121b479e65f..b742ec635b2 100644
--- a/backends/cadence/generic/operators/op_transposed_convolution.cpp
+++ b/backends/cadence/generic/operators/op_transposed_convolution.cpp
@@ -16,12 +16,12 @@ namespace generic {
 namespace native {
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
 using ::impl::generic::kernels::quantize;
+using std::optional;
 
 // This implements a generic 2d transposed_conv kernel that operates on raw
 // pointers. The version handles both quantized and fp32 convolutions.
diff --git a/backends/cadence/hifi/operators/op_mean.cpp b/backends/cadence/hifi/operators/op_mean.cpp
index 514813fbe05..ccd54e80698 100644
--- a/backends/cadence/hifi/operators/op_mean.cpp
+++ b/backends/cadence/hifi/operators/op_mean.cpp
@@ -17,8 +17,8 @@ using executorch::aten::RuntimeContext;
 using executorch::aten::ScalarType;
 using executorch::aten::Tensor;
 using executorch::runtime::ArrayRef;
+using std::optional;
 using torch::executor::Error;
-using torch::executor::optional;
 
 namespace impl {
 namespace HiFi {
diff --git a/backends/cadence/hifi/operators/op_quantized_conv1d_nlc.cpp b/backends/cadence/hifi/operators/op_quantized_conv1d_nlc.cpp
index 5171c2908bc..9d363469f74 100644
--- a/backends/cadence/hifi/operators/op_quantized_conv1d_nlc.cpp
+++ b/backends/cadence/hifi/operators/op_quantized_conv1d_nlc.cpp
@@ -238,7 +238,7 @@ void quantized_conv1d_nlc_per_tensor_out(
     int64_t output_zero_point,
     int64_t out_multiplier,
     int64_t out_shift,
-    __ET_UNUSED const ::executorch::aten::optional<Tensor>& offset,
+    __ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
   // HiFi nnlib kernels only support dilation=1.
   // Fall back to generic implementation for dilation > 1.
diff --git a/backends/cadence/hifi/operators/op_quantized_conv2d_nhwc_out.cpp b/backends/cadence/hifi/operators/op_quantized_conv2d_nhwc_out.cpp
index ea3a756f995..86ef244711d 100644
--- a/backends/cadence/hifi/operators/op_quantized_conv2d_nhwc_out.cpp
+++ b/backends/cadence/hifi/operators/op_quantized_conv2d_nhwc_out.cpp
@@ -17,7 +17,7 @@ using Tensor = executorch::aten::Tensor;
 using KernelRuntimeContext = torch::executor::KernelRuntimeContext;
 using ScalarType = executorch::aten::ScalarType;
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
+using std::optional;
 
 namespace impl {
 namespace HiFi {
diff --git a/backends/cadence/hifi/operators/op_quantized_depthwise_conv1d_nlc.cpp b/backends/cadence/hifi/operators/op_quantized_depthwise_conv1d_nlc.cpp
index 4299990b52a..a8e2b42d77d 100644
--- a/backends/cadence/hifi/operators/op_quantized_depthwise_conv1d_nlc.cpp
+++ b/backends/cadence/hifi/operators/op_quantized_depthwise_conv1d_nlc.cpp
@@ -206,7 +206,7 @@ void quantized_depthwise_conv1d_nlc_per_tensor_out(
         output_zero_point,
         out_multiplier,
         out_shift,
-        ::executorch::aten::optional<Tensor>(),
+        std::optional<Tensor>(),
         out);
     return;
   }
diff --git a/backends/cadence/hifi/operators/op_quantized_matmul_out.h b/backends/cadence/hifi/operators/op_quantized_matmul_out.h
index c53a07b58aa..a567c7f650d 100644
--- a/backends/cadence/hifi/operators/op_quantized_matmul_out.h
+++ b/backends/cadence/hifi/operators/op_quantized_matmul_out.h
@@ -21,7 +21,7 @@ ::executorch::aten::Tensor& quantized_matmul_out(
     int64_t X_zero_point,
     const ::executorch::aten::Tensor& Y,
     int64_t Y_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& bias,
+    const std::optional<::executorch::aten::Tensor>& bias,
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
diff --git a/backends/cadence/hifi/operators/op_softmax_f32_f32.cpp b/backends/cadence/hifi/operators/op_softmax_f32_f32.cpp
index 074ff29b301..907156af1f7 100644
--- a/backends/cadence/hifi/operators/op_softmax_f32_f32.cpp
+++ b/backends/cadence/hifi/operators/op_softmax_f32_f32.cpp
@@ -22,7 +22,7 @@ inline Tensor& _softmax_f32_f32_out(
     KernelRuntimeContext& ctx,
     const Tensor& in,
     int64_t dim,
-    ::executorch::aten::optional<bool> half_to_float,
+    std::optional<bool> half_to_float,
     Tensor& out) {
   constexpr int kNnlibMaxDim = 16;
 
@@ -146,7 +146,7 @@ Tensor& softmax_f32_f32_out(
     KernelRuntimeContext& ctx,
     const Tensor& in,
     int64_t dim,
-    ::executorch::aten::optional<bool> half_to_float,
+    std::optional<bool> half_to_float,
     Tensor& out) {
   return _softmax_f32_f32_out(ctx, in, dim, half_to_float, out);
 }
diff --git a/backends/cadence/hifi/operators/operators.h b/backends/cadence/hifi/operators/operators.h
index 3ca505d40cb..fa6847f744b 100644
--- a/backends/cadence/hifi/operators/operators.h
+++ b/backends/cadence/hifi/operators/operators.h
@@ -72,7 +72,7 @@ void quantized_linear_out(
     const ::executorch::aten::Tensor& out_multiplier,
     const ::executorch::aten::Tensor& out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 void quantized_linear_per_tensor_out(
@@ -85,7 +85,7 @@ void quantized_linear_per_tensor_out(
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 void quantized_conv2d_nhwc_out(
@@ -158,7 +158,7 @@ void quantized_conv2d_nhwc_per_tensor_out(
     int64_t output_zero_point,
     int64_t out_multiplier,
     int64_t out_shift,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& offset,
+    const std::optional<::executorch::aten::Tensor>& offset,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& cat_out(
diff --git a/backends/cadence/vision/operators/op_quantized_conv_out.cpp b/backends/cadence/vision/operators/op_quantized_conv_out.cpp
index be4b34bff03..aaba9f5696d 100644
--- a/backends/cadence/vision/operators/op_quantized_conv_out.cpp
+++ b/backends/cadence/vision/operators/op_quantized_conv_out.cpp
@@ -582,7 +582,7 @@ void quantized_conv2d_nhwc_per_tensor_out(
     int64_t output_zero_point,
     int64_t out_multiplier,
     int64_t out_shift,
-    ET_UNUSED const ::executorch::aten::optional<Tensor>& offset,
+    ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
   quantized_conv_per_tensor_out(
       ctx,
diff --git a/backends/cadence/vision/operators/op_quantized_fully_connected_out.cpp b/backends/cadence/vision/operators/op_quantized_fully_connected_out.cpp
index 29aa8906414..c53f7f7667a 100644
--- a/backends/cadence/vision/operators/op_quantized_fully_connected_out.cpp
+++ b/backends/cadence/vision/operators/op_quantized_fully_connected_out.cpp
@@ -13,10 +13,10 @@ namespace impl {
 namespace vision {
 namespace native {
 
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 void quantized_fully_connected_out(
     __ET_UNUSED KernelRuntimeContext& ctx,
diff --git a/backends/cadence/vision/operators/op_quantized_linear_out.cpp b/backends/cadence/vision/operators/op_quantized_linear_out.cpp
index b6b7cdd17bc..7b3daed8ef6 100644
--- a/backends/cadence/vision/operators/op_quantized_linear_out.cpp
+++ b/backends/cadence/vision/operators/op_quantized_linear_out.cpp
@@ -84,7 +84,7 @@ void quantized_linear_out(
     const Tensor& out_multiplier,
     const Tensor& out_shift,
     int64_t out_zero_point,
-    __ET_UNUSED const executorch::aten::optional<Tensor>& offset,
+    __ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
   // TODO: refactor to use switch case as quantized_linear_per_tensor_out
   if (out.scalar_type() == executorch::aten::ScalarType::Byte) {
@@ -127,7 +127,7 @@ void quantized_linear_per_tensor_out(
     const int64_t out_multiplier,
     const int64_t out_shift,
     const int64_t out_zero_point,
-    __ET_UNUSED const executorch::aten::optional<Tensor>& offset,
+    __ET_UNUSED const std::optional<Tensor>& offset,
     Tensor& out) {
 #define typed_quantized_linear_per_tensor(ctype, dtype) \
   case executorch::aten::ScalarType::dtype: {           \
diff --git a/backends/cadence/vision/operators/op_quantized_matmul_out.cpp b/backends/cadence/vision/operators/op_quantized_matmul_out.cpp
index 54a303288c3..e63ae5bdda1 100644
--- a/backends/cadence/vision/operators/op_quantized_matmul_out.cpp
+++ b/backends/cadence/vision/operators/op_quantized_matmul_out.cpp
@@ -60,7 +60,7 @@ void inline _typed_quantized_matmul(
     int64_t X_zero_point,
     const Tensor& Y,
     int64_t Y_zero_point,
-    const executorch::aten::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
@@ -114,7 +114,7 @@ void quantized_matmul_out(
     int64_t X_zero_point,
     const Tensor& Y,
     int64_t Y_zero_point,
-    const executorch::aten::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     int64_t out_multiplier,
     int64_t out_shift,
     int64_t out_zero_point,
diff --git a/backends/cadence/vision/operators/op_softmax.cpp b/backends/cadence/vision/operators/op_softmax.cpp
index 58ca33c6a0b..6b93709b226 100644
--- a/backends/cadence/vision/operators/op_softmax.cpp
+++ b/backends/cadence/vision/operators/op_softmax.cpp
@@ -50,7 +50,7 @@ Tensor& _softmax_out(
   // Adjust for negative dim
   dim = dim < 0 ? dim + executorch::runtime::nonzero_dim(in) : dim;
 
-  const executorch::aten::optional<int64_t>& dim_t = dim;
+  const std::optional<int64_t>& dim_t = dim;
   const size_t d = ET_NORMALIZE_IX(dim_t.value(), in.dim());
   const size_t size = in.size(d);
 
diff --git a/backends/cadence/vision/operators/operators.h b/backends/cadence/vision/operators/operators.h
index 8b5db4161eb..1c756c0b237 100644
--- a/backends/cadence/vision/operators/operators.h
+++ b/backends/cadence/vision/operators/operators.h
@@ -31,7 +31,7 @@ using ::executorch::runtime::getLeadingDims;
 inline __attribute__((always_inline)) void linear_(
     const ::executorch::aten::Tensor& input,
     const ::executorch::aten::Tensor& weight,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& bias,
+    const std::optional<::executorch::aten::Tensor>& bias,
     ::executorch::aten::Tensor& output) {
   const float* __restrict__ input_data = input.const_data_ptr<float>();
   const float* __restrict__ weight_data = weight.const_data_ptr<float>();
diff --git a/backends/cortex_m/ops/cmsis_scratch_buffer_context.h b/backends/cortex_m/ops/cmsis_scratch_buffer_context.h
index 656309abcee..807cf18cebc 100644
--- a/backends/cortex_m/ops/cmsis_scratch_buffer_context.h
+++ b/backends/cortex_m/ops/cmsis_scratch_buffer_context.h
@@ -49,7 +49,7 @@ class CMSISScratchBufferContext final {
       Tensor& scratch_buffer,
       const Tensor& weights,
       const Tensor& weight_zero_point,
-      const torch::executor::optional<Tensor>& bias)
+      const std::optional<Tensor>& bias)
       : scratch_ptr_(scratch_buffer.mutable_data_ptr<int8_t>()),
         total_size_(scratch_buffer.size(0)),
         base_ptr_(reinterpret_cast<uint8_t*>(scratch_ptr_)),
diff --git a/backends/cortex_m/ops/op_quantized_conv2d.cpp b/backends/cortex_m/ops/op_quantized_conv2d.cpp
index 3d4f19e10d0..13e8b132410 100644
--- a/backends/cortex_m/ops/op_quantized_conv2d.cpp
+++ b/backends/cortex_m/ops/op_quantized_conv2d.cpp
@@ -19,7 +19,7 @@ bool validate_conv2d_arguments(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Tensor& output,
     const Int64ArrayRef& stride,
     const Int64ArrayRef& padding,
@@ -103,7 +103,7 @@ Tensor& quantized_conv2d_out(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Int64ArrayRef stride,
     const Int64ArrayRef padding,
     const Int64ArrayRef dilation,
diff --git a/backends/cortex_m/ops/op_quantized_depthwise_conv2d.cpp b/backends/cortex_m/ops/op_quantized_depthwise_conv2d.cpp
index a8e1fc21ed7..0793606de44 100644
--- a/backends/cortex_m/ops/op_quantized_depthwise_conv2d.cpp
+++ b/backends/cortex_m/ops/op_quantized_depthwise_conv2d.cpp
@@ -19,7 +19,7 @@ bool validate_depthwise_conv2d_arguments(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Tensor& output,
     const Int64ArrayRef& stride,
     const Int64ArrayRef& padding,
@@ -140,7 +140,7 @@ Tensor& quantized_depthwise_conv2d_out(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Int64ArrayRef stride,
     const Int64ArrayRef padding,
     const Int64ArrayRef dilation,
diff --git a/backends/cortex_m/ops/op_quantized_linear.cpp b/backends/cortex_m/ops/op_quantized_linear.cpp
index 7448058de8e..c92ec493cd5 100644
--- a/backends/cortex_m/ops/op_quantized_linear.cpp
+++ b/backends/cortex_m/ops/op_quantized_linear.cpp
@@ -18,8 +18,8 @@ Tensor& quantized_linear_out(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weights,
-    const torch::executor::optional<Tensor>& bias,
-    const torch::executor::optional<Tensor>& kernel_sum,
+    const std::optional<Tensor>& bias,
+    const std::optional<Tensor>& kernel_sum,
     const int64_t input_offset,
     const int64_t filter_offset,
     const int64_t output_offset,
diff --git a/backends/cortex_m/ops/op_quantized_transpose_conv2d.cpp b/backends/cortex_m/ops/op_quantized_transpose_conv2d.cpp
index e7ecbc7c7b4..04d57d4c693 100644
--- a/backends/cortex_m/ops/op_quantized_transpose_conv2d.cpp
+++ b/backends/cortex_m/ops/op_quantized_transpose_conv2d.cpp
@@ -21,7 +21,7 @@ bool validate_transpose_conv2d_arguments(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Tensor& output,
     const Tensor& requantize_multipliers,
     const Tensor& requantize_shifts) {
@@ -88,7 +88,7 @@ Tensor& quantized_transpose_conv2d_out(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& weight,
-    const torch::executor::optional<Tensor>& bias,
+    const std::optional<Tensor>& bias,
     const Int64ArrayRef stride,
     const Int64ArrayRef padding,
     const Int64ArrayRef output_padding,
diff --git a/codegen/api/et_cpp.py b/codegen/api/et_cpp.py
index 88f1eb83fe0..1a823b98a5b 100644
--- a/codegen/api/et_cpp.py
+++ b/codegen/api/et_cpp.py
@@ -278,7 +278,7 @@ def default_expr(d: str, t: Type) -> str:
 
     if isinstance(t, OptionalType):
         if d == "None":
-            return "torch::executor::nullopt"
+            return "std::nullopt"
 
         return default_expr(d, t.elem)
 
diff --git a/codegen/api/types/types.py b/codegen/api/types/types.py
index 712d7e5e341..07c7f7e196a 100644
--- a/codegen/api/types/types.py
+++ b/codegen/api/types/types.py
@@ -59,7 +59,7 @@ class OptionalCType(CType):
 
     def cpp_type(self, *, strip_ref: bool = False) -> str:
         # Do not pass `strip_ref` recursively.
-        return f"torch::executor::optional<{self.elem.cpp_type()}>"
+        return f"std::optional<{self.elem.cpp_type()}>"
 
     def remove_const_ref(self) -> CType:
         return OptionalCType(self.elem.remove_const_ref())
diff --git a/extension/aten_util/make_aten_functor_from_et_functor.h b/extension/aten_util/make_aten_functor_from_et_functor.h
index 8e1c2bf0143..c7a823c26df 100644
--- a/extension/aten_util/make_aten_functor_from_et_functor.h
+++ b/extension/aten_util/make_aten_functor_from_et_functor.h
@@ -67,12 +67,12 @@ struct type_map<torch::executor::Tensor> final {
 
 // Optional.
 template <class T>
-struct type_map<torch::executor::optional<T>> final {
+struct type_map<std::optional<T>> final {
   using type = std::optional<typename type_map<T>::type>;
 };
 
 template <class T>
-struct type_map<torch::executor::optional<T>&> final {
+struct type_map<std::optional<T>&> final {
   using type = std::optional<typename type_map<T>::type>&;
 };
 
@@ -177,7 +177,7 @@ struct type_convert<
                 typename remove_const_ref<AOptional>::type::value_type>> &&
         std::is_same_v<
             typename remove_const_ref<EOptional>::type,
-            torch::executor::optional<
+            std::optional<
                 typename remove_const_ref<EOptional>::type::value_type>>>>
     final {
  public:
diff --git a/extension/aten_util/test/make_aten_functor_from_et_functor_test.cpp b/extension/aten_util/test/make_aten_functor_from_et_functor_test.cpp
index b76596b9963..3abc84897ce 100644
--- a/extension/aten_util/test/make_aten_functor_from_et_functor_test.cpp
+++ b/extension/aten_util/test/make_aten_functor_from_et_functor_test.cpp
@@ -32,8 +32,8 @@ Tensor& add_1_out(const Tensor& a, Tensor& out) {
 }
 
 Tensor& add_optional_scalar_out(
-    torch::executor::optional<int64_t> s1,
-    torch::executor::optional<int64_t> s2,
+    std::optional<int64_t> s1,
+    std::optional<int64_t> s2,
     Tensor& out) {
   if (s1.has_value()) {
     out.mutable_data_ptr<int64_t>()[0] += s1.value();
@@ -45,8 +45,8 @@ Tensor& add_optional_scalar_out(
 }
 
 Tensor& add_optional_tensor_out(
-    torch::executor::optional<torch::executor::Tensor> s1,
-    torch::executor::optional<torch::executor::Tensor> s2,
+    std::optional<torch::executor::Tensor> s1,
+    std::optional<torch::executor::Tensor> s2,
     Tensor& out) {
   if (s1.has_value()) {
     out.mutable_data_ptr<int64_t>()[0] +=
@@ -78,8 +78,7 @@ Tensor& sum_arrayref_tensor_out(
 }
 
 Tensor& sum_arrayref_optional_tensor_out(
-    torch::executor::ArrayRef<
-        torch::executor::optional<torch::executor::Tensor>> a,
+    torch::executor::ArrayRef<std::optional<torch::executor::Tensor>> a,
     Tensor& out) {
   for (int i = 0; i < a.size(); i++) {
     if (a[i].has_value()) {
@@ -169,20 +168,19 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestTypeMap_Tuple_TensorRef3x) {
 TEST_F(MakeATenFunctorFromETFunctorTest, TestTypeMap_Optionals) {
   // Scalar.
   EXPECT_TRUE((std::is_same<
-               type_map<torch::executor::optional<int64_t>>::type,
+               type_map<std::optional<int64_t>>::type,
                std::optional<int64_t>>::value));
   // Tensor.
+  EXPECT_TRUE((std::is_same<
+               type_map<std::optional<torch::executor::Tensor>>::type,
+               std::optional<at::Tensor>>::value));
+  // ArrayRef.
   EXPECT_TRUE(
       (std::is_same<
-          type_map<torch::executor::optional<torch::executor::Tensor>>::type,
-          std::optional<at::Tensor>>::value));
-  // ArrayRef.
-  EXPECT_TRUE((std::is_same<
-               type_map<torch::executor::optional<
-                   torch::executor::ArrayRef<int64_t>>>::type,
-               std::optional<c10::ArrayRef<int64_t>>>::value));
+          type_map<std::optional<torch::executor::ArrayRef<int64_t>>>::type,
+          std::optional<c10::ArrayRef<int64_t>>>::value));
   EXPECT_TRUE((std::is_same<
-               type_map<torch::executor::optional<
+               type_map<std::optional<
                    torch::executor::ArrayRef<torch::executor::Tensor>>>::type,
                std::optional<c10::ArrayRef<at::Tensor>>>::value));
 }
@@ -198,13 +196,13 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestTypeMap_ArrayRef) {
           type_map<torch::executor::ArrayRef<torch::executor::Tensor>>::type,
           c10::ArrayRef<at::Tensor>>::value));
   // Optionals.
+  EXPECT_TRUE(
+      (std::is_same<
+          type_map<torch::executor::ArrayRef<std::optional<int64_t>>>::type,
+          c10::ArrayRef<std::optional<int64_t>>>::value));
   EXPECT_TRUE((std::is_same<
                type_map<torch::executor::ArrayRef<
-                   torch::executor::optional<int64_t>>>::type,
-               c10::ArrayRef<std::optional<int64_t>>>::value));
-  EXPECT_TRUE((std::is_same<
-               type_map<torch::executor::ArrayRef<
-                   torch::executor::optional<torch::executor::Tensor>>>::type,
+                   std::optional<torch::executor::Tensor>>>::type,
                c10::ArrayRef<std::optional<at::Tensor>>>::value));
 }
 
@@ -253,17 +251,16 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_OptionalScalar) {
   // Convert optional at to et.
   auto optional_at_in = std::optional<int64_t>();
   auto optional_et =
-      type_convert<std::optional<int64_t>, torch::executor::optional<int64_t>>(
+      type_convert<std::optional<int64_t>, std::optional<int64_t>>(
           optional_at_in)
           .call();
   EXPECT_TRUE(
-      (std::is_same<decltype(optional_et), torch::executor::optional<int64_t>>::
-           value));
+      (std::is_same<decltype(optional_et), std::optional<int64_t>>::value));
 
   // Convert optional et to at.
-  auto optional_et_in = torch::executor::optional<int64_t>();
+  auto optional_et_in = std::optional<int64_t>();
   auto optional_at_out =
-      type_convert<torch::executor::optional<int64_t>, std::optional<int64_t>>(
+      type_convert<std::optional<int64_t>, std::optional<int64_t>>(
           optional_et_in)
           .call();
   EXPECT_TRUE(
@@ -273,20 +270,19 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_OptionalScalar) {
 TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_OptionalTensor) {
   // Convert optional at to et.
   auto optional_at_in = std::optional<at::Tensor>();
-  auto optional_et =
-      type_convert<
-          std::optional<at::Tensor>,
-          torch::executor::optional<torch::executor::Tensor>>(optional_at_in)
-          .call();
+  auto optional_et = type_convert<
+                         std::optional<at::Tensor>,
+                         std::optional<torch::executor::Tensor>>(optional_at_in)
+                         .call();
   EXPECT_TRUE((std::is_same<
                decltype(optional_et),
-               torch::executor::optional<torch::executor::Tensor>>::value));
+               std::optional<torch::executor::Tensor>>::value));
 
   // Convert optional et to at.
   torch::executor::testing::TensorFactory<ScalarType::Int> tf;
-  auto et_in = torch::executor::optional<torch::executor::Tensor>(tf.ones({3}));
+  auto et_in = std::optional<torch::executor::Tensor>(tf.ones({3}));
   auto optional_at_out = type_convert<
-                             torch::executor::optional<torch::executor::Tensor>,
+                             std::optional<torch::executor::Tensor>,
                              std::optional<at::Tensor>>(optional_et)
                              .call();
   EXPECT_TRUE(
@@ -519,9 +515,8 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
   const std::optional<int64_t> const_optional_at_in =
       std::optional<int64_t>(42);
   auto const_optional_et =
-      type_convert<
-          const std::optional<int64_t>,
-          torch::executor::optional<int64_t>>(const_optional_at_in)
+      type_convert<const std::optional<int64_t>, std::optional<int64_t>>(
+          const_optional_at_in)
           .call();
   EXPECT_TRUE(const_optional_et.has_value());
   EXPECT_EQ(const_optional_et.value(), 42);
@@ -529,7 +524,7 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
   // Test optional scalar reference conversion
   std::optional<int64_t> optional_at_ref_in = std::optional<int64_t>(24);
   auto optional_et_from_ref =
-      type_convert<std::optional<int64_t>&, torch::executor::optional<int64_t>>(
+      type_convert<std::optional<int64_t>&, std::optional<int64_t>>(
           optional_at_ref_in)
           .call();
   EXPECT_TRUE(optional_et_from_ref.has_value());
@@ -539,9 +534,8 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
   const std::optional<int64_t> const_optional_at_ref_in =
       std::optional<int64_t>(84);
   auto const_optional_et_from_ref =
-      type_convert<
-          const std::optional<int64_t>&,
-          torch::executor::optional<int64_t>>(const_optional_at_ref_in)
+      type_convert<const std::optional<int64_t>&, std::optional<int64_t>>(
+          const_optional_at_ref_in)
           .call();
   EXPECT_TRUE(const_optional_et_from_ref.has_value());
   EXPECT_EQ(const_optional_et_from_ref.value(), 84);
@@ -551,8 +545,7 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
       std::optional<at::Tensor>(torch::tensor({5}));
   auto const_optional_tensor_converter = type_convert<
       const std::optional<at::Tensor>,
-      torch::executor::optional<torch::executor::Tensor>>(
-      const_optional_tensor_at_in);
+      std::optional<torch::executor::Tensor>>(const_optional_tensor_at_in);
   auto const_optional_tensor_et = const_optional_tensor_converter.call();
   EXPECT_TRUE(const_optional_tensor_et.has_value());
   EXPECT_EQ(const_optional_tensor_et.value().const_data_ptr<int64_t>()[0], 5);
@@ -562,8 +555,7 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
       std::optional<at::Tensor>(torch::tensor({7}));
   auto optional_tensor_converter_from_ref = type_convert<
       std::optional<at::Tensor>&,
-      torch::executor::optional<torch::executor::Tensor>>(
-      optional_tensor_at_ref_in);
+      std::optional<torch::executor::Tensor>>(optional_tensor_at_ref_in);
   auto optional_tensor_et_from_ref = optional_tensor_converter_from_ref.call();
   EXPECT_TRUE(optional_tensor_et_from_ref.has_value());
   EXPECT_EQ(
@@ -574,8 +566,7 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
       std::optional<at::Tensor>(torch::tensor({9}));
   auto const_optional_tensor_converter_from_ref = type_convert<
       const std::optional<at::Tensor>&,
-      torch::executor::optional<torch::executor::Tensor>>(
-      const_optional_tensor_at_ref_in);
+      std::optional<torch::executor::Tensor>>(const_optional_tensor_at_ref_in);
   auto const_optional_tensor_et_from_ref =
       const_optional_tensor_converter_from_ref.call();
   EXPECT_TRUE(const_optional_tensor_et_from_ref.has_value());
@@ -586,9 +577,8 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
   // Test empty const optional conversions
   const std::optional<int64_t> empty_const_optional_at_in = std::nullopt;
   auto empty_const_optional_et =
-      type_convert<
-          const std::optional<int64_t>,
-          torch::executor::optional<int64_t>>(empty_const_optional_at_in)
+      type_convert<const std::optional<int64_t>, std::optional<int64_t>>(
+          empty_const_optional_at_in)
           .call();
   EXPECT_FALSE(empty_const_optional_et.has_value());
 
@@ -597,7 +587,7 @@ TEST_F(MakeATenFunctorFromETFunctorTest, TestConvert_ConstRefOptionals) {
   auto empty_const_optional_tensor_et =
       type_convert<
           const std::optional<at::Tensor>,
-          torch::executor::optional<torch::executor::Tensor>>(
+          std::optional<torch::executor::Tensor>>(
           empty_const_optional_tensor_at_in)
           .call();
   EXPECT_FALSE(empty_const_optional_tensor_et.has_value());
diff --git a/extension/flat_tensor/flat_tensor_data_map.cpp b/extension/flat_tensor/flat_tensor_data_map.cpp
index 845778f45c2..342e29e63fc 100644
--- a/extension/flat_tensor/flat_tensor_data_map.cpp
+++ b/extension/flat_tensor/flat_tensor_data_map.cpp
@@ -49,7 +49,7 @@ bool is_aligned(const void* data) {
 }
 
 Result<const flat_tensor_flatbuffer::NamedData*> get_named_data(
-    executorch::aten::string_view key,
+    std::string_view key,
     const flatbuffers::Vector<
         flatbuffers::Offset<flat_tensor_flatbuffer::NamedData>>* named_data,
     const flatbuffers::Vector<
@@ -127,7 +127,7 @@ Result<const TensorLayout> create_tensor_layout(
 } // namespace
 
 ET_NODISCARD Result<const TensorLayout> FlatTensorDataMap::get_tensor_layout(
-    executorch::aten::string_view key) const {
+    std::string_view key) const {
   Result<uint64_t> segment_end_offset = get_segment_end_offset(header_);
   if (!segment_end_offset.ok()) {
     return segment_end_offset.error();
@@ -144,7 +144,7 @@ ET_NODISCARD Result<const TensorLayout> FlatTensorDataMap::get_tensor_layout(
 }
 
 ET_NODISCARD Result<FreeableBuffer> FlatTensorDataMap::get_data(
-    executorch::aten::string_view key) const {
+    std::string_view key) const {
   Result<uint64_t> segment_end_offset = get_segment_end_offset(header_);
   if (!segment_end_offset.ok()) {
     return segment_end_offset.error();
@@ -170,7 +170,7 @@ ET_NODISCARD Result<FreeableBuffer> FlatTensorDataMap::get_data(
 }
 
 ET_NODISCARD Error FlatTensorDataMap::load_data_into(
-    ET_UNUSED executorch::aten::string_view key,
+    ET_UNUSED std::string_view key,
     ET_UNUSED void* buffer,
     ET_UNUSED size_t size) const {
   Result<uint64_t> segment_end_offset = get_segment_end_offset(header_);
diff --git a/extension/flat_tensor/flat_tensor_data_map.h b/extension/flat_tensor/flat_tensor_data_map.h
index 751e312f7ef..7b66eeab470 100644
--- a/extension/flat_tensor/flat_tensor_data_map.h
+++ b/extension/flat_tensor/flat_tensor_data_map.h
@@ -54,7 +54,7 @@ class FlatTensorDataMap final
   ET_NODISCARD
   executorch::runtime::Result<
       const executorch::ET_RUNTIME_NAMESPACE::TensorLayout>
-  get_tensor_layout(executorch::aten::string_view key) const override;
+  get_tensor_layout(std::string_view key) const override;
 
   /**
    * Retrieve read-only data for the specified key.
@@ -65,7 +65,7 @@ class FlatTensorDataMap final
    */
   ET_NODISCARD
   executorch::runtime::Result<executorch::runtime::FreeableBuffer> get_data(
-      executorch::aten::string_view key) const override;
+      std::string_view key) const override;
 
   /**
    * Loads the data of the specified tensor into the provided buffer.
@@ -78,7 +78,7 @@ class FlatTensorDataMap final
    * @returns an Error indicating if the load was successful.
    */
   ET_NODISCARD executorch::runtime::Error load_data_into(
-      executorch::aten::string_view key,
+      std::string_view key,
       void* buffer,
       size_t size) const override;
 
diff --git a/extension/named_data_map/merged_data_map.cpp b/extension/named_data_map/merged_data_map.cpp
index 630395e006c..d76f741fbf4 100644
--- a/extension/named_data_map/merged_data_map.cpp
+++ b/extension/named_data_map/merged_data_map.cpp
@@ -12,13 +12,13 @@
 #include <unordered_map>
 #include <vector>
 
-using executorch::aten::string_view;
 using executorch::ET_RUNTIME_NAMESPACE::NamedDataMap;
 using executorch::ET_RUNTIME_NAMESPACE::TensorLayout;
 using executorch::runtime::Error;
 using executorch::runtime::FreeableBuffer;
 using executorch::runtime::Result;
 using executorch::runtime::Span;
+using std::string_view;
 
 namespace executorch::extension {
 namespace ET_MERGED_DATA_MAP_NAMESPACE {
diff --git a/extension/named_data_map/merged_data_map.h b/extension/named_data_map/merged_data_map.h
index 42490ec3d58..cc291b4d093 100644
--- a/extension/named_data_map/merged_data_map.h
+++ b/extension/named_data_map/merged_data_map.h
@@ -48,7 +48,7 @@ class MergedDataMap final
   ET_NODISCARD
   executorch::runtime::Result<
       const executorch::ET_RUNTIME_NAMESPACE::TensorLayout>
-  get_tensor_layout(executorch::aten::string_view key) const override;
+  get_tensor_layout(std::string_view key) const override;
 
   /**
    * Retrieve read-only data for the specified key.
@@ -59,7 +59,7 @@ class MergedDataMap final
    */
   ET_NODISCARD
   executorch::runtime::Result<executorch::runtime::FreeableBuffer> get_data(
-      executorch::aten::string_view key) const override;
+      std::string_view key) const override;
 
   /**
    * Loads the data of the specified tensor into the provided buffer.
@@ -72,7 +72,7 @@ class MergedDataMap final
    * @returns an Error indicating if the load was successful.
    */
   ET_NODISCARD executorch::runtime::Error load_data_into(
-      executorch::aten::string_view key,
+      std::string_view key,
       void* buffer,
       size_t size) const override;
 
diff --git a/kernels/portable/cpu/op_index_put.cpp b/kernels/portable/cpu/op_index_put.cpp
index 812d3e8fab3..519842db598 100644
--- a/kernels/portable/cpu/op_index_put.cpp
+++ b/kernels/portable/cpu/op_index_put.cpp
@@ -19,8 +19,7 @@ namespace executor {
 namespace native {
 
 using Tensor = executorch::aten::Tensor;
-using TensorOptList =
-    executorch::aten::ArrayRef<executorch::aten::optional<Tensor>>;
+using TensorOptList = executorch::aten::ArrayRef<std::optional<Tensor>>;
 
 Tensor& index_put_out(
     KernelRuntimeContext& ctx,
diff --git a/kernels/portable/cpu/op_native_dropout.cpp b/kernels/portable/cpu/op_native_dropout.cpp
index fae7928568d..dc72fb54599 100644
--- a/kernels/portable/cpu/op_native_dropout.cpp
+++ b/kernels/portable/cpu/op_native_dropout.cpp
@@ -17,7 +17,7 @@ std::tuple<Tensor&, Tensor&> native_dropout_out(
     KernelRuntimeContext& ctx,
     const Tensor& input,
     double prob,
-    torch::executor::optional<bool> train,
+    std::optional<bool> train,
     Tensor& out,
     Tensor& mask) {
   std::tuple<Tensor&, Tensor&> ret(out, mask);
diff --git a/kernels/test/op_native_dropout_test.cpp b/kernels/test/op_native_dropout_test.cpp
index 931205f54a5..fec00c87862 100644
--- a/kernels/test/op_native_dropout_test.cpp
+++ b/kernels/test/op_native_dropout_test.cpp
@@ -25,7 +25,7 @@ class OpNativeDropoutTest : public OperatorTest {
   void op_native_dropout_out(
       const Tensor& self,
       double prob,
-      executorch::aten::optional<bool> train,
+      std::optional<bool> train,
       Tensor& out,
       Tensor& mask) {
     torch::executor::aten::native_dropout_outf(
diff --git a/runtime/core/exec_aten/exec_aten.h b/runtime/core/exec_aten/exec_aten.h
index f539414aec9..bdac41ba509 100644
--- a/runtime/core/exec_aten/exec_aten.h
+++ b/runtime/core/exec_aten/exec_aten.h
@@ -8,6 +8,9 @@
 
 #pragma once
 
+#include <optional>
+#include <string_view>
+
 #include <executorch/runtime/core/error.h> // @manual
 #include <executorch/runtime/core/result.h> // @manual
 #include <executorch/runtime/core/tensor_shape_dynamism.h> // @manual
@@ -41,11 +44,9 @@
 #include <executorch/runtime/core/portable_type/complex.h> // @manual
 #include <executorch/runtime/core/portable_type/device.h> // @manual
 #include <executorch/runtime/core/portable_type/half.h> // @manual
-#include <executorch/runtime/core/portable_type/optional.h> // @manual
 #include <executorch/runtime/core/portable_type/qint_types.h> // @manual
 #include <executorch/runtime/core/portable_type/scalar.h> // @manual
 #include <executorch/runtime/core/portable_type/scalar_type.h> // @manual
-#include <executorch/runtime/core/portable_type/string_view.h> // @manual
 #include <executorch/runtime/core/portable_type/tensor.h> // @manual
 #include <executorch/runtime/core/portable_type/tensor_options.h> // @manual
 
@@ -183,8 +184,7 @@ using quint2x4 = torch::executor::quint2x4;
 using IntArrayRef = torch::executor::IntArrayRef;
 
 template <typename T>
-using OptionalArrayRef =
-    torch::executor::optional<torch::executor::ArrayRef<T>>;
+using OptionalArrayRef = std::optional<torch::executor::ArrayRef<T>>;
 using OptionalIntArrayRef = OptionalArrayRef<int64_t>;
 
 using torch::executor::compute_numel;
diff --git a/runtime/core/exec_aten/util/scalar_type_util.h b/runtime/core/exec_aten/util/scalar_type_util.h
index f48b50a0786..10b6ed2c6ac 100644
--- a/runtime/core/exec_aten/util/scalar_type_util.h
+++ b/runtime/core/exec_aten/util/scalar_type_util.h
@@ -47,11 +47,11 @@ using ScalarType = at::ScalarType;
 } // namespace executorch
 #else // !USE_ATEN_LIB
 #include <executorch/runtime/core/portable_type/scalar_type.h>
-#include <executorch/runtime/core/portable_type/string_view.h>
+#include <string_view>
 namespace executorch {
 namespace aten {
 using ScalarType = torch::executor::ScalarType;
-using string_view = torch::executor::string_view;
+using string_view = std::string_view;
 } // namespace aten
 } // namespace executorch
 #endif // USE_ATEN_LIB
diff --git a/runtime/core/named_data_map.h b/runtime/core/named_data_map.h
index c6b6aa4bb7b..dbd5b21a66f 100644
--- a/runtime/core/named_data_map.h
+++ b/runtime/core/named_data_map.h
@@ -31,7 +31,7 @@ class NamedDataMap {
    * @return Result containing TensorLayout.
    */
   ET_NODISCARD virtual Result<const TensorLayout> get_tensor_layout(
-      executorch::aten::string_view key) const = 0;
+      std::string_view key) const = 0;
   /**
    * Get data by key.
    *
@@ -39,7 +39,7 @@ class NamedDataMap {
    * @return Result containing a FreeableBuffer.
    */
   ET_NODISCARD virtual Result<FreeableBuffer> get_data(
-      executorch::aten::string_view key) const = 0;
+      std::string_view key) const = 0;
 
   /**
    * Loads data corresponding to the key into the provided buffer.
@@ -51,10 +51,8 @@ class NamedDataMap {
    * `size` bytes of memory.
    * @returns an Error indicating if the load was successful.
    */
-  ET_NODISCARD virtual Error load_data_into(
-      executorch::aten::string_view key,
-      void* buffer,
-      size_t size) const = 0;
+  ET_NODISCARD virtual Error
+  load_data_into(std::string_view key, void* buffer, size_t size) const = 0;
 
   /**
    * Get the number of keys in the NamedDataMap.
diff --git a/runtime/core/portable_type/optional.h b/runtime/core/portable_type/optional.h
deleted file mode 100644
index 31ad06fd093..00000000000
--- a/runtime/core/portable_type/optional.h
+++ /dev/null
@@ -1,36 +0,0 @@
-/*
- * Copyright (c) Meta Platforms, Inc. and affiliates.
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree.
- */
-
-#pragma once
-
-#include <optional>
-
-namespace executorch {
-namespace runtime {
-namespace etensor {
-
-// NOLINTNEXTLINE(misc-unused-using-decls)
-using std::nullopt;
-// NOLINTNEXTLINE(misc-unused-using-decls)
-using std::nullopt_t;
-// NOLINTNEXTLINE(misc-unused-using-decls)
-using std::optional;
-
-} // namespace etensor
-} // namespace runtime
-} // namespace executorch
-
-namespace torch {
-namespace executor {
-// TODO(T197294990): Remove these deprecated aliases once all users have moved
-// to the new `::executorch` namespaces.
-using ::executorch::runtime::etensor::nullopt;
-using ::executorch::runtime::etensor::nullopt_t;
-using ::executorch::runtime::etensor::optional;
-} // namespace executor
-} // namespace torch
diff --git a/runtime/core/portable_type/string_view.h b/runtime/core/portable_type/string_view.h
deleted file mode 100644
index 8e28fa022cc..00000000000
--- a/runtime/core/portable_type/string_view.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*
- * Copyright (c) Meta Platforms, Inc. and affiliates.
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree.
- */
-
-#pragma once
-
-#include <string_view>
-
-namespace executorch {
-namespace runtime {
-namespace etensor {
-
-using std::string_view;
-
-} // namespace etensor
-} // namespace runtime
-} // namespace executorch
-
-namespace torch {
-namespace executor {
-// TODO(T197294990): Remove these deprecated aliases once all users have moved
-// to the new `::executorch` namespaces.
-using ::executorch::runtime::etensor::string_view;
-} // namespace executor
-} // namespace torch
diff --git a/runtime/core/portable_type/targets.bzl b/runtime/core/portable_type/targets.bzl
index 0a368bd6379..bfa2ecd2a30 100644
--- a/runtime/core/portable_type/targets.bzl
+++ b/runtime/core/portable_type/targets.bzl
@@ -14,11 +14,9 @@ def define_common_targets():
         srcs = ["tensor_impl.cpp"],
         exported_headers = [
             "tensor_options.h",
-            "optional.h",
             "scalar.h",
             "tensor.h",
             "tensor_impl.h",
-            "string_view.h",
             "device.h",
         ],
         # Only should be depended on by kernel_types:kernel_types, but various suffixes like Android and Static
diff --git a/runtime/core/portable_type/test/CMakeLists.txt b/runtime/core/portable_type/test/CMakeLists.txt
index b1f57a93ab5..8aa27d4d877 100644
--- a/runtime/core/portable_type/test/CMakeLists.txt
+++ b/runtime/core/portable_type/test/CMakeLists.txt
@@ -23,7 +23,6 @@ set(_test_srcs
     bfloat16_test.cpp
     dont_shadow_complex_test.c
     half_test.cpp
-    optional_test.cpp
     scalar_test.cpp
     tensor_impl_test.cpp
     tensor_test.cpp
diff --git a/runtime/core/portable_type/test/optional_test.cpp b/runtime/core/portable_type/test/optional_test.cpp
deleted file mode 100644
index 60d835d7439..00000000000
--- a/runtime/core/portable_type/test/optional_test.cpp
+++ /dev/null
@@ -1,142 +0,0 @@
-/*
- * Copyright (c) Meta Platforms, Inc. and affiliates.
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree.
- */
-
-#include <executorch/runtime/core/portable_type/optional.h>
-
-#include <string>
-#include <type_traits>
-
-#include <executorch/test/utils/DeathTest.h>
-#include <gtest/gtest.h>
-
-using namespace ::testing;
-using executorch::runtime::etensor::nullopt;
-using executorch::runtime::etensor::optional;
-
-// Test that optional::value_type matches the template parameter type.
-static_assert(
-    std::is_same<optional<int32_t>::value_type, int32_t>::value,
-    "Unexpected optional::value_type");
-static_assert(
-    std::is_same<optional<std::string>::value_type, std::string>::value,
-    "Unexpected optional::value_type");
-
-TEST(TestOptional, DefaultHasNoValue) {
-  optional<int32_t> o;
-  EXPECT_FALSE(o.has_value());
-}
-
-TEST(TestOptional, NulloptHasNoValue) {
-  optional<int32_t> o(nullopt);
-  EXPECT_FALSE(o.has_value());
-}
-
-TEST(TestOptional, ValueOfEmptyOptionalShouldThrow) {
-  optional<int32_t> o;
-  EXPECT_FALSE(o.has_value());
-
-  EXPECT_THROW({ (void)o.value(); }, std::bad_optional_access);
-}
-
-TEST(TestOptional, IntValue) {
-  optional<int32_t> o(15);
-  EXPECT_TRUE(o.has_value());
-  EXPECT_EQ(o.value(), 15);
-}
-
-TEST(TestOptional, NonTrivialValueType) {
-  optional<std::string> o("hey");
-  EXPECT_TRUE(o.has_value());
-  EXPECT_EQ(o.value(), "hey");
-}
-
-TEST(TestOptional, ConstValue) {
-  const optional<std::string> o("hey");
-  auto s = o.value(); // If this compiles, we're good.
-  EXPECT_EQ(o.value(), "hey");
-}
-
-TEST(TestOptional, CopyCtorWithValue) {
-  optional<int32_t> o1(15);
-  optional<int32_t> o2(o1);
-
-  EXPECT_TRUE(o2.has_value());
-  EXPECT_EQ(o2.value(), 15);
-}
-
-TEST(TestOptional, CopyCtorWithNoValue) {
-  optional<int32_t> o1;
-  optional<int32_t> o2(o1);
-
-  EXPECT_FALSE(o2.has_value());
-}
-
-TEST(TestOptional, CopyAssignTrivial) {
-  optional<int32_t> o1(1);
-  optional<int32_t> o2(2);
-  o1 = o2;
-
-  EXPECT_EQ(o1.value(), 2);
-}
-
-TEST(TestOptional, CopyAssignNonTrivial) {
-  optional<std::string> o1("abcde");
-  optional<std::string> o2("foo");
-  o1 = o2;
-
-  EXPECT_EQ(o1.value(), "foo");
-}
-
-TEST(TestOptional, CopyAssignNone) {
-  optional<int32_t> o1(2);
-  optional<int32_t> o2;
-  o1 = o2;
-  EXPECT_FALSE(o1.has_value());
-}
-
-TEST(TestOptional, MoveCtorWithNoValue) {
-  optional<int32_t> o1;
-  optional<int32_t> o2(std::move(o1));
-
-  EXPECT_FALSE(o2.has_value());
-}
-
-TEST(TestOptional, MoveCtorWithValue) {
-  optional<int32_t> o1(15);
-  optional<int32_t> o2(std::move(o1));
-
-  EXPECT_TRUE(o2.has_value());
-  EXPECT_EQ(o2.value(), 15);
-}
-
-TEST(TestOptional, MoveCtorNonTrivialType) {
-  optional<std::string> o1("abc");
-  optional<std::string> o2(std::move(o1));
-
-  EXPECT_TRUE(o2.has_value());
-  EXPECT_EQ(o2.value(), "abc");
-}
-
-optional<int32_t> function_returning_optional_of(int32_t value) {
-  return value;
-}
-
-TEST(TestOptional, ImplicitReturnOfValue) {
-  auto o = function_returning_optional_of(21);
-  EXPECT_TRUE(o.has_value());
-  EXPECT_EQ(o.value(), 21);
-}
-
-optional<int32_t> function_returning_nullopt() {
-  return nullopt;
-}
-
-TEST(TestOptional, ImplicitReturnOfNullopt) {
-  auto o = function_returning_nullopt();
-  EXPECT_FALSE(o.has_value());
-}
diff --git a/runtime/core/portable_type/test/targets.bzl b/runtime/core/portable_type/test/targets.bzl
index a6671d7d400..ce831f89327 100644
--- a/runtime/core/portable_type/test/targets.bzl
+++ b/runtime/core/portable_type/test/targets.bzl
@@ -15,14 +15,6 @@ def define_common_targets():
         ],
     )
 
-    runtime.cxx_test(
-        name = "optional_test",
-        srcs = ["optional_test.cpp"],
-        deps = [
-            "//executorch/runtime/core/portable_type:portable_type",
-        ],
-    )
-
     runtime.cxx_test(
         name = "tensor_test",
         srcs = ["tensor_test.cpp"],
diff --git a/runtime/executor/merged_data_map.h b/runtime/executor/merged_data_map.h
index d5ae97057f2..aae92d90a9b 100644
--- a/runtime/executor/merged_data_map.h
+++ b/runtime/executor/merged_data_map.h
@@ -57,7 +57,7 @@ class MergedDataMap final : public NamedDataMap {
    */
   ET_NODISCARD
   Result<const TensorLayout> get_tensor_layout(
-      executorch::aten::string_view key) const override {
+      std::string_view key) const override {
     auto layout = first_->get_tensor_layout(key);
     if (layout.ok()) {
       return layout.get();
@@ -76,8 +76,7 @@ class MergedDataMap final : public NamedDataMap {
    * @return error if the key is not present or data cannot be loaded.
    */
   ET_NODISCARD
-  Result<FreeableBuffer> get_data(
-      executorch::aten::string_view key) const override {
+  Result<FreeableBuffer> get_data(std::string_view key) const override {
     auto data = first_->get_data(key);
     if (data.error() != Error::NotFound) {
       return data;
@@ -97,7 +96,7 @@ class MergedDataMap final : public NamedDataMap {
    * @returns an Error indicating if the load was successful.
    */
   ET_NODISCARD Error load_data_into(
-      ET_UNUSED executorch::aten::string_view key,
+      ET_UNUSED std::string_view key,
       ET_UNUSED void* buffer,
       ET_UNUSED size_t size) const override {
     return Error::NotImplemented;
diff --git a/runtime/executor/pte_data_map.cpp b/runtime/executor/pte_data_map.cpp
index 881bfd5165a..e35745e7689 100644
--- a/runtime/executor/pte_data_map.cpp
+++ b/runtime/executor/pte_data_map.cpp
@@ -26,8 +26,7 @@ namespace internal {
 }
 
 ET_NODISCARD
-Result<FreeableBuffer> PteDataMap::get_data(
-    executorch::aten::string_view key) const {
+Result<FreeableBuffer> PteDataMap::get_data(std::string_view key) const {
   for (uint32_t i = 0; i < named_data_->size(); i++) {
     const auto* named_data_item = named_data_->Get(i);
     ET_CHECK_OR_RETURN_ERROR(
diff --git a/runtime/executor/pte_data_map.h b/runtime/executor/pte_data_map.h
index b4b46a6b541..36d33ae3945 100644
--- a/runtime/executor/pte_data_map.h
+++ b/runtime/executor/pte_data_map.h
@@ -79,7 +79,7 @@ class PteDataMap final : public NamedDataMap {
    */
   ET_NODISCARD
   Result<const TensorLayout> get_tensor_layout(
-      ET_UNUSED executorch::aten::string_view key) const override {
+      ET_UNUSED std::string_view key) const override {
     return Error::NotImplemented;
   }
 
@@ -91,14 +91,13 @@ class PteDataMap final : public NamedDataMap {
    * @return error if the key is not present or data cannot be loaded.
    */
   ET_NODISCARD
-  Result<FreeableBuffer> get_data(
-      executorch::aten::string_view key) const override;
+  Result<FreeableBuffer> get_data(std::string_view key) const override;
 
   /**
    * The PteDataMap currently does not implement load_into.
    */
   ET_NODISCARD Error load_data_into(
-      ET_UNUSED executorch::aten::string_view key,
+      ET_UNUSED std::string_view key,
       ET_UNUSED void* buffer,
       ET_UNUSED size_t size) const override {
     return Error::NotImplemented;
diff --git a/runtime/executor/test/method_meta_test.cpp b/runtime/executor/test/method_meta_test.cpp
index a1991a0562c..1324a40cf52 100644
--- a/runtime/executor/test/method_meta_test.cpp
+++ b/runtime/executor/test/method_meta_test.cpp
@@ -38,7 +38,7 @@ class TensorInfoTestFriend final {
       Span<const uint8_t> dim_order,
       executorch::aten::ScalarType scalar_type,
       const bool is_memory_planned,
-      executorch::aten::string_view name) {
+      std::string_view name) {
     return TensorInfo::create(
                Span<const int32_t>(sizes.data(), sizes.size()),
                Span<const uint8_t>(dim_order.data(), dim_order.size()),
@@ -236,7 +236,7 @@ TEST_F(MethodMetaTest, TensorInfoSizeOverflow) {
           Span<const uint8_t>(dim_order.data(), dim_order.size()),
           executorch::aten::ScalarType::Float,
           false, // is_memory_planned
-          executorch::aten::string_view{nullptr, 0}),
+          std::string_view{nullptr, 0}),
       "");
 }
 
diff --git a/test/utils/OSSTestConfig.json b/test/utils/OSSTestConfig.json
index c0877aac924..d7d0cb08567 100644
--- a/test/utils/OSSTestConfig.json
+++ b/test/utils/OSSTestConfig.json
@@ -75,7 +75,6 @@
             "bfloat16_test.cpp",
             "dont_shadow_complex_test.c",
             "half_test.cpp",
-            "optional_test.cpp",
             "scalar_test.cpp",
             "tensor_impl_test.cpp",
             "tensor_test.cpp"