Update base for Update on "[ExecuTorch][#10447] Extend PyBundledModule with extension.BundledModule"

# Context

This issue is a step of #9638.

In #9638, we want to have `extension.Module` as the single source of implementation in `pybindings`, which means that `pybindings.PyModule` should use `extension.Module` rather than its own `pybindings.Module`.

# Proposal

Now that we have `extension.BundledModule` ready, we want to test it out by having our existing `PyBundledModule` to extend it, and let `verify_result_with_bundled_expected_output` to use it, so that we can test out the whole thing with https://github.com/pytorch/executorch/blob/fb45e19055a92d2a91a4d4b7008e135232cbb14b/devtools/bundled_program/test/test_end2end.py

Differential Revision: [D73564127](https://our.internmc.facebook.com/intern/diff/D73564127/)

[ghstack-poisoned]

Author: Gasoonjia

backends/cadence/fusion_g3/operators/op_clamp.cpp

@@ -21,13 +21,13 @@
 #include <executorch/kernels/portable/cpu/util/math_util.h>
 #include <executorch/runtime/kernel/kernel_includes.h>
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::canCast;
 using ::executorch::runtime::Error;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace cadence {
 namespace impl {

backends/cadence/fusion_g3/operators/op_dequantize.cpp

@@ -24,7 +24,7 @@ using ::executorch::runtime::Error;
 using ::executorch::runtime::KernelRuntimeContext;
 
 template <typename T>
-using optional = ::executorch::aten::optional<T>;
+using optional = std::optional<T>;
 /* ScalarType in Executorch do not have support for below data types.
  * So, creating a placeholder for these data types. Once, ScalarTypes is
  * updated to have support for below data types, these can be removed and
@@ -51,7 +51,7 @@ void check_dequantize_per_tensor_args(
     int64_t quant_min,
     int64_t quant_max,
     ScalarType dtype,
-    ::executorch::aten::optional<ScalarType>& out_dtype,
+    std::optional<ScalarType>& out_dtype,
     Tensor& out) {
   ET_CHECK_MSG(
       input.scalar_type() == ScalarType::Byte ||
@@ -93,7 +93,7 @@ Tensor& dequantize_impl(
     float* scale_data,
     int* zero_point_data,
     int* axis,
-    ::executorch::aten::optional<ScalarType> out_dtype) {
+    std::optional<ScalarType> out_dtype) {
   const ::executorch::aten::ArrayRef<Tensor::SizesType> input_size =
       input.sizes();
 
@@ -260,8 +260,8 @@ Tensor& dequantize_impl(
           }
         }
 
-        ::executorch::aten::optional<::executorch::aten::ArrayRef<int64_t>>
-            optional_dim_list{::executorch::aten::ArrayRef<int64_t>{
+        std::optional<::executorch::aten::ArrayRef<int64_t>> optional_dim_list{
+            ::executorch::aten::ArrayRef<int64_t>{
                 dims, size_t(input.dim() - 1)}};
 
 // Actual dequantization logic
@@ -466,8 +466,8 @@ Tensor& dequantize_impl(
           }
         }
 
-        ::executorch::aten::optional<::executorch::aten::ArrayRef<int64_t>>
-            optional_dim_list{::executorch::aten::ArrayRef<int64_t>{
+        std::optional<::executorch::aten::ArrayRef<int64_t>> optional_dim_list{
+            ::executorch::aten::ArrayRef<int64_t>{
                 dims, size_t(input.dim() - 1)}};
 
 // Actual dequantization logic
@@ -600,7 +600,7 @@ Tensor& dequantize_per_tensor_tensor_args_out(
     int64_t quant_min,
     int64_t quant_max,
     ScalarType dtype,
-    ::executorch::aten::optional<ScalarType> out_dtype,
+    std::optional<ScalarType> out_dtype,
     Tensor& out) {
 #ifdef OP_ARG_CHECK
   ET_CHECK_MSG(
@@ -639,12 +639,12 @@ Tensor& dequantize_per_channel_out(
     KernelRuntimeContext& context,
     const Tensor& input,
     const Tensor& scale,
-    const ::executorch::aten::optional<Tensor>& opt_zero_points,
+    const std::optional<Tensor>& opt_zero_points,
     int64_t axis,
     int64_t quant_min,
     int64_t quant_max,
     ScalarType dtype,
-    ::executorch::aten::optional<ScalarType> out_dtype,
+    std::optional<ScalarType> out_dtype,
     Tensor& out) {
   if (axis < 0) {
     axis += executorch::runtime::nonzero_dim(input);

backends/cadence/fusion_g3/operators/op_div.cpp

@@ -19,14 +19,14 @@
 #include <executorch/runtime/kernel/kernel_includes.h>
 #include <executorch/runtime/platform/assert.h>
 
-using ::executorch::aten::optional;
 using ::executorch::aten::Scalar;
 using ::executorch::aten::ScalarType;
-using ::executorch::aten::string_view;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::canCast;
 using ::executorch::runtime::Error;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
+using std::string_view;
 
 namespace cadence {
 namespace impl {
@@ -686,4 +686,4 @@ Tensor& div_scalar_mode_out(
 } // namespace native
 } // namespace G3
 } // namespace impl
-} // namespace cadence
+} // namespace cadence

backends/cadence/fusion_g3/operators/op_mean.cpp

@@ -17,11 +17,11 @@
 #include <executorch/runtime/platform/assert.h>
 
 using ::executorch::aten::ArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::Error;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace cadence {
 namespace impl {

backends/cadence/fusion_g3/operators/op_native_layer_norm.cpp

@@ -19,11 +19,11 @@
 #include <executorch/runtime/kernel/kernel_includes.h>
 
 using ::executorch::aten::IntArrayRef;
-using ::executorch::aten::optional;
 using ::executorch::aten::ScalarType;
 using ::executorch::aten::Tensor;
 using ::executorch::runtime::Error;
 using ::executorch::runtime::KernelRuntimeContext;
+using std::optional;
 
 namespace cadence {
 namespace impl {

backends/cadence/fusion_g3/operators/op_quantize.cpp

@@ -329,8 +329,8 @@ Tensor& quantize_impl(
           }
         }
 
-        ::executorch::aten::optional<::executorch::aten::ArrayRef<int64_t>>
-            optional_dim_list{::executorch::aten::ArrayRef<int64_t>{
+        std::optional<::executorch::aten::ArrayRef<int64_t>> optional_dim_list{
+            ::executorch::aten::ArrayRef<int64_t>{
                 dims, size_t(input.dim() - 1)}};
 
 // Actual quantization logic
@@ -534,8 +534,8 @@ Tensor& quantize_impl(
           }
         }
 
-        ::executorch::aten::optional<::executorch::aten::ArrayRef<int64_t>>
-            optional_dim_list{::executorch::aten::ArrayRef<int64_t>{
+        std::optional<::executorch::aten::ArrayRef<int64_t>> optional_dim_list{
+            ::executorch::aten::ArrayRef<int64_t>{
                 dims, size_t(input.dim() - 1)}};
 
 // Actual quantization logic

backends/cadence/fusion_g3/operators/op_slice_copy.cpp

@@ -37,8 +37,8 @@ Tensor& slice_copy_Tensor_out(
     KernelRuntimeContext& ctx,
     const Tensor& in,
     int64_t dim,
-    ::executorch::aten::optional<int64_t> start_val,
-    ::executorch::aten::optional<int64_t> end_val,
+    std::optional<int64_t> start_val,
+    std::optional<int64_t> end_val,
     int64_t step,
     Tensor& out) {
   (void)ctx;

backends/cadence/fusion_g3/operators/operators.h

@@ -47,13 +47,12 @@ ::executorch::aten::Tensor& dequantize_per_channel_out(
     ::executorch::runtime::KernelRuntimeContext& context,
     const ::executorch::aten::Tensor& input,
     const ::executorch::aten::Tensor& scale,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>&
-        opt_zero_points,
+    const std::optional<::executorch::aten::Tensor>& opt_zero_points,
     int64_t axis,
     int64_t quant_min,
     int64_t quant_max,
     ::executorch::aten::ScalarType dtype,
-    ::executorch::aten::optional<::executorch::aten::ScalarType> out_dtype,
+    std::optional<::executorch::aten::ScalarType> out_dtype,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& dequantize_per_tensor_out(
@@ -64,7 +63,7 @@ ::executorch::aten::Tensor& dequantize_per_tensor_out(
     int64_t quant_min,
     int64_t quant_max,
     ::executorch::aten::ScalarType dtype,
-    ::executorch::aten::optional<::executorch::aten::ScalarType> out_dtype,
+    std::optional<::executorch::aten::ScalarType> out_dtype,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& div_out(
@@ -77,7 +76,7 @@ ::executorch::aten::Tensor& div_out_mode(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& a,
     const ::executorch::aten::Tensor& b,
-    ::executorch::aten::optional<::executorch::aten::string_view> mode,
+    std::optional<std::string_view> mode,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& div_scalar_out(
@@ -90,7 +89,7 @@ ::executorch::aten::Tensor& div_scalar_mode_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& a,
     const ::executorch::aten::Scalar& b,
-    ::executorch::aten::optional<::executorch::aten::string_view> mode,
+    std::optional<std::string_view> mode,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& exp_out(
@@ -101,10 +100,9 @@ ::executorch::aten::Tensor& exp_out(
 ::executorch::aten::Tensor& mean_dim_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& in,
-    ::executorch::aten::optional<::executorch::aten::ArrayRef<int64_t>>
-        dim_list,
+    std::optional<::executorch::aten::ArrayRef<int64_t>> dim_list,
     bool keepdim,
-    ::executorch::aten::optional<::executorch::aten::ScalarType> dtype,
+    std::optional<::executorch::aten::ScalarType> dtype,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& mul_out(
@@ -127,8 +125,8 @@ native_layer_norm_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& input,
     ::executorch::aten::IntArrayRef normalized_shape,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& weight,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& bias,
+    const std::optional<::executorch::aten::Tensor>& weight,
+    const std::optional<::executorch::aten::Tensor>& bias,
     double eps,
     ::executorch::aten::Tensor& out,
     ::executorch::aten::Tensor& mean_out,
@@ -165,8 +163,8 @@ ::executorch::aten::Tensor& slice_copy_Tensor_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& in,
     int64_t dim,
-    ::executorch::aten::optional<int64_t> start_val,
-    ::executorch::aten::optional<int64_t> end_val,
+    std::optional<int64_t> start_val,
+    std::optional<int64_t> end_val,
     int64_t step,
     ::executorch::aten::Tensor& out);
 
@@ -226,15 +224,15 @@ ::executorch::aten::Tensor& where_out(
 ::executorch::aten::Tensor& clamp_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& in,
-    const ::executorch::aten::optional<::executorch::aten::Scalar>& min_opt,
-    const ::executorch::aten::optional<::executorch::aten::Scalar>& max_opt,
+    const std::optional<::executorch::aten::Scalar>& min_opt,
+    const std::optional<::executorch::aten::Scalar>& max_opt,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& clamp_tensor_out(
     ::executorch::runtime::KernelRuntimeContext& ctx,
     const ::executorch::aten::Tensor& in,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& min_opt,
-    const ::executorch::aten::optional<::executorch::aten::Tensor>& max_opt,
+    const std::optional<::executorch::aten::Tensor>& min_opt,
+    const std::optional<::executorch::aten::Tensor>& max_opt,
     ::executorch::aten::Tensor& out);
 
 ::executorch::aten::Tensor& transpose_copy_int_out(

backends/cadence/hifi/operators/op_clamp.cpp

@@ -51,8 +51,8 @@ namespace native {
 Tensor& clamp_Tensor_out(
     RuntimeContext& ctx,
     const Tensor& in,
-    const executorch::aten::optional<Tensor>& min_opt,
-    const executorch::aten::optional<Tensor>& max_opt,
+    const std::optional<Tensor>& min_opt,
+    const std::optional<Tensor>& max_opt,
     Tensor& out) {
   (void)ctx;
 
@@ -325,8 +325,8 @@ Tensor& clamp_Tensor_out(
 Tensor& clamp_tensor_out(
     RuntimeContext& ctx,
     const Tensor& in,
-    const executorch::aten::optional<Tensor>& min_opt,
-    const executorch::aten::optional<Tensor>& max_opt,
+    const std::optional<Tensor>& min_opt,
+    const std::optional<Tensor>& max_opt,
     Tensor& out) {
   return clamp_Tensor_out(ctx, in, min_opt, max_opt, out);
 }

backends/cadence/hifi/operators/op_div.cpp

@@ -178,7 +178,7 @@ Tensor& div_out_mode(
     RuntimeContext& ctx,
     const Tensor& a,
     const Tensor& b,
-    executorch::aten::optional<executorch::aten::string_view> mode,
+    std::optional<std::string_view> mode,
     Tensor& out) {
   ET_KERNEL_CHECK(
       ctx,