pytorch/torch/csrc/dynamo/guards.cpp

#include <ATen/PythonTorchFunctionTLS.h>
#include <ATen/autocast_mode.h>
#include <c10/core/SafePyObject.h>
#include <c10/core/impl/PyInterpreter.h>
#define PY_SSIZE_T_CLEAN
#include <ATen/EmptyTensor.h>
#include <ATen/SparseCsrTensorUtils.h>
#include <c10/util/flat_hash_map.h>
#include <torch/csrc/autograd/grad_mode.h>
#include <torch/csrc/autograd/utils/wrap_outputs.h>
#include <torch/csrc/dynamo/guards.h>
#include <torch/csrc/inductor/inductor_ops.h>
#include <torch/csrc/utils/disable_torch_function.h>
#include <torch/csrc/utils/python_arg_parser.h>
#include <torch/csrc/utils/python_compat.h>
#include <torch/csrc/utils/python_numbers.h>
#include <torch/csrc/utils/python_symnode.h>
#include <torch/csrc/utils/pythoncapi_compat.h>
#include <torch/extension.h>

#include <torch/csrc/dynamo/debug_macros.h>

#include <nlohmann/json.hpp>

#ifdef USE_CUDA
#include <ATen/cuda/EmptyTensor.h>
#endif

#ifdef USE_XPU
#include <ATen/xpu/EmptyTensor.h>
#endif

#include <chrono>
#include <sstream>
#include <tuple>
#include <utility>

// Certain CPython data structures are defined in `.c` files in earlier Python
// versions, e.g., for TupleIteratorGetItemAccessor, we need a fast way to
// retrieve the underlying tuple and access the item. Before Python 3.12
// version, the data structure is in tupleobject.c file -
// https://github.com/python/cpython/blob/9afc6d102d16080535325f645849cd84eb04d57d/Objects/tupleobject.c#L1058-L1062
//
// To handle the older python versions, we manually copy the struct here and
// manually cast it to this new struct. For newer versions, the struct is
// included in the header file.
#if IS_PYTHON_3_12_PLUS

#define Py_BUILD_CORE
#include <internal/pycore_range.h> // _PyRangeIterObject
#include <internal/pycore_tuple.h> // _PyTupleIterObject
#undef Py_BUILD_CORE

#else

// Manually create _PyTupleIterObject struct
typedef struct {
  PyObject_HEAD
  Py_ssize_t it_index;
  PyTupleObject* it_seq; /* Set to NULL when iterator is exhausted */
} _PyTupleIterObject;

// Copied from CPython, and given a unified name for different Python versions.
// https://github.com/python/cpython/blob/7f71003b222ad398713514c2b55d34dc05dba6bc/Objects/rangeobject.c#L765-L771
typedef struct {
  PyObject_HEAD
  // NOTE for Python 3.12+, `index` is removed, and `start` is updated in place
  // instead, upon each `next(...)` call. See
  // https://github.com/python/cpython/pull/27986
  long index;
  long start;
  long step;
  long len;
} _PyRangeIterObject;

#endif // IS_PYTHON_3_12_PLUS

namespace torch::dynamo {

// Macro to skip addition of duplicate guards like EQUALS_MATCH
#define SKIP_IF_GUARD_ALREADY_PRESENT(name) \
  if (self.is_leaf_guard_present(name)) {   \
    return;                                 \
  }                                         \
  self.insert_leaf_guard(name);

TensorCheck::TensorCheck(
    const LocalState& state,
    PyTypeObject* pt,
    const at::Tensor& v,
    c10::DispatchKeySet dispatch_key_set,
    std::vector<std::optional<c10::SymInt>> dynamic_dims_sizes,
    std::vector<std::optional<c10::SymInt>> dynamic_dims_strides)
    : pytype(pt),
      dispatch_key_(state.apply(dispatch_key_set).raw_repr()),
      dtype_(v.dtype().toScalarType()),
      device_index_(v.device().index()),
      requires_grad_(v.requires_grad()),
      sizes_(std::move(dynamic_dims_sizes)),
      strides_(std::move(dynamic_dims_strides)),
      dim_(static_cast<int64_t>(sizes_.size())) {
  // TODO(voz): In cases where sizes_ and strides_ are fully dynamic, should
  // we just treat this as optional?
}

TensorCheck::TensorCheck(
    const LocalState& state,
    PyTypeObject* pt,
    c10::DispatchKeySet dispatch_key_set,
    at::ScalarType dtype,
    at::DeviceIndex device_index,
    bool requires_grad,
    std::vector<std::optional<c10::SymInt>> dynamic_dims_sizes,
    std::vector<std::optional<c10::SymInt>> dynamic_dims_strides)
    : pytype(pt),
      dispatch_key_(state.apply(dispatch_key_set).raw_repr()),
      dtype_(dtype),
      device_index_(device_index),
      requires_grad_(requires_grad),
      sizes_(std::move(dynamic_dims_sizes)),
      strides_(std::move(dynamic_dims_strides)),
      dim_(static_cast<int64_t>(sizes_.size())) {}

// See note in guards.py [Note - On Export Tensor Guards]
// Logic parallel to here must be maintained in python
bool TensorCheck::check(const LocalState& state, const at::Tensor& v) {
  // In terms of a sparse_csr tensor, it does not support strides information
  c10::SymIntArrayRef sym_strides(std::vector<SymInt>(v.ndimension(), -1));
  bool does_not_support_stride = v.layout() == c10::kSparseCsr ||
      v.layout() == c10::kSparseCsc || v.layout() == c10::kSparseBsc ||
      v.layout() == c10::kSparseBsr;
  if (!does_not_support_stride) {
    sym_strides = v.sym_strides();
  }

  return check(
      state,
      v.key_set(),
      v.dtype().toScalarType(),
      v.device(),
      v.sym_sizes(),
      sym_strides,
      v.requires_grad());
}

bool TensorCheck::check(
    const LocalState& state,
    const c10::DispatchKeySet& dispatch_key_set,
    const at::ScalarType& dtype,
    const c10::Device& device,
    const c10::SymIntArrayRef& sym_sizes,
    const c10::SymIntArrayRef& sym_strides,
    const bool& requires_grad) {
  if (dispatch_key_ != state.apply(dispatch_key_set).raw_repr() ||
      dtype_ != dtype || device_index_ != device.index() ||
      requires_grad_ != requires_grad) {
    return false;
  }

  auto ndim = sym_sizes.size();
  if (ndim != static_cast<size_t>(dim_)) {
    return false;
  }

  const auto& sizes = sym_sizes;
  const auto& strides = sym_strides;
  for (auto i : c10::irange(ndim)) {
    auto known_size = sizes_[i];
    auto known_stride = strides_[i];
    if (known_size.has_value()) {
      if (known_size.value() != sizes[i]) {
        return false;
      }
    }
    if (known_stride.has_value()) {
      if (known_stride.value() != strides[i]) {
        return false;
      }
    }
  }
  return true;
}

std::string TensorCheck::check_verbose(
    const LocalState& state,
    const at::Tensor& v,
    const std::string& tensor_name) {
  std::stringstream fail_reason;
  fail_reason << "tensor '" << tensor_name << "' ";
  if (dispatch_key_ != state.apply(v.key_set()).raw_repr()) {
    // return fmt::format("tensor dispatch key mismatch. expected {}, actual
    // {}", dispatch_key_, state.apply(v.key_set()).raw_repr());
    fail_reason << "dispatch key set mismatch. expected "
                << c10::DispatchKeySet(c10::DispatchKeySet::RAW, dispatch_key_)
                << ", actual " << state.apply(v.key_set());
    return fail_reason.str();
  } else if (dtype_ != v.dtype().toScalarType()) {
    // return fmt::format("tensor dtype mismatch. expected {}, actual {}",
    // dtype_, v.dtype().toScalarType());
    fail_reason << "dtype mismatch. expected " << dtype_ << ", actual "
                << v.dtype().toScalarType();
    return fail_reason.str();
  } else if (device_index_ != v.device().index()) {
    fail_reason << "Tensor device index mismatch. Expected device index to be "
                << device_index_ << ", actual " << v.device().index();
    return fail_reason.str();
  } else if (requires_grad_ != v.requires_grad()) {
    // return fmt::format("tensor requires_grad mismatch. expected {}",
    // requires_grad_);
    fail_reason << "requires_grad mismatch. expected requires_grad="
                << requires_grad_;
    return fail_reason.str();
  }
  auto ndim = v.ndimension();
  if (ndim != dim_) {
    // return fmt::format("tensor rank mismatch. expected {}, actual {}",
    // sizes_.size(), ndim);
    fail_reason << "rank mismatch. expected " << sizes_.size() << ", actual "
                << ndim;
    return fail_reason.str();
  }
  const auto& sizes = v.sym_sizes();
  for (auto i : c10::irange(ndim)) {
    auto known_size = sizes_[i];
    if (known_size.has_value() && (known_size.value() != sizes[i])) {
      fail_reason << "size mismatch at index " << i << ". expected "
                  << known_size.value() << ", actual " << sizes[i];
      return fail_reason.str();
    }
  }
  const bool supports_stride =
      !v.is_sparse() && !at::sparse_csr::is_sparse_compressed(v);
  if (supports_stride) {
    const auto& strides = v.sym_strides();
    for (auto i : c10::irange(ndim)) {
      auto known_stride = strides_[i];
      if (known_stride.has_value() && known_stride.value() != strides[i]) {
        fail_reason << "stride mismatch at index " << i << ". expected "
                    << known_stride.value() << ", actual " << strides[i];
        return fail_reason.str();
      }
    }
  }
  return "";
}

namespace {

typedef std::vector<TensorCheck> ChecksList;

typedef struct {
  PyObject_HEAD
  ChecksList* checks;
} TensorGuards;

static void TensorGuards_dealloc(TensorGuards* self) {
  if (self->checks != nullptr) {
    delete self->checks;
    self->checks = nullptr;
  }
  Py_TYPE(self)->tp_free((PyObject*)self);
}

static PyObject* TensorGuards_new(
    PyTypeObject* type,
    PyObject* args,
    PyObject* kwds) {
  TensorGuards* self = (TensorGuards*)type->tp_alloc(type, 0);
  if (self != nullptr) {
    self->checks = new ChecksList();
  }
  return (PyObject*)self;
}

static std::vector<std::optional<c10::SymInt>> wrapIntegersInOptional(
    const c10::SymIntArrayRef& intArray) {
  std::vector<std::optional<c10::SymInt>> optVec(intArray.size());
  std::transform(
      intArray.begin(),
      intArray.end(),
      optVec.begin(),
      [](const c10::SymInt& value) { return value; });
  return optVec;
}

static std::vector<std::optional<c10::SymInt>> pyListToVecOptInt(
    PyObject* pyList) {
  std::vector<std::optional<c10::SymInt>> vec;
  Py_ssize_t size = PyList_Size(pyList);
  for (Py_ssize_t i = 0; i < size; i++) {
    PyObject* item = PyList_GetItem(pyList, i);
    auto handle = py::handle(item);
    if (item == Py_None) {
      vec.emplace_back(std::nullopt);
    } else if (torch::is_symint(handle)) {
      vec.emplace_back(py::cast<c10::SymInt>(handle));
    } else {
      int64_t value = PyLong_AsLongLong(item);
      if (value == -1 && PyErr_Occurred()) {
        PyErr_SetString(
            PyExc_TypeError,
            "Size or stride list item is not a valid integer.");
        TORCH_CHECK(false, "Size or stride list item is not a valid integer.");
      }
      vec.emplace_back(c10::SymInt(value));
    }
  }
  return vec;
}

static std::vector<std::vector<std::optional<c10::SymInt>>> get_dynamic_dims(
    PyObject* dynamic_dims_py) {
  std::vector<std::vector<std::optional<c10::SymInt>>> per_tensor_dynamic_dims;
  if (dynamic_dims_py != Py_None) {
    Py_ssize_t size = PyList_Size(dynamic_dims_py);
    for (Py_ssize_t i = 0; i < size; i++) {
      PyObject* py_list = PyList_GetItem(dynamic_dims_py, i);
      std::vector<std::optional<c10::SymInt>> vec = pyListToVecOptInt(py_list);
      per_tensor_dynamic_dims.push_back(std::move(vec));
    }
  }
  return per_tensor_dynamic_dims;
}

static int TensorGuards_init(
    TensorGuards* self,
    PyObject* args,
    PyObject* kwds) {
  if (!PyTuple_CheckExact(args)) {
    PyErr_SetString(PyExc_TypeError, "expected tuple()");
    return -1;
  }
  // Top level structure is List[List[Union[int, None]]]
  PyObject* dynamic_dims_sizes_py =
      PyDict_GetItemString(kwds, "dynamic_dims_sizes");
  if (dynamic_dims_sizes_py == nullptr) {
    PyErr_SetString(PyExc_TypeError, "missing dynamic_dims_sizes=...");
    return -1;
  }
  PyObject* dynamic_dims_strides_py =
      PyDict_GetItemString(kwds, "dynamic_dims_strides");
  if (dynamic_dims_strides_py == nullptr) {
    PyErr_SetString(PyExc_TypeError, "missing dynamic_dims_strides=...");
    return -1;
  }

  // dynamic_dims_strides/sizes_py is None when dynamic_shapes=False - this is
  // an optimization to avoid invoking .size()/.stride() in python needlessly
  std::vector<std::vector<std::optional<c10::SymInt>>>
      per_tensor_dynamic_dims_sizes = get_dynamic_dims(dynamic_dims_sizes_py);
  std::vector<std::vector<std::optional<c10::SymInt>>>
      per_tensor_dynamic_dims_strides =
          get_dynamic_dims(dynamic_dims_strides_py);

  auto& checks = *self->checks;
  auto len = PyTuple_GET_SIZE(args);
  checks.reserve(len);
  LocalState state;

  for (auto i : c10::irange(len)) {
    PyObject* item = PyTuple_GET_ITEM(args, i);
    if (!THPVariable_CheckExact(item) && !THPVariable_Check(item)) {
      PyErr_SetString(PyExc_TypeError, "expected Tensor()");
      return -1;
    }
    auto tensor = THPVariable_Unpack(item);
    std::vector<std::optional<c10::SymInt>> tensor_dims_size =
        per_tensor_dynamic_dims_sizes.empty()
        ? wrapIntegersInOptional(tensor.sym_sizes())
        : per_tensor_dynamic_dims_sizes[i];
    std::vector<std::optional<c10::SymInt>> tensor_dims_stride =
        per_tensor_dynamic_dims_strides.empty()
        ? wrapIntegersInOptional(tensor.sym_strides())
        : per_tensor_dynamic_dims_strides[i];

    checks.emplace_back(
        state,
        Py_TYPE(item),
        std::move(tensor),
        tensor.key_set(),
        std::move(tensor_dims_size),
        std::move(tensor_dims_stride));
  }
  return 0;
}

PyObject* TensorGuards_check(
    TensorGuards* self,
    PyObject* args,
    PyObject* kwargs) {
  if (!PyTuple_CheckExact(args)) {
    PyErr_SetString(PyExc_TypeError, "expected tuple()");
    return nullptr;
  }
  auto& checks = *self->checks;
  auto len = PyTuple_GET_SIZE(args);

  // kwargs is just ignored here

  if (static_cast<decltype(len)>(checks.size()) != len) {
    PyErr_SetString(PyExc_TypeError, "wrong length");
    return nullptr;
  }

  LocalState state;
  // Note - all the tensors that make it to guards must be unique. Dynamo
  // builder handles guarding for positive aliases (X is Y). However, we do not
  // create guards for negative alias (X is not Y) as that is an N^2
  // relationship. Instead, we rely on the uniqueness upstream to verify, at
  // check_fn time (this function).
  ska::flat_hash_map<PyObject*, std::nullptr_t> unique_tensors;
  for (auto i : c10::irange(len)) {
    PyObject* item = PyTuple_GET_ITEM(args, i);

    if (Py_TYPE(item) != checks[i].pytype) {
      Py_RETURN_FALSE;
    }
    auto insertion = unique_tensors.insert({item, nullptr});
    if (!insertion.second) {
      // Violates uniqueness
      Py_RETURN_FALSE;
    }
    if (!checks[i].check(state, THPVariable_Unpack(item))) {
      Py_RETURN_FALSE;
    }
  }

  Py_RETURN_TRUE;
}

PyObject* TensorGuards_check_verbose(
    TensorGuards* self,
    PyObject* args,
    PyObject* kwargs) {
  if (!PyTuple_CheckExact(args)) {
    PyErr_SetString(PyExc_TypeError, "expected tuple()");
    return nullptr;
  }
  auto& checks = *self->checks;
  auto len = PyTuple_GET_SIZE(args);

  if (static_cast<decltype(len)>(checks.size()) != len) {
    PyErr_SetString(PyExc_TypeError, "wrong length");
    return nullptr;
  }

  PyObject* tensor_check_names_py =
      PyDict_GetItemString(kwargs, "tensor_check_names");
  if (tensor_check_names_py == nullptr) {
    PyErr_SetString(PyExc_TypeError, "missing tensor_check_names kwarg");
    return nullptr;
  }

  if (!PyList_Check(tensor_check_names_py)) {
    PyErr_SetString(PyExc_TypeError, "tensor_check_names kwarg must be a list");
    return nullptr;
  }

  auto names_size = PyList_Size(tensor_check_names_py);
  if (names_size != static_cast<decltype(names_size)>(checks.size())) {
    PyErr_SetString(
        PyExc_TypeError,
        "tensor_check_names should be the same size as # tensors");
    return nullptr;
  }

  std::vector<std::string> tensor_check_names;
  tensor_check_names.reserve(names_size);
  for (auto i : c10::irange(names_size)) {
    PyObject* value = PyList_GetItem(tensor_check_names_py, i);
    if (!PyUnicode_Check(value)) {
      PyErr_SetString(
          PyExc_TypeError, "tensor_check_names must only contain strings");
      return nullptr;
    }
    tensor_check_names.emplace_back(PyUnicode_AsUTF8(value));
  }

  LocalState state;
  ska::flat_hash_map<PyObject*, std::nullptr_t> unique_tensors;
  for (auto i : c10::irange(len)) {
    PyObject* item = PyTuple_GET_ITEM(args, i);
    if (Py_TYPE(item) != checks[i].pytype) {
      std::stringstream fail_reason;
      PyObject* type_str = PyObject_Str(PyObject_Type(item));
      fail_reason << "expected type of '" << tensor_check_names[i]
                  << "' to be a tensor type, ";
      if (!type_str) {
        fail_reason << "but found a different type";
      } else {
        fail_reason << "' but found " << PyUnicode_AsUTF8(type_str);
      }
      return Py_BuildValue("s", fail_reason.str().c_str());
    }

    auto insertion = unique_tensors.insert({item, nullptr});
    if (!insertion.second) {
      std::stringstream fail_reason;
      fail_reason << "Duplicate tensor found where not expected! ";
      fail_reason << tensor_check_names[i]
                  << "should not alias to anything, but is aliased";
      return Py_BuildValue("s", fail_reason.str().c_str());
    }
    std::string fail_reason = checks[i].check_verbose(
        state, THPVariable_Unpack(item), tensor_check_names[i]);
    if (!fail_reason.empty()) {
      return Py_BuildValue("s", fail_reason.c_str());
    }
  }

  Py_RETURN_TRUE;
}

// NOLINTNEXTLINE(modernize-avoid-c-arrays,cppcoreguidelines-avoid-c-arrays)
static PyMethodDef TensorGuards_methods[] = {
    {"check",
     (PyCFunction)(void*)TensorGuards_check,
     METH_VARARGS | METH_KEYWORDS,
     ""},
    {"check_verbose",
     (PyCFunction)(void*)TensorGuards_check_verbose,
     METH_VARARGS | METH_KEYWORDS,
     "verbose fail reasons for failed checks"},
    {nullptr} /* Sentinel */
};

static PyTypeObject TensorGuardsType = { PyVarObject_HEAD_INIT(nullptr, 0)
};

struct AutocastState {
  static constexpr auto& DEVICES = at::autocast::_AUTOCAST_SUPPORTED_DEVICES;
  std::array<bool, DEVICES.size()> enabled{};
  std::array<at::ScalarType, DEVICES.size()> dtype{};
  bool cache_enabled;

  AutocastState() {
    for (size_t i = 0; i < DEVICES.size(); i++) {
      enabled[i] = at::autocast::is_autocast_enabled(DEVICES[i]);
      dtype[i] = at::autocast::get_autocast_dtype(DEVICES[i]);
    }
    cache_enabled = at::autocast::is_autocast_cache_enabled();
  }

  bool operator==(const AutocastState& o) const {
    for (size_t i = 0; i < DEVICES.size(); i++) {
      // If disabled audocast, autocast_dtype comparison not occur
      if (enabled[i] == false && o.enabled[i] == false) {
        continue;
      }
      if (enabled[i] != o.enabled[i] || dtype[i] != o.dtype[i]) {
        return false;
      }
    }
    if (cache_enabled != o.cache_enabled) {
      return false;
    }
    return true;
  }

  template <typename T>
  friend void to_json(T& json_j, const AutocastState& json_t) {
    json_j["enabled"] = json_t.enabled;
    json_j["dtype"] = json_t.dtype;
    json_j["cached_enabled"] = json_t.cache_enabled;
  }

  template <typename T>
  friend void from_json(const T& json_j, AutocastState& json_t) {
    json_t.enabled = json_j.at("enabled");
    json_t.dtype = json_j.at("dtype");
    json_t.cache_enabled = json_j.at("cached_enabled");
  }
};

// TODO (janimesh) - Remove the PyObject_HEAD part when C++ guard manager is
// merged.
// NOLINTNEXTLINE(cppcoreguidelines-pro-type-member-init)
struct GlobalStateGuard {
  PyObject_HEAD

  inline void init() {
    auto& ctx = at::globalContext();
    _grad_mode = at::GradMode::is_enabled();
    _autocast_state = AutocastState();
    // The below two flags disambiguate
    // if torch function disabled state is
    // 1) enabled, 2) all disabled, 3) subclasses disabled
    // we guard on the stack separately
    _torch_function = torch::torch_function_enabled();
    _torch_function_all_disabled = at::impl::torch_function_all_disabled();
    _deterministic_algorithms = ctx.deterministicAlgorithms();
    _deterministic_algorithms_warn_only = ctx.deterministicAlgorithmsWarnOnly();
    _allow_tf32 = ctx.float32Precision("cuda", "matmul") == "tf32";
    _allow_fp16_reduce = ctx.allowFP16ReductionCuBLAS();
    _allow_bf16_reduce = ctx.allowBF16ReductionCuBLAS();
    _num_threads = at::get_num_threads();
    _default_dtype = at::get_default_dtype();
  }

  inline bool check() const {
    auto& ctx = at::globalContext();
    return (_grad_mode == at::GradMode::is_enabled() &&
            _autocast_state == AutocastState() &&
            _torch_function == torch::torch_function_enabled() &&
            _torch_function_all_disabled ==
                at::impl::torch_function_all_disabled() &&
            _deterministic_algorithms == ctx.deterministicAlgorithms() &&
            _deterministic_algorithms_warn_only ==
                ctx.deterministicAlgorithmsWarnOnly() &&
            _allow_tf32 == (ctx.float32Precision("cuda", "matmul") == "tf32") &&
            _allow_fp16_reduce == ctx.allowFP16ReductionCuBLAS() &&
            _allow_bf16_reduce == ctx.allowBF16ReductionCuBLAS() &&
            _num_threads == at::get_num_threads()) &&
        _default_dtype == at::get_default_dtype();
  }

  inline std::string reason() const {
    std::ostringstream os;
    auto& ctx = at::globalContext();
    if (_grad_mode != at::GradMode::is_enabled())
      os << "grad_mode ";
    if (!(_autocast_state == AutocastState()))
      os << "autocast ";
    if (_torch_function != torch::torch_function_enabled())
      os << "torch_function ";
    if (_deterministic_algorithms != ctx.deterministicAlgorithms())
      os << "deterministic_algorithms ";
    if (_deterministic_algorithms_warn_only !=
        ctx.deterministicAlgorithmsWarnOnly())
      os << "deterministic_algorithms_warn_only ";
    if (_allow_tf32 != (ctx.float32Precision("cuda", "matmul") == "tf32"))
      os << "allow_tf32 ";
    if (_allow_fp16_reduce != ctx.allowFP16ReductionCuBLAS())
      os << "allow_fp16_reduce ";
    if (_allow_bf16_reduce != ctx.allowBF16ReductionCuBLAS())
      os << "allow_bf16_reduce ";
    if (_num_threads != at::get_num_threads())
      os << "num_threads ";
    if (_default_dtype != at::get_default_dtype())
      os << "default_dtype ";
    return os.str();
  }

  template <typename T>
  friend void to_json(T& json_j, const GlobalStateGuard& json_t) {
    json_j["grad_mode"] = json_t._grad_mode;
    json_j["autocast_state"] = json_t._autocast_state;
    json_j["torch_function"] = json_t._torch_function;
    json_j["torch_function_all_disabled"] = json_t._torch_function_all_disabled;
    json_j["deterministic_algorithms"] = json_t._deterministic_algorithms;
    json_j["deterministic_algorithms_warn_only"] =
        json_t._deterministic_algorithms_warn_only;
    json_j["allow_tf32"] = json_t._allow_tf32;
    json_j["allow_fp16_reduce"] = json_t._allow_fp16_reduce;
    json_j["allow_bf16_reduce"] = json_t._allow_bf16_reduce;
    json_j["num_threads"] = json_t._num_threads;
    json_j["default_dtype"] = json_t._default_dtype.toScalarType();
  }

  template <typename T>
  friend void from_json(const T& json_j, GlobalStateGuard& json_t) {
    json_t._grad_mode = json_j.at("grad_mode");
    json_t._autocast_state = json_j.at("autocast_state");
    json_t._torch_function = json_j.at("torch_function");
    json_t._torch_function_all_disabled =
        json_j.at("torch_function_all_disabled");
    json_t._deterministic_algorithms = json_j.at("deterministic_algorithms");
    json_t._deterministic_algorithms_warn_only =
        json_j.at("deterministic_algorithms_warn_only");
    json_t._allow_tf32 = json_j.at("allow_tf32");
    json_t._allow_fp16_reduce = json_j.at("allow_fp16_reduce");
    json_t._allow_bf16_reduce = json_j.at("allow_bf16_reduce");
    json_t._num_threads = json_j.at("num_threads");
    json_t._default_dtype =
        caffe2::TypeMeta::fromScalarType(json_j.at("default_dtype"));
  }

  bool _grad_mode;
  AutocastState _autocast_state;
  bool _torch_function;
  bool _torch_function_all_disabled;
  bool _deterministic_algorithms;
  bool _deterministic_algorithms_warn_only;
  bool _allow_tf32;
  bool _allow_fp16_reduce;
  bool _allow_bf16_reduce;
  int _num_threads;
  caffe2::TypeMeta _default_dtype;
  // TODO(jansel): we should guard on more state as inductor starts using it
};

int GlobalStateGuard_init(
    GlobalStateGuard* self,
    PyObject* args,
    PyObject* kwargs) {
  self->init();
  return 0;
}

PyObject* GlobalStateGuard_check(
    GlobalStateGuard* self,
    PyObject* args,
    PyObject* kwargs) {
  if (self->check()) {
    Py_RETURN_TRUE;
  } else {
    Py_RETURN_FALSE;
  }
}

PyObject* GlobalStateGuard_reason(
    GlobalStateGuard* self,
    PyObject* args,
    PyObject* kwargs) {
  return PyUnicode_FromString(self->reason().c_str());
}

PyObject* GlobalStateGuard_dump(
    GlobalStateGuard* self,
    PyObject* args,
    PyObject* kwargs) {
  return PyUnicode_FromString(nlohmann::json(*self).dump().c_str());
}

PyObject* GlobalStateGuard_load(
    GlobalStateGuard* self,
    PyObject* args,
    PyObject* kwargs) {
  char* json;
  if (!PyArg_ParseTuple(args, "s", &json)) {
    throw std::runtime_error("Cannot parse as json string.");
  }
  nlohmann::json::parse(json).get_to(*self);
  Py_RETURN_NONE;
}

// NOLINTNEXTLINE(*array*)
static PyMethodDef GlobalStateGuard_methods[] = {
    {"check",
     (PyCFunction)(void*)GlobalStateGuard_check,
     METH_NOARGS,
     "Return true if global state was the same as at creation time"},
    {"reason",
     (PyCFunction)(void*)GlobalStateGuard_reason,
     METH_NOARGS,
     "Return string reason for guard check failing"},
    {"dump",
     (PyCFunction)(void*)GlobalStateGuard_dump,
     METH_NOARGS,
     "Return serialized json format"},
    {"load",
     (PyCFunction)(void*)GlobalStateGuard_load,
     METH_VARARGS,
     "Parse serialized json format"},
    {nullptr}};
static PyTypeObject GlobalStateGuardType = { PyVarObject_HEAD_INIT(nullptr, 0)
};

static PyObject* check_type_id(PyObject* dummy, PyObject* args) {
  // faster `lambda obj, expected: id(type(obj)) == expected`
  PyObject* obj = nullptr;
  unsigned long long expected = 0;
  if (!PyArg_ParseTuple(args, "OK", &obj, &expected)) {
    return nullptr;
  }
  // NOLINTNEXTLINE(performance-no-int-to-ptr)
  if (Py_TYPE(obj) == (void*)expected) {
    Py_RETURN_TRUE;
  } else {
    Py_RETURN_FALSE;
  }
}

static PyObject* check_obj_id(PyObject* dummy, PyObject* args) {
  // faster `lambda obj, expected: id(obj) == expected`
  PyObject* obj = nullptr;
  unsigned long long expected = 0;
  if (!PyArg_ParseTuple(args, "OK", &obj, &expected)) {
    return nullptr;
  }
  // NOLINTNEXTLINE(performance-no-int-to-ptr)
  if (obj == (void*)expected) {
    Py_RETURN_TRUE;
  } else {
    Py_RETURN_FALSE;
  }
}

#if IS_PYTHON_3_12_PLUS

static std::unordered_map<PyObject*, uint64_t> dict_version_map;
static int dict_version_watcher_id;
static uint64_t global_dict_version_id = 1;
static int dict_version_watch_callback(
    PyDict_WatchEvent event,
    PyObject* dict,
    PyObject* key,
    PyObject* new_value) noexcept {
  if (event == PyDict_EVENT_DEALLOCATED) {
    dict_version_map.erase(dict);
  } else if (event != PyDict_EVENT_CLONED) {
    dict_version_map[dict] = global_dict_version_id++;
  }
  return 0;
}

#endif

static uint64_t get_dict_version_unchecked(PyObject* dict) {
#if IS_PYTHON_3_12_PLUS

  if (PyDict_Watch(dict_version_watcher_id, dict)) {
    throw std::runtime_error("failed to add version watcher to dict!");
  }
  if (!dict_version_map.count(dict)) {
    dict_version_map[dict] = global_dict_version_id++;
  }
  return dict_version_map[dict];

#else

  return ((PyDictObject*)dict)->ma_version_tag;

#endif
}

static PyObject* dict_version(PyObject* dummy, PyObject* args) {
  // Retrieves the version of a dictionary.
  PyObject* obj = nullptr;
  if (!PyArg_ParseTuple(args, "O", &obj)) {
    return nullptr;
  }
  if (!PyDict_Check(obj)) {
    return nullptr;
  }
  return THPUtils_packUInt64(get_dict_version_unchecked(obj));
}

static PyObject* assert_size_stride(PyObject* dummy, PyObject* args) {
  /*
   Assert that a given tensor has a given size/stride, but ignore strides
   of size==1 dimensions.  Implemented in C++ as this is on the hot path.
  */
  PyObject* item = nullptr;
  PyObject* size = nullptr;
  PyObject* stride = nullptr;
  const char* op_name = nullptr;

  if (!PyArg_ParseTuple(args, "OOO|s", &item, &size, &stride, &op_name)) {
    return nullptr;
  }
  if (!THPVariable_CheckExact(item) && !THPVariable_Check(item)) {
    std::stringstream msg;
    msg << "expected Tensor()";
    if (op_name) {
      msg << " for op: " << op_name;
    }
    PyErr_SetString(PyExc_TypeError, msg.str().c_str());
    return nullptr;
  }
  if (!PyTuple_CheckExact(size) || !PyTuple_CheckExact(stride)) {
    std::stringstream msg;
    msg << "expected tuple()";
    if (op_name) {
      msg << " for op: " << op_name;
    }
    PyErr_SetString(PyExc_TypeError, msg.str().c_str());
    return nullptr;
  }
  at::Tensor tensor = THPVariable_Unpack(item);
  int64_t ndim = tensor.ndimension();
  if (PyTuple_GET_SIZE(size) != ndim || PyTuple_GET_SIZE(stride) != ndim) {
    std::stringstream msg;
    msg << "wrong number of dimensions" << ndim;
    if (op_name) {
      msg << " for op: " << op_name;
    }
    PyErr_SetString(PyExc_AssertionError, msg.str().c_str());
    return nullptr;
  }

  // We may add the size/stride assert at compile time due to unbacked symint,
  // but at runtime, the tensor can be empty.
  if (tensor.numel() == 0) {
    Py_RETURN_TRUE;
  }

  std::stringstream msg;
  int num_errors = 0;
  for (auto i : c10::irange(ndim)) {
    int64_t want_size = THPUtils_unpackLong(PyTuple_GET_ITEM(size, i));
    int64_t want_stride = THPUtils_unpackLong(PyTuple_GET_ITEM(stride, i));
    int64_t actual_size = tensor.size(i);
    int64_t actual_stride = tensor.stride(i);
    if (want_size != actual_size ||
        // ignore stride differences when size is 1
        (want_stride != actual_stride && actual_size > 1)) {
      if (num_errors > 0)
        msg << "; ";
      msg << "expected size " << actual_size << "==" << want_size << ", stride "
          << actual_stride << "==" << want_stride << " at dim=" << i;
      num_errors++;
    }
  }

  if (num_errors) {
    if (op_name) {
      msg << "\nError in op: " << op_name;
    }
    msg << "\nThis error most often comes from a incorrect fake (aka meta) kernel for a custom op.";
    msg << "\nUse torch.library.opcheck to test your custom op.";
    msg << "\nSee https://pytorch.org/docs/stable/library.html#torch.library.opcheck";
    PyErr_SetString(PyExc_AssertionError, msg.str().c_str());
    return nullptr;
  }

  Py_RETURN_TRUE;
}

static PyObject* assert_alignment(PyObject* dummy, PyObject* args) {
  /*
   * Asserts that a given tensor meets certain alignment.
   * This C++ version of torch._inductor.utils.tensor_is_aligned
   */
  PyObject* item = nullptr;
  unsigned long alignment = 0;
  const char* op_name = nullptr;

  if (!PyArg_ParseTuple(args, "Ok|s", &item, &alignment, &op_name)) {
    return nullptr;
  }
  if (!THPVariable_CheckExact(item) && !THPVariable_Check(item)) {
    std::stringstream msg;
    msg << "expected Tensor()";
    if (op_name) {
      msg << " for op: " << op_name;
    }
    PyErr_SetString(PyExc_TypeError, msg.str().c_str());
    return nullptr;
  }
  if (alignment == 0) {
    std::stringstream msg;
    msg << "alignment cannot be 0";
    if (op_name) {
      msg << " in op: " << op_name;
    }
    PyErr_SetString(PyExc_AssertionError, msg.str().c_str());
    return nullptr;
  }

  at::Tensor tensor = THPVariable_Unpack(item);

  int64_t storage_offset = tensor.storage_offset();
  size_t itemsize = tensor.itemsize();
  if (storage_offset * itemsize % alignment != 0) {
    std::stringstream msg;
    if (op_name) {
      msg << "\nError in op: " << op_name;
    }
    msg << "\nExpect the tensor to be " << alignment
        << " bytes aligned. Fail due to storage_offset=" << storage_offset
        << " itemsize=" << itemsize;
    PyErr_SetString(PyExc_AssertionError, msg.str().c_str());
    return nullptr;
  }

  Py_RETURN_TRUE;
}

template <typename T>
static void unwrap_size_tuple(PyObject* obj, T& output) {
  TORCH_CHECK(PyTuple_CheckExact(obj));
  size_t len = PyTuple_GET_SIZE(obj);
  output.reserve(len);
  for (size_t i = 0; i < len; ++i) {
    auto result = PyLong_AsSsize_t(PyTuple_GET_ITEM(obj, i));
    TORCH_CHECK(result >= 0);
    output.emplace_back(result);
  }
}

template <typename T>
static void _parse_empty_strided_args(
    PyObject* args,
    T& sizes,
    T& strides,
    at::ScalarType& dtype) {
  TORCH_CHECK(PyTuple_CheckExact(args));
  TORCH_CHECK(PyTuple_GET_SIZE(args) == 3);
  // note PyTuple_GET_ITEM returns a borrowed ref, so no need for refcounts
  unwrap_size_tuple(PyTuple_GET_ITEM(args, 0), sizes);
  unwrap_size_tuple(PyTuple_GET_ITEM(args, 1), strides);
  PyObject* py_dtype = PyTuple_GET_ITEM(args, 2);
  TORCH_CHECK(THPDtype_Check(py_dtype));
  dtype = reinterpret_cast<THPDtype*>(py_dtype)->scalar_type;
}

static PyObject* _empty_strided_device(
    PyObject* dummy,
    PyObject* args,
    c10::DeviceType device_type) {
  HANDLE_TH_ERRORS;
  at::SmallVector<int64_t, 8> sizes;
  at::SmallVector<int64_t, 8> strides;
  at::ScalarType dtype{at::ScalarType::Undefined};
  _parse_empty_strided_args(args, sizes, strides, dtype);
  if (device_type == c10::DeviceType::CPU) {
    return THPVariable_Wrap(
        at::detail::empty_strided_cpu(sizes, strides, dtype));
  }
#ifdef USE_CUDA
  else if (device_type == c10::DeviceType::CUDA) {
    return THPVariable_Wrap(at::detail::empty_strided_cuda(
        sizes, strides, dtype, c10::DeviceType::CUDA));
  }
#endif
#ifdef USE_XPU
  else if (device_type == c10::DeviceType::XPU) {
    return THPVariable_Wrap(at::detail::empty_strided_xpu(
        sizes, strides, dtype, c10::DeviceType::XPU));
  }
#endif
  else {
    TORCH_CHECK(
        false, "PyTorch compiled without support for the specified device.");
  }

  END_HANDLE_TH_ERRORS;
}

static PyObject* _empty_strided_cpu(PyObject* dummy, PyObject* args) {
  // at::empty_strided is surprising slow.  This is a lower-overhead
  // version that saves ~2us on every allocation.
  return _empty_strided_device(dummy, args, c10::DeviceType::CPU);
}

static PyObject* _empty_strided_cuda(PyObject* dummy, PyObject* args) {
  // at::empty_strided is surprising slow.  This is lower-overhead.
  return _empty_strided_device(dummy, args, c10::DeviceType::CUDA);
}

static PyObject* _empty_strided_xpu(PyObject* dummy, PyObject* args) {
  // at::empty_strided is surprising slow.  This is lower-overhead.
  return _empty_strided_device(dummy, args, c10::DeviceType::XPU);
}

static PyObject* _reinterpret_tensor(PyObject* dummy, PyObject* args) {
  HANDLE_TH_ERRORS;
  static PythonArgParser parser(
      {"_reinterpret_tensor(Tensor base, IntArrayRef sizes, IntArrayRef strides, int64_t offset_increment=0)"},
      /*traceable=*/true);

  ParsedArgs<4> parsed_args;
  auto r = parser.parse(args, /*kwargs=*/nullptr, parsed_args);

  Tensor self = r.tensor(0);
  auto sizes = r.intlist(1);
  auto strides = r.intlist(2);
  auto offset_increment = r.toInt64(3);

  auto res = torch::inductor::_reinterpret_tensor(
      self, sizes, strides, offset_increment);
  return torch::autograd::utils::wrap(res);

  END_HANDLE_TH_ERRORS;
}

// NOLINTNEXTLINE(modernize-avoid-c-arrays,cppcoreguidelines-avoid-c-arrays)
static PyMethodDef _methods[] = {
    {"check_type_id", check_type_id, METH_VARARGS, nullptr},
    {"check_obj_id", check_obj_id, METH_VARARGS, nullptr},
    {"assert_size_stride", assert_size_stride, METH_VARARGS, nullptr},
    {"assert_alignment", assert_alignment, METH_VARARGS, nullptr},
    {"dict_version", dict_version, METH_VARARGS, nullptr},
    {"_empty_strided_cpu", _empty_strided_cpu, METH_VARARGS, nullptr},
    {"_empty_strided_cuda", _empty_strided_cuda, METH_VARARGS, nullptr},
    {"_empty_strided_xpu", _empty_strided_xpu, METH_VARARGS, nullptr},
    {"_reinterpret_tensor", _reinterpret_tensor, METH_VARARGS, nullptr},
    {nullptr, nullptr, 0, nullptr}};

static struct PyModuleDef _module = {
    PyModuleDef_HEAD_INIT,
    "torch._C._dynamo.guards",
    "Module containing checks on tensors",
    -1,
    _methods};

std::string get_exception_message() {
  PyObject *ptype = nullptr, *pvalue = nullptr, *ptraceback = nullptr;
  PyErr_Fetch(&ptype, &pvalue, &ptraceback);

  PyObject* exc_message_pyobj = PyObject_Str(pvalue);
  std::string exc_message = PyUnicode_AsUTF8(exc_message_pyobj);

  Py_DECREF(exc_message_pyobj);
  Py_XDECREF(ptype);
  Py_XDECREF(pvalue);
  Py_XDECREF(ptraceback);
  return exc_message;
}

bool is_immutable_object(py::handle example_value) {
  py::object config_module = py::module_::import("torch._dynamo.config");

  bool is_tensor_immutable =
      config_module.attr("skip_tensor_guards_with_matching_dict_tags")
          .cast<bool>();

  if (PyTuple_Check(example_value.ptr())) {
    // Check that each element is immutable
    for (Py_ssize_t i = 0; i < PyTuple_Size(example_value.ptr()); ++i) {
      if (!is_immutable_object(
              py::handle(PyTuple_GetItem(example_value.ptr(), i)))) {
        return false;
      }
    }
    return true;
  }

  return PyLong_Check(example_value.ptr()) ||
      PyFloat_Check(example_value.ptr()) || PyBool_Check(example_value.ptr()) ||
      PyUnicode_Check(example_value.ptr()) ||
      (is_tensor_immutable && THPVariable_Check(example_value.ptr()));
}

bool is_parameter(py::handle tensor) {
  py::object parameter = py::module::import("torch.nn").attr("Parameter");
  return py::isinstance(tensor, parameter);
}

/**
 * Dispatches metadata functions to the methods that return integer values,
 * i.e. used whenever static shapes are being used.
 *
 * These are used by the tensor storage overlapping check. Even though their
 * symbolic counterpart does work whenever static shapes are being used, the
 * introduced overhead might significantly worsen the performance.
 */
struct StaticMeta {
  static int64_t numel(const Tensor& t) {
    return t.numel();
  }

  static int64_t storage_offset(const Tensor& t) {
    return t.storage_offset();
  }

  static int64_t size(const Tensor& t, int64_t i) {
    return t.size(i);
  }

  static int64_t stride(const Tensor& t, int64_t i) {
    return t.stride(i);
  }
};

/**
 * Dispatches metadata functions to the methods that return c10::SymInt
 * values, i.e. used whenever dynamic shapes are being used.
 */
struct DynamicMeta {
  static SymInt numel(const Tensor& t) {
    return t.sym_numel();
  }

  static SymInt storage_offset(const Tensor& t) {
    return t.sym_storage_offset();
  }

  static SymInt size(const Tensor& t, int64_t i) {
    return t.sym_size(i);
  }

  static SymInt stride(const Tensor& t, int64_t i) {
    return t.sym_stride(i);
  }
};

/**
 * Assumption: x and y are known to share a storage, and we are trying to
 * determine if their memory is actually completely disjoint, based on
 * sizes/strides/storage_offset
 *
 * "Meta" should be one of the "*Meta" classes above. They dictate which
 * version of the metadata functions we should be using (symbolic vs.
 * concrete). Even though they have the same apparent behavior, the symbolic
 * version introduces a bit of overhead. Such an overhead might end up
 * becoming relevant if it's run enough times.
 */
template <class Meta>
bool tensors_definitely_do_not_overlap(const Tensor& x, const Tensor& y) {
  if (x.is_same(y)) {
    return false;
  }
  if (Meta::numel(x) == 0 || Meta::numel(y) == 0) {
    return true;
  }

  // Make x always on the left
  if (Meta::storage_offset(x) > Meta::storage_offset(y)) {
    return tensors_definitely_do_not_overlap<Meta>(y, x);
  }

  // Short-circuit in the "obvious" overlapping case: both tensors are
  // contiguous
  if (x.is_contiguous() && y.is_contiguous()) {
    if (Meta::storage_offset(x) + Meta::numel(x) > Meta::storage_offset(y)) {
      // definitely overlap
      return false;
    } else {
      // definitely no overlap
      return true;
    }
  }

  // Short-circuit: if last memory address of x is < start of y, then not
  // overlapping.
  auto x_last = Meta::storage_offset(x);
  for (int64_t i = 0; i < x.dim(); i++) {
    x_last += (Meta::size(x, i) - 1) * Meta::stride(x, i);
  }
  if (x_last < Meta::storage_offset(y)) {
    return true;
  }

  if (x.dim() == 2 && y.dim() == 2 && Meta::stride(x, 1) == 1 &&
      Meta::stride(y, 1) == 1) {
    // This cases is needed for the shampoo optimizer.
    // All tensors are 2d (non-contiguous), have the same outer stride, and have
    // an inner stride of 1 (so rows are contiguous)
    if (Meta::stride(x, 0) == Meta::stride(y, 0)) {
      auto offset_delta = Meta::storage_offset(y) - Meta::storage_offset(x);
      if (offset_delta < Meta::size(x, 1)) {
        // definitely overlaps (row 0 of y overlaps with row 0 of x)
        // Example:
        //   base = torch.arange(32).reshape(4, 8)
        //   x = base.narrow(1, 0, 4)
        //     x: size=(4, 4), stride=(8, 1), offset=0
        //   y = base.narrow(1, 3, 4)
        //     y: size=(4, 4), stride=(8, 1), offset=3
        return false;
      }
      auto x_total_elems_covered =
          Meta::stride(x, 0) * (Meta::size(x, 0) - 1) + Meta::size(x, 1);
      if (x_total_elems_covered <= offset_delta) {
        // definitely does not overlap (last byte of x is before start of y)
        // Example:
        //   x: size=(4, 4), stride=(8, 1), offset=0 (last byte is 27)
        //   y: size=(4, 4), stride=(8, 1), offset=28 (start byte is 28)
        return true;
      }
      // At this point, we want to check if the 0th row of y
      // overlaps with **some** row of x.
      // We can check this by shifting y backward by the shared stride,
      // repeatedly, until the first row of y is before the first row of x. Then
      // we can check if these rows overlap. We can accomplish this by modding
      // our offset by the stride.
      auto offset_delta_mod = offset_delta % Meta::stride(x, 0);
      // Example:
      // 0 1 2 3
      // 9 10 11 12
      // 18 19 20 21
      // 27 28 29 30
      //   x: size=(4, 4), stride=(9, 1), offset=0
      //   y: size=(4, 4), stride=(9, 1), offset=22 (this would not overlap)
      //   y: size=(4, 4), stride=(9, 1), offset=23 (this would not overlap)
      //   y: size=(4, 4), stride=(9, 1), offset=24 (this would overlap)
      //   y: size=(4, 4), stride=(9, 1), offset=25 (this would overlap)
      // If the interval [modded_offset, modded_offset + x_size] falls entirely
      // without
      if (offset_delta_mod + Meta::size(y, 1) <= Meta::stride(x, 0)) {
        return true;
      }
    }
  }
  return false;
}

/**
 * Computes the indices of the tensors that might overlap.
 *
 * Checks which of the given tensors have overlapping storages with ANY of
 * the other tensors.
 *
 * So, for example, if tensor 1 overlaps with tensor 2, and tensor 3 with
 * tensor 4, all of them will be in the output of this function. Even if
 * tensor 1 and 4 don't overlap.
 */
template <class Meta>
std::unordered_set<int64_t> compute_overlapping_tensors(
    const std::vector<Tensor>& tensors) {
  std::unordered_set<int64_t> aliased_tensor_indices;
  for (int64_t i = 0; i < static_cast<int64_t>(tensors.size()); i++) {
    const auto& tensor_i = tensors[i];
    for (int64_t j = 0; j < i; j++) {
      if (!tensors_definitely_do_not_overlap<Meta>(tensor_i, tensors[j])) {
        aliased_tensor_indices.insert(i);
        aliased_tensor_indices.insert(j);
      }
    }
  }
  return aliased_tensor_indices;
}

/**
 * Checks whether the storage overlapping relation is preserved.
 *
 * At this point, `non_overlapping` represents the tensors that should not
 * have overlapping storages. Similarly, `overlapping` represents the tensors
 * that should have overlapping storage in some way (or that we can't be sure).
 *
 * This function checks whether the assumption above is true or not.
 */
bool check_overlapping(
    const std::vector<Tensor>& overlapping,
    const std::vector<Tensor>& non_overlapping) {
  // Merge the tensor lists.
  std::vector<Tensor> tensors;
  tensors.reserve(overlapping.size() + non_overlapping.size());
  tensors.insert(tensors.end(), overlapping.begin(), overlapping.end());
  tensors.insert(tensors.end(), non_overlapping.begin(), non_overlapping.end());
  // Check what is the current storage overlapping relation.
  auto indices = compute_overlapping_tensors<StaticMeta>(tensors);
  // Check that the set of indices of tensors that might overlap is equal to
  // the indices of the first `overlapping.size()` tensors. That's because
  // `overlapping` tensors were in the beginning of `tensors` list.
  auto range = c10::irange(overlapping.size());
  return indices.size() == overlapping.size() &&
      std::all_of(range.begin(), range.end(), [&](int64_t i) {
           return indices.count(i) == 1;
         });
}

/**
 * Class responsible for collecting and checking the storage overlap relations.
 *
 * The way GuardManager is implemented, when STORAGE_OVERLAPPING guard check is
 * run on a given tensor, we don't know if it is an overlapping or
 * non-overlapping tensor. There's no order to which GuardManager runs the guard
 * check so that we can split it in 2.
 *
 * Since we are only interested in the classification of each tensor (not
 * necessarily the order), we can just issue 2 STORAGE_OVERLAPPING guards
 * representing the overlapping tensors and the non-overlapping ones.
 *
 * In order to collect the information from both guards (so that we can call
 * `check_overlapping` function correctly), we need this class which stores
 * both kinds of tensors, and knows when it has collected each one of them.
 */
class StorageOverlapChecker {
 public:
  StorageOverlapChecker(
      size_t expected_overlapping,
      size_t expected_non_overlapping)
      : _expected_overlapping(expected_overlapping),
        _expected_non_overlapping(expected_non_overlapping) {}

  /**
   * Adds a tensor to the corresponding storage, based on whether it should be
   * an `overlapping` tensor or not.
   */
  void add(PyObject* obj, bool overlapping) {
    // Just check that `obj` is actually a tensor, so that we can keep it alive
    // by incrementing its ref-count.
    TORCH_CHECK(THPVariable_CheckExact(obj) || THPVariable_Check(obj));
    Py_INCREF(obj);
    _get(overlapping).push_back(obj);
  }

  void reset(bool overlapping) {
    auto& vec = _get(overlapping);
    for (auto item : vec) {
      Py_DECREF(item);
    }
    vec.clear();
  }

  /**
   * Maybe checks the storage overlapping relation.
   *
   * Before actually calling `check_overlapping` function, this function makes
   * sure it has collected all expected tensors.
   */
  bool maybe_check() {
    TORCH_CHECK(_expected_overlapping >= _overlapping.size());
    TORCH_CHECK(_expected_non_overlapping >= _non_overlapping.size());
    if (_expected_overlapping == _overlapping.size() &&
        _expected_non_overlapping == _non_overlapping.size()) {
      // Transform each list of PyObject* into an actual list of Tensors.
      auto overlapping_tensors =
          _tensors_from(_overlapping, _expected_overlapping);
      auto non_overlapping_tensors =
          _tensors_from(_non_overlapping, _expected_non_overlapping);
      return check_overlapping(overlapping_tensors, non_overlapping_tensors);
    } else {
      // If we haven't collected them all yet, keep on running.
      return true;
    }
  }

 private:
  /**
   * Returns a reference to the container that corresponds to the given
   * overlapping relation.
   */
  std::vector<PyObject*>& _get(bool overlapping) {
    return overlapping ? _overlapping : _non_overlapping;
  }

  /**
   * Transforms a given list of PyObject* into a list of Tensor.
   */
  std::vector<Tensor> _tensors_from(
      const std::vector<PyObject*>& objects,
      size_t size) {
    std::vector<Tensor> tensors;
    tensors.reserve(size);
    std::transform(
        objects.begin(),
        objects.end(),
        std::back_inserter(tensors),
        [=](PyObject* obj) { return THPVariable_Unpack(obj); });
    return tensors;
  }

  // Expected number of possibly overlapping tensors.
  size_t _expected_overlapping;
  // Expected number of non-overlapping tensors.
  size_t _expected_non_overlapping;
  // Collected possibly overlapping tensors.
  std::vector<PyObject*> _overlapping;
  // Collected non-overlapping tensors.
  std::vector<PyObject*> _non_overlapping;
};

/**
 * Stores relevant guard debug information, e.g., failure str for a LeafGuard
 * failure. The data structure is also accessible in Python.
 */

class GuardDebugInfo {
 public:
  GuardDebugInfo(
      bool result,
      py::list verbose_code_parts,
      int num_guards_executed)
      : result(result),
        verbose_code_parts(std::move(verbose_code_parts)),
        num_guards_executed(num_guards_executed) {}

  // This constructor is used when guard succeeds.
  GuardDebugInfo(bool result, int num_guards_executed)
      : result(result), num_guards_executed(num_guards_executed) {}

  GuardDebugInfo(
      bool result,
      const std::string& failed_reason,
      int num_guards_executed)
      : GuardDebugInfo(result, num_guards_executed) {
    verbose_code_parts.append(failed_reason);
  }

  std::string to_string() {
    std::stringstream ss;
    ss << "GuardDebugInfo(\n"
       << "result=" << result << ",\n"
       << "verbose_code_parts=" << verbose_code_parts << ",\n"
       << "num_guards_executed=" << num_guards_executed << ")\n";
    return ss.str();
  }

  // Whether the guard passed or failed.
  bool result;

  // This is a list of verbose_code_parts for the failed guard. When there are
  // more than one verbose_code_parts, then recompilation reasoning infra on the
  // Python side can iterate over this list and eval each string to pinpoint the
  // exact code part that failed.
  py::list verbose_code_parts;

  // Total number of executed guards so far. This is helpful in debugging if
  // shuffling is working.
  int num_guards_executed;
};

class GuardManager;
class RootGuardManager;
class DictGuardManager;

/**
 * Base class for the leaf guard in the GuardManager hierarchy.
 */
class LeafGuard {
 public:
  // Most guards do not need root guard manager.
  LeafGuard(py::object verbose_code_parts)
      : _verbose_code_parts(std::move(verbose_code_parts)) {}

  // Guards like TENSOR_MATCH require root_guard_manager to access local_state
  // shared across all leaf guards.
  LeafGuard(RootGuardManager* root_guard_manager, py::object verbose_code_parts)
      : _root_guard_manager(root_guard_manager),
        _verbose_code_parts(std::move(verbose_code_parts)) {}

  // check function could be called from python. This is useful for debugging
  // purpose.
  bool check(py::handle value) {
    return check_nopybind(value.ptr());
  }

  GuardDebugInfo check_verbose(py::handle value) {
    return check_verbose_nopybind(value.ptr());
  }

  virtual GuardDebugInfo check_verbose_nopybind(
      PyObject* value) { // borrowed ref
    bool result = check_nopybind(value);
    if (!result) {
      return GuardDebugInfo(result, _verbose_code_parts, 0);
    }
    return GuardDebugInfo(true, 0);
  }

  py::list verbose_code_parts() {
    return _verbose_code_parts;
  }

  // This is on the hot path and avoids any refcounting code from pybind. This
  // is not exposed to Python and can only be called from C++.
  virtual bool check_nopybind(PyObject* value) = 0;
  virtual bool check_nopybind(FrameLocalsMapping* map) {
    // throw std::runtime_error("fallback to python");
    // Could fallback to running check on the Python dict (lazily constructed)
    return check_nopybind((PyObject*)map->to_dict());
  }

  virtual ~LeafGuard() = default;

 protected:
  // RootGuardManager has state that is common across all guards like
  // LocalState.
  RootGuardManager* _root_guard_manager{nullptr};

 private:
  // This is set while constructing the leaf guard. This is used for identifying
  // the cause of recompilation.
  py::list _verbose_code_parts;
};

/**
 * Represents a leaf guard that accepts the python guard check function. We
 * would like to have most of the guards in C++ (to avoid a Python function
 * call).  But, it will take some time to reach that goal. Also, there might be
 * cases where its too tedious to write an equivalent C++ guard.
 *
 * LAMBDA_GUARD allows us to gradually move to C++. We can start from all
 * guards of type PythonLambaGuard and incrementally move expensive guards to
 * C++.
 */
class LAMBDA_GUARD : public LeafGuard {
 public:
  LAMBDA_GUARD(py::object guard_check_fn, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    if (py::isinstance<py::function>(guard_check_fn)) {
      _guard_check_fn = py::cast<py::function>(std::move(guard_check_fn));
    } else {
      throw py::type_error("LAMBDA_GUARD expects (callable, str)");
    }
  }

  // Runs the lambda function with the current f_locals value.
  bool check_nopybind(PyObject* value) override { // borrowed ref
    PyObject* x = PyObject_CallOneArg(_guard_check_fn.ptr(), value); // new ref
    if (x == nullptr) {
      // An exception is caught in the lambda function.
      PyErr_Clear();
      return false;
    }
    bool result = PyObject_IsTrue(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(PyObject* value) override {
    PyObject* x = PyObject_CallOneArg(_guard_check_fn.ptr(), value); // new ref
    if (x == nullptr) {
      // An exception is caught in the lambda function.
      std::string exc_message = get_exception_message();
      PyErr_Clear();
      return GuardDebugInfo(false, exc_message, 0);
    }
    bool result = PyObject_IsTrue(x);
    Py_DECREF(x);
    if (result) {
      return GuardDebugInfo(true, 0);
    }
    return GuardDebugInfo(false, verbose_code_parts(), 0);
  }

 private:
  // The user provided lambda function for check_fn.
  py::function _guard_check_fn;
};

class TYPE_MATCH : public LeafGuard {
 public:
  // type_id = id(type(obj))
  TYPE_MATCH(py::object type_id, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _expected(py::cast<intptr_t>(std::move(type_id))) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // NOLINTNEXTLINE(performance-no-int-to-ptr)
    return Py_TYPE(value) == (void*)_expected;
  }

 private:
  // id of the type of the original object.
  intptr_t _expected;
};

class ID_MATCH : public LeafGuard {
 public:
  // obj_id = id(obj)
  ID_MATCH(py::object obj_id, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _expected(py::cast<intptr_t>(std::move(obj_id))) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // NOLINTNEXTLINE(performance-no-int-to-ptr)
    return value == (void*)_expected;
  }

 private:
  // id of the original object.
  intptr_t _expected;
};

class NONE_MATCH : public LeafGuard {
 public:
  NONE_MATCH(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return value == Py_None;
  }
};

class TRUE_MATCH : public LeafGuard {
 public:
  TRUE_MATCH(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return value == Py_True;
  }
};

class FALSE_MATCH : public LeafGuard {
 public:
  FALSE_MATCH(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return value == Py_False;
  }
};

class EQUALS_MATCH : public LeafGuard {
 public:
  EQUALS_MATCH(py::object value, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _value(value),
        _value_type(Py_TYPE(value.ptr())) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // Fast path - pointer equality check. Pointer equality checks are ok
    // because objects guarded with EQUALS_MATCH are immutable.
    if (value != _value.ptr()) {
      // Check type
      if (Py_TYPE(value) != _value_type) {
        return false;
      }
      int result = PyObject_RichCompareBool(value, _value.ptr(), Py_EQ);
      // Check for exception
      if (result == -1) {
        PyErr_Clear();
        return false;
      }
      return result;
    }
    return true;
  }

 private:
  // value to compare against. This is py::object so that we hold on to the
  // original value and prevent garbage collection. We run EQUALS_MATCH only on
  // selected objects which do not have high memory footprint, so holding on to
  // these objects is ok.
  py::object _value;

  // Type of the value
  PyTypeObject* _value_type;
};

class RANGE_ITERATOR_MATCH : public LeafGuard {
 public:
  RANGE_ITERATOR_MATCH(
      py::object start,
      py::object stop,
      py::object step,
      py::object type_id,
      py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _type_id(py::cast<intptr_t>(std::move(type_id))) {
    PyObject* start_obj = start.ptr();
    PyObject* stop_obj = stop.ptr();
    PyObject* step_obj = step.ptr();
    _start = THPUtils_unpackLong(start_obj);
    _stop = THPUtils_unpackLong(stop_obj);
    _step = THPUtils_unpackLong(step_obj);
    TORCH_CHECK(
        !PyErr_Occurred(), "values of start/stop/step must fit in a long type");
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // Do a type match first.
    // NOLINTNEXTLINE(performance-no-int-to-ptr)
    if (Py_TYPE(value) != (void*)_type_id) {
      return false;
    }
    _PyRangeIterObject* iter = (_PyRangeIterObject*)value;

#if IS_PYTHON_3_12_PLUS
    long start = iter->start;
#else
    long start = iter->start + iter->index * iter->step;
#endif // IS_PYTHON_3_12_PLUS

    long stop = iter->start + iter->len * iter->step;
    return start == _start && stop == _stop && iter->step == _step;
  }

 private:
  intptr_t _type_id;
  // Normalized representation of a range iterator.
  long _start;
  long _stop;
  long _step;
};

class TUPLE_ITERATOR_LEN : public LeafGuard {
 public:
  TUPLE_ITERATOR_LEN(
      py::object length,
      py::object type_id,
      py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _length(py::cast<Py_ssize_t>(std::move(length))),
        _type_id(py::cast<intptr_t>(std::move(type_id))) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // Do a type match first.
    // NOLINTNEXTLINE(performance-no-int-to-ptr)
    if (Py_TYPE(value) != (void*)_type_id) {
      return false;
    }
    _PyTupleIterObject* it = (_PyTupleIterObject*)value;
    Py_ssize_t length = 0;
    if (it->it_seq)
      length = PyTuple_GET_SIZE(it->it_seq) - it->it_index;
    return length == _length;
  }

 private:
  // Length of the guarded list
  Py_ssize_t _length;
  intptr_t _type_id;
};

class LENGTH_CHECK : public LeafGuard {
 public:
  LENGTH_CHECK(py::object value, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _length(py::cast<Py_ssize_t>(std::move(value))) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // PySequence_Length returns -1 if the object is not a sequence. So, we
    // don't have to test for PySequence_Check.
    return PySequence_Length(value) == _length;
  }

 private:
  // Length of the guarded list
  Py_ssize_t _length;
};

class DICT_LENGTH : public LeafGuard {
 public:
  DICT_LENGTH(py::object value, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _length(py::cast<Py_ssize_t>(std::move(value))) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return PyDict_Check(value) && PyDict_Size(value) == _length;
  }

 private:
  // Length of the guarded dict
  Py_ssize_t _length;
};

class NOT_NONE : public LeafGuard {
 public:
  NOT_NONE(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return value != Py_None;
  }
};

class MAPPING_KEYS_MATCH : public LeafGuard {
 public:
  MAPPING_KEYS_MATCH(py::object value, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    // This is ok to stash in the state because we only support
    // MappingProxyType objects with constant keys. So, the mem overhead is
    // negligible.
    _keys = py::list(value.attr("keys")());
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    PyObject* keys = PyMapping_Keys(value); // new ref
    int result = PyObject_RichCompareBool(keys, _keys.ptr(), Py_EQ);
    Py_DECREF(keys);
    return result;
  }

 private:
  py::object _keys;
};

class DEFAULT_DEVICE : public LeafGuard {
 public:
  DEFAULT_DEVICE(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    py::handle device_module = py::module::import("torch.utils._device");
    // Save the dict using py::object
    _utils_device_dict = device_module.attr("__dict__");
    _device = _utils_device_dict["CURRENT_DEVICE"];
  }

  template <typename T>
  bool check_nopybind_template(T* value) { // borrowed ref
    // Create a static interned string. Interned string is faster than creating
    // a new string every time. Even though its a new reference, we don't dec
    // ref it. Interned strings are used for things like variable names and are
    // leaked by design.
    static PyObject* current_device_str =
        PyUnicode_InternFromString("CURRENT_DEVICE");
    PyObject* device = PyDict_GetItem(
        _utils_device_dict.ptr(), current_device_str); // borrowed ref
    if (device != _device.ptr()) {
      int result = PyObject_RichCompareBool(device, _device.ptr(), Py_EQ);
      if (result == -1) {
        PyErr_Clear();
        return false;
      }
      return result;
    }
    return true;
  }

  bool check_nopybind(PyObject* value) override {
    return check_nopybind_template(value);
  }

  bool check_nopybind(FrameLocalsMapping* value) override {
    return check_nopybind_template(value);
  }

 private:
  // Save the current device and the module dict during the guard construction.
  py::object _utils_device_dict;
  py::object _device;
};

class GLOBAL_STATE : public LeafGuard {
 public:
  GLOBAL_STATE(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    _guard = std::make_unique<GlobalStateGuard>();
    _guard->init();
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // Ignore value arg, this is just to satisfy the interface.
    return _guard->check();
  }

  bool check_nopybind(FrameLocalsMapping* value) override {
    // Ignore value arg, this is just to satisfy the interface.
    return _guard->check();
  }

  GuardDebugInfo check_verbose_nopybind(PyObject* value) override {
    if (!_guard->check()) {
      return GuardDebugInfo(
          false, "GLOBAL_STATE changed: " + _guard->reason(), 0);
    }
    return GuardDebugInfo(true, 1);
  }

 private:
  std::unique_ptr<GlobalStateGuard> _guard;
};

// Checks that an attr is absent in the object. We don't need the opposite
// HASATTR guard because we can just rely on GetAttrGuardAccessor to act as
// HASATTR guard.
class NO_HASATTR : public LeafGuard {
 public:
  NO_HASATTR(py::object attr_name, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _attr_name(std::move(attr_name)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    return PyObject_HasAttr(value, _attr_name.ptr()) == 0;
  }

 private:
  py::object _attr_name;
};

// Checks that dict contains or does not contain a key. This happens for
// PythonSysModulesVariable tracker.
// TODO(janimesh) - Check if we can use DictGuardManager. The downside could be
// large number of keys for sys module, so DICT_CONTAINS might still end up
// being faster.
class DICT_CONTAINS : public LeafGuard {
 public:
  DICT_CONTAINS(bool contains, py::object key, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _contains(contains ? 1 : 0),
        _key(std::move(key)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    int result = PyDict_Check(value) && PyDict_Contains(value, _key.ptr());
    if (result == -1) {
      PyErr_Clear();
      return false;
    }
    return result == _contains;
  }

 private:
  int _contains;
  py::object _key;
};

/**
 * Relational guards compare more than one value. We implement Relational
 * guards by capturing some state in the guard object. For example for tensor
 * aliasing guards - tensor X is not tensor Y - we construct one leaf guard
 * and and install it at as a leaf of two guard managers (one for X and
 * another for Y). Therefore, this guard is run twice. In the first
 * invocation, it saves the first value (state) and returns True. In the
 * second invocation, it compares the saved value with the new value and
 * returns True if they do not alias.
 *
 * We have to be careful about resetting in case the other guards fail and we
 * have some state in the relational guard. This is done by virtual method
 * reset_state(). This is called by the RootGuardManager before it exits.
 *
 */
class RelationalGuard : public LeafGuard {
 public:
  RelationalGuard(py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {}

  // reset the relational guard state on guard failure. This is called by the
  // guard manager.
  virtual void reset_state() = 0;
};

/**
 * Checks that object x is object y.
 */
class OBJECT_ALIASING : public RelationalGuard {
 public:
  OBJECT_ALIASING(py::object verbose_code_parts)
      : RelationalGuard(std::move(verbose_code_parts)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    if (_is_first_call) {
      _first_tensor = value;
      _is_first_call = false;
      return true;
    }
    return _first_tensor == value;
  }

  void reset_state() final {
    _is_first_call = true;
  }

 private:
  bool _is_first_call{true};
  PyObject* _first_tensor{nullptr};
};

/**
 * Checks that none of the tensors alias.
 */
class NO_TENSOR_ALIASING : public RelationalGuard {
 public:
  NO_TENSOR_ALIASING(
      const py::list& tensor_names,
      py::object verbose_code_parts)
      : RelationalGuard(std::move(verbose_code_parts)),
        _tensor_names(tensor_names) {
    _unique_tensors.reserve(tensor_names.size());
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    auto insertion = _unique_tensors.insert({value, nullptr});
    if (!insertion.second) {
      // No need to clear _unique_tensors, reset_state will do
      // it.
      return false;
    }
    return true;
  }

  GuardDebugInfo check_verbose_nopybind(PyObject* value) override {
    bool result = check_nopybind(value);

    if (!result) {
      return GuardDebugInfo(
          false, "Duplicate tensor found where not expected!", 0);
    }
    return GuardDebugInfo(true, 1);
  }

  void reset_state() final {
    _unique_tensors.clear();
  }

 private:
  py::list _tensor_names;
  ska::flat_hash_map<PyObject*, std::nullptr_t> _unique_tensors;
};

/**
 * Checks the storage overlapping relation of input tensors.
 *
 * This guard is always installed in pairs: one for the possibly overlapping
 * tensors, and another one for the non-overlapping tensors. This is so we can
 * correctly identify the given tensor in the check method as one of the 2
 * classes mentioned above.
 *
 * In the end, the one responsible for storing and checking is the
 * `StorageOverlapChecker` class.
 */
class STORAGE_OVERLAPPING : public RelationalGuard {
 public:
  STORAGE_OVERLAPPING(
      bool overlapping,
      std::shared_ptr<StorageOverlapChecker> checker,
      py::object verbose_code_parts)
      : RelationalGuard(std::move(verbose_code_parts)),
        _overlapping(overlapping),
        _checker(std::move(checker)) {}

  bool check_nopybind(PyObject* value) override {
    _checker->add(value, _overlapping);
    return _checker->maybe_check();
  }

  void reset_state() final {
    _checker->reset(_overlapping);
  }

 private:
  // Flag that indicates which kind of tensor this guard is collecting:
  //   1. Possibly overlapping tensors; or
  //   2. Non-overlapping tensors.
  bool _overlapping;
  // Actual checker for this guard.
  std::shared_ptr<StorageOverlapChecker> _checker;
};

/**
 * Symbolic Shape Guard.
 */
class SYMBOLIC_SHAPE_GUARD : public RelationalGuard {
 public:
  SYMBOLIC_SHAPE_GUARD(
      py::int_ nargs_int,
      py::int_ nargs_float,
      py::int_ py_addr,
      py::object py_addr_keep_alive,
      py::object verbose_code_parts)
      : RelationalGuard(std::move(verbose_code_parts)),
        _py_addr_keep_alive(std::move(py_addr_keep_alive)) {
    _nargs_int = PyLong_AsSize_t(nargs_int.ptr());
    _nargs_float = PyLong_AsSize_t(nargs_float.ptr());
    _nargs = _nargs_int + _nargs_float;
    if (PyErr_Occurred()) {
      throw py::value_error(
          "SYMBOLIC_SHAPE_GUARD expected a non-negative number of arguments.");
    }
    uintptr_t addr = PyLong_AsUnsignedLongLong(py_addr.ptr());
    if (PyErr_Occurred()) {
      throw py::value_error(
          "SYMBOLIC_SHAPE_GUARD expected an address to a C function.");
    }
    _guard_check_fn = reinterpret_cast<int8_t (*)(int64_t*, double*)>(addr);
    _args_int = std::vector<int64_t>(_nargs_int);
    _args_float = std::vector<double>(_nargs_float);
  }

  bool check_nopybind(PyObject* value) override {
    // We know that these arguments came from
    // IndexedSource(TensorPropertyGuard) and therefore no need to check that
    // the value is a Tuple[int, int].
    PyObject* py_idx = PyTuple_GET_ITEM(value, 0);
    PyObject* py_val = PyTuple_GET_ITEM(value, 1);
    size_t iarg = PyLong_AsSize_t(py_idx);
    if (iarg < _nargs_int) {
      if (!PyLong_Check(py_val)) {
        return false;
      }
      _args_int[iarg] = PyLong_AsLongLong(py_val);
    } else {
      if (!PyFloat_Check(py_val)) {
        return false;
      }
      _args_float[iarg - _nargs_int] = PyFloat_AS_DOUBLE(py_val);
    }
    _args_seen++;

    if (_args_seen == _nargs) {
      _args_seen = 0;
      return _guard_check_fn(_args_int.data(), _args_float.data());
    } else {
      // We don't have all the values yet. Return true until we get all.
      return true;
    }
  }

  GuardDebugInfo check_verbose_nopybind(PyObject* value) override {
    if (!PyTuple_Check(value)) {
      return GuardDebugInfo(false, "Non tuple found!", 0);
    } else if (PyTuple_Size(value) != 2) {
      return GuardDebugInfo(false, "Tuple of size not 2 found!", 0);
    } else {
      PyObject* py_idx = PyTuple_GET_ITEM(value, 0);
      PyObject* py_val = PyTuple_GET_ITEM(value, 1);
      if (!PyLong_Check(py_idx)) {
        return GuardDebugInfo(false, "Non integer index found!", 0);
      }
      size_t iarg = PyLong_AsSize_t(py_idx);
      if (iarg >= _nargs) {
        return GuardDebugInfo(false, "Index out of bounds!", 0);
      } else if (iarg < _nargs_int && !PyLong_Check(py_val)) {
        return GuardDebugInfo(false, "Non integer found!", 0);
      } else if (iarg >= _nargs_int && !PyFloat_Check(py_val)) {
        return GuardDebugInfo(false, "Non float found!", 0);
      }
    }
    bool result = check_nopybind(value);

    if (!result) {
      std::string msg = "\"Shape guard failed with values: ";
      for (auto v : _args_int) {
        msg += std::to_string(v) + ",";
      }
      for (auto v : _args_float) {
        msg += std::to_string(v) + ",";
      }
      msg.pop_back();
      msg += "\"";
      auto msgs = py::list();
      for (auto code_part : verbose_code_parts()) {
        msgs.append(code_part);
      }
      msgs.append(msg);
      return GuardDebugInfo(false, msgs, 0);
    }
    return GuardDebugInfo(true, 1);
  }

  void reset_state() final {
    _args_seen = 0;
  }

 private:
  py::object _py_addr_keep_alive;
  size_t _args_seen{0}, _nargs_float, _nargs_int, _nargs;
  std::vector<int64_t> _args_int;
  std::vector<double> _args_float;
  std::function<int8_t(int64_t*, double*)> _guard_check_fn;
};

class DYNAMIC_INDICES : public LeafGuard {
  // C++ equivalent of
  //  code.append(
  //      f"(({tensor_name}._dynamo_dynamic_indices.issubset({value._dynamo_dynamic_indices}))
  //      if hasattr({tensor_name}, '_dynamo_dynamic_indices') else True)"  #
  //      noqa: B950
  //  )
 public:
  DYNAMIC_INDICES(py::set dynamic_indices, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)),
        _dynamic_indices(std::move(dynamic_indices)) {}

  bool check_nopybind(PyObject* value) override { // borrowed ref
    // Make an interned string
    static PyObject* dynamic_indices_str =
        PyUnicode_InternFromString("_dynamo_dynamic_indices");
    PyObject* indices = PyObject_GetAttr(value, dynamic_indices_str); // new ref
    if (indices == nullptr) {
      // Attr absent. Clear exception.
      PyErr_Clear();
      // This is true deliberately. If hasattr fails, we return true.
      return true;
    }

    static PyObject* issubset_str = PyUnicode_InternFromString("issubset");
    PyObject* call_result = PyObject_CallMethodObjArgs(
        indices, issubset_str, _dynamic_indices.ptr(), nullptr); // new ref
    bool result = PyObject_IsTrue(call_result);
    Py_DECREF(call_result);
    Py_DECREF(indices);
    return result;
  }

 private:
  py::set _dynamic_indices;
};

class DICT_VERSION : public LeafGuard {
 public:
  DICT_VERSION(py::object value, py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    if (!PyDict_Check(value.ptr())) {
      throw py::type_error("DICT_VERSION expects a dict");
    }
    _tag = get_dict_version_unchecked(value.ptr());
  }
  bool check_nopybind(PyObject* value) override { // borrowed ref
    return PyDict_Check(value) && get_dict_version_unchecked(value) == _tag;
  }

  // Saved dict version.
  uint64_t _tag;
};

// GuardManager can be a pointer to DictGuardManager, but at this point the
// compiler does not know that DictGuardManager is a derived class of
// GuardManager (no way to define inheritance relationships in forward
// declarations), so we forward declare a factory function and define it when
// both DictGuardManager and GuardManager are fully defined.
std::unique_ptr<GuardManager> make_guard_manager(
    RootGuardManager* root,
    std::string source,
    py::handle example_value,
    py::handle guard_manager_enum);

GuardManager* clone_guard_manager(
    GuardManager* from,
    RootGuardManager* root,
    const py::function& clone_filter_fn);
void add_relational_guard_resetter_to_cloned_root(
    RootGuardManager* root,
    std::shared_ptr<RelationalGuard> guard);

/**
 * Base class representing a pair of accessor and the associated guard
 * manager. The accessor defines how to access the child value from the
 * py::object given to the parent check function.
 *
 * GuardAccessors can be considered equivalent to name() method of Source
 * objects in guards.py. In python, name() method returns a str which we can
 * then eval in f_locals and f_globals to retrieve the actual py object.
 * GuardAccessor serves the same purpose. The minor difference is that
 * GuardManager is a tree structure, so a GuardAccessor just has to retrieve
 * the value in the next level in this tree and pass it to the child
 * GuardAccessor.
 *
 * GuardAccessor also owns the GuardManager associated with the retrieved
 * value from the GuardAccessor.
 */
class GuardAccessor {
 public:
  GuardAccessor(
      RootGuardManager* root,
      py::object accessor_key,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum);

  // Return by reference as GuardAccessor owns the GuardManager.
  std::unique_ptr<GuardManager>& get_guard_manager() {
    return _guard_manager;
  }

  bool matches_key(const py::handle& key) const {
    return _accessor_key.equal(key);
  }

  std::string get_source() {
    return _source;
  }

  // matches_dict_tag is used by the DictGetItemGuardAccessor to skip the guard
  // subtree on immutable dict getitems.
  virtual bool check_nopybind(PyObject* obj, bool matches_dict_tag = false) = 0;
  virtual bool check_nopybind(FrameLocalsMapping* map, bool matches_dict_tag) {
    // throw std::runtime_error("fallback to python");
    // Could fallback to running check on the Python dict (lazily constructed)
    return check_nopybind((PyObject*)map->to_dict(), matches_dict_tag);
  }
  virtual GuardDebugInfo check_verbose_nopybind(PyObject* obj) = 0;
  virtual std::string repr() const = 0;

  virtual ~GuardAccessor() = default;

  // Cloning related functions
  GuardAccessor(GuardManager* guard_manager, GuardAccessor* from);

  virtual GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) = 0;

  void clone_visitor(GuardAccessor* to) {
    to->_source = this->_source;
    to->_accessor_key = this->_accessor_key;
  }

  template <typename DerivedGuardAccessor>
  GuardAccessor* clone_common(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) {
    GuardManager* cloned_mgr = clone_guard_manager(
        get_guard_manager().get(), cloned_root, clone_filter_fn);
    if (cloned_mgr == nullptr) {
      return nullptr;
    }
    DerivedGuardAccessor* cloned_accessor =
        new DerivedGuardAccessor(cloned_mgr, (DerivedGuardAccessor*)this);
    return cloned_accessor;
  }

 protected:
  // Guard manager corresponding to the retrieved value from the
  // GuardAccessor.
  std::unique_ptr<GuardManager> _guard_manager;
  // accessor key could be py::str for getattr, getitem or py::function for
  // lambda accessor. It is a py::object because we need to keep these accessor
  // keys alive.
  py::object _accessor_key;

  // A string that can be eval'd on f_locals or f_globals to access the variable
  // value. Only used for debugging.
  std::string _source;
};

/**
 * GuardManager encapsulates all the guards related to a particular
 * py::object. It is a tree structure and consists of 1) Leaf guards - Guards
 * that are run on the user given object 2) Accessors - Guard accessors (like
 * getattr, getitem) to access the next value in the tree hierarchy. Accessor
 * object also holds the child GuardManager.
 *
 * Lets look at an example to understand how it works.
 * class Pair:
 *     int x = 1;
 *     int y = 2;
 *
 * At compile time
 * >> guard_mananger = GuardManager()
 * >> guard_mananger.x.add_lambda_guard(
 *        lambda x: isinstance(x, Pair),
 *        lambda x: f"expected Pair, found {type(x)}"
 *    )
 * >> guard_mananger.x.add_lambda_guard(lambda x: x == 1, lambda x: f"found
 * {x}, expected 1")
 * >> guard_mananger.y.add_lambda_guard(lambda x: x == 2, lambda x: f"found
 * {x}, expected 2")
 *
 * At runtime
 * >> guard_mananger.check(Pair())
 *
 * At compile time we build the tree structure. When we do `guard_manager.x`,
 * it creates an AttrGuardAccessorNode, initializes a child guard manager with
 * this accessor node, and adds it as a child. When we do
 * `guard_manager.x.add_lambda_guard`, we call add_lambda_guard on the newly
 * created guard manager and register a new leaf guard on it.
 *
 * At runtime, the accessor node has an important function of providing a way
 * to access the value for the child guard. In the above example,
 * guard_manager.x adds an AttrGuardAccessorNode with attr_name x. When check
 * function is called, parent GuardManager calls getattr(value, "x") on its
 * value passed to the check function to call the check function of the child
 * guard manager.
 *
 * Performance optimization for fail fast - An optimization for runtime here is
 * to sort the execution of child guards depending on the failure count.  This
 * ensures that we run the guards that are more prone to fail statistically
 * first. This can improve the cache lookup time when we have multiple cache
 * entries.
 */

// NOLINTNEXTLINE(cppcoreguidelines-special-member-functions)
class GuardManager {
 public:
  GuardManager() = delete;
  GuardManager(RootGuardManager* root, std::string source)
      : _root(root), _source(std::move(source)), _is_dict(false) {}

  GuardManager(
      RootGuardManager* root,
      std::string source,
      py::handle example_value)
      : _root(root),
        _source(std::move(source)),
        _is_dict(py::isinstance<py::dict>(example_value)) {
    if (_is_dict) {
      _dict_tag = get_dict_version_unchecked(example_value.ptr());
    }
  }

  GuardManager(const GuardManager& m) = delete;
  GuardManager& operator=(const GuardManager&) = delete;
  virtual ~GuardManager() = default;

  RootGuardManager* get_root() {
    return _root;
  }

  std::string get_source() {
    return _source;
  }

  virtual void add_leaf_guard(std::shared_ptr<LeafGuard> leaf_guard) {
    _leaf_guards.emplace_back(std::move(leaf_guard));
  }

 public:
  // For cloning
  GuardManager(RootGuardManager* root, std::string source, bool is_dict)
      : _root(root), _source(std::move(source)), _is_dict(is_dict) {}

  void clone_common(
      RootGuardManager* cloned_root,
      GuardManager* cloned_mgr,
      const py::function& clone_filter_fn) {
    for (const auto& guard : _leaf_guards) {
      cloned_mgr->_leaf_guards.emplace_back(guard);
      if (std::shared_ptr<RelationalGuard> relational_guard =
              std::dynamic_pointer_cast<RelationalGuard>(guard)) {
        add_relational_guard_resetter_to_cloned_root(
            cloned_root, relational_guard);
      }
    }

    for (const auto& accessor : _accessors) {
      GuardAccessor* cloned_accessor =
          accessor->clone(cloned_root, clone_filter_fn);
      if (cloned_accessor != nullptr) {
        cloned_mgr->_accessors.emplace_back(
            std::unique_ptr<GuardAccessor>(cloned_accessor));
      }
    }
  }

  virtual GuardManager* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) {
    if (!py::cast<bool>(clone_filter_fn(this))) {
      return nullptr;
    }
    GuardManager* cloned_mgr = new GuardManager(cloned_root, _source, _is_dict);
    clone_common(cloned_root, cloned_mgr, clone_filter_fn);
    return cloned_mgr;
  }

  /**
   * Adds a new guard manager with appropriate Accessor. If the accessor is
   * already present, we just return the guard manager.
   */
  template <typename GuardAccessorT>
  GuardManager* get_child_manager(
      const py::object& accessor_key,
      const std::string& source,
      py::handle example_value,
      py::handle guard_manager_enum) {
    // accessor_key type depends on the GuardAccessorT
    // for example for GetAttrGuardAccessor - py::str name

    // Return the manager if the guard accessor exists
    for (const auto& accessor : _accessors) {
      if (accessor->matches_key(accessor_key) &&
          source == accessor->get_source()) {
        return accessor->get_guard_manager().get();
      }
    }

    // Construct a new guard accessor
    _accessors.emplace_back(std::make_unique<GuardAccessorT>(
        _root,
        std::move(accessor_key),
        source,
        example_value,
        guard_manager_enum));
    return _accessors.back()->get_guard_manager().get();
  }

  // Runs the leaf guards check and then child managers check function.
  //
  // NB: There is some code DUPLICATION between this and check_verbose
  // function. This is intentional. check function is in the hot path and is
  // kept very simple. The purpose of check_verbose function is to get guard
  // failure reasoning to understand recompilations. check_verbose function
  // does not change the state of the guard, e.g., it does not shuffle the
  // guards and does not change the fail count. For simplicity, we duplicate
  // the code here.
  template <typename T>
  bool check_nopybind_template(T* value) { // borrowed ref

    if (!this->check_leaf_guards_nopybind(value)) {
      return false;
    }

    return this->check_accessors_nopybind(value);
  }

  virtual bool check_nopybind(PyObject* value) {
    return check_nopybind_template(value);
  }

  virtual bool check_nopybind(FrameLocalsMapping* value) {
    return check_nopybind_template(value);
  }

  template <typename T>
  bool check_leaf_guards_nopybind(T* value) {
    // Iterate over leaf guards
    for (const auto& guard : _leaf_guards) {
      if (!guard->check_nopybind(value)) { // early exit
        _fail_count += 1;
        // no need of sorting, just return.
        return false;
      }
    }

    return true;
  }

  template <typename T>
  bool check_accessors_nopybind(T* value) {
    bool matches_dict_tag = false;
    uint64_t new_tag = 0;
    if constexpr (std::is_same_v<T, PyObject>) {
      if (_is_dict) {
        // Check if the dict tag matches. If it does, propagate to the child
        // accessors. This will pass to the child manager via
        // DictGetItemGuardManager.
        // Relational Guards need to keep state, so do not send matches_dict_tag
        // to avoid early exits when dict_tag matches and the object is
        // immutable.
        new_tag = get_dict_version_unchecked(value);
        matches_dict_tag = (new_tag == _dict_tag);
      }
    }

    // Iterate over accessors.
    bool result = true;
    bool failed_on_first = true;
    for (const auto& accessor : _accessors) {
      if (!accessor->check_nopybind(value, matches_dict_tag)) { // early exit
        _fail_count += 1;
        result = false;
        // need to sort, so break the loop.
        break;
      }
      failed_on_first = false;
    }

    // failed_on_first is just an optimization to avoid sorting if we are
    // failing on the first accessor itself. This is helpful when we have
    // already sorted the guards once, and dont need to sort again.
    if (!result && !failed_on_first) {
      // Inplace sort the child guards by fail count. This moves the guard
      // with higher fail count earlier in the queue, and enables fail fast
      // for the next check_verbose.

      // An alternate implementation was to use priority queue directly on
      // _accessors, but it was rejected because of the complexity of
      // popping and creating a new pq on each run_guards. Moreover, this sort
      // is happening on the unhappy path when check_verbose guard
      // fails. So, its probably ok.
      std::sort(
          _accessors.begin(),
          _accessors.end(),
          [](const std::unique_ptr<GuardAccessor>& a,
             const std::unique_ptr<GuardAccessor>& b) {
            return a->get_guard_manager()->fail_count() >
                b->get_guard_manager()->fail_count();
          });
    }

    if (_is_dict && result) {
      // If result is true, reset the _dict_tag. This is useful if there is a
      // mutation on the dict but it does not change the attr values (like
      // swapping).
      _dict_tag = new_tag;
    }

    return result;
  }

  // This function has some code duplication with function check. This is
  // deliberate to keep check function simple and fast.
  virtual GuardDebugInfo check_verbose_nopybind(
      PyObject* value) { // borrowed ref
    int num_guards_executed = 0;

    const GuardDebugInfo& debug_info =
        check_leaf_guards_verbose_nopybind(value, num_guards_executed);
    if (!debug_info.result) {
      return debug_info;
    }

    return check_accessors_verbose_nopybind(value, num_guards_executed);
  }

  GuardDebugInfo check_leaf_guards_verbose_nopybind(
      PyObject* value,
      int& num_guards_executed) {
    // Iterate over leaf guards
    for (const auto& guard : _leaf_guards) {
      const GuardDebugInfo& debug_info = guard->check_verbose_nopybind(value);
      num_guards_executed++;
      if (!debug_info.result) {
        return GuardDebugInfo(
            false, debug_info.verbose_code_parts, num_guards_executed);
      }
    }

    return GuardDebugInfo(true, num_guards_executed);
  }

  GuardDebugInfo check_accessors_verbose_nopybind(
      PyObject* value,
      int& num_guards_executed) {
    // Iterate over accessors
    for (const auto& accessor : _accessors) {
      const GuardDebugInfo& debug_info =
          accessor->check_verbose_nopybind(value);
      num_guards_executed += debug_info.num_guards_executed;
      if (!debug_info.result) {
        return GuardDebugInfo(
            false, debug_info.verbose_code_parts, num_guards_executed);
      }
    }

    return GuardDebugInfo(true, num_guards_executed);
  }

  bool has_no_accessors() {
    return _accessors.empty();
  }

  int64_t fail_count() const {
    return _fail_count;
  }

  // DEBUG function - Returning raw pointers because we can't return unique_ptr
  // and pybind does not accept a unique_ptr reference return type.
  virtual std::vector<GuardAccessor*> get_accessors() const {
    std::vector<GuardAccessor*> ret;
    ret.reserve(_accessors.size());
    for (const auto& accessor : _accessors) {
      ret.emplace_back(accessor.get());
    }
    return ret;
  }

  // DEBUG function - Returning raw pointers because we can't return unique_ptr
  // and pybind does not accept a unique_ptr reference return type.
  virtual std::vector<GuardManager*> get_child_managers() {
    std::vector<GuardManager*> ret;
    ret.reserve(_accessors.size());
    for (const auto& accessor : _accessors) {
      ret.emplace_back(accessor->get_guard_manager().get());
    }
    return ret;
  }

  // DEBUG function - Returning raw pointers because we can't return unique_ptr
  // and pybind does not accept a unique_ptr reference return type.
  std::vector<LeafGuard*> get_leaf_guards() const {
    std::vector<LeafGuard*> ret;
    ret.reserve(_leaf_guards.size());
    for (const auto& guard : _leaf_guards) {
      ret.push_back(guard.get());
    }
    return ret;
  }

  bool is_leaf_guard_present(const std::string& guard_name) {
    return _inserted_leaf_guards.find(guard_name) !=
        _inserted_leaf_guards.end();
  }

  void insert_leaf_guard(const std::string& guard_name) {
    _inserted_leaf_guards.insert(guard_name);
  }

  void add_permitted_leaf_guard(std::shared_ptr<LeafGuard> leaf_guard) {
    // Selectively called for permitted guards. This is used by DictGuardManager
    // which overrides the add_leaf_guard manager to throw runtime error.
    GuardManager::add_leaf_guard(std::move(leaf_guard));
  }

 protected:
  // Keeps a count of how many times this guard manager check function returns
  // False. This is used for sorting optimization.
  int64_t _fail_count{0};

 private:
  // Root of the guard manager, this is the used to install the relational
  // guard resetters.
  RootGuardManager* _root;

  // A string that can be used to eval on f_locals or f_globals to get the
  // value. This is used only to pass on debugging information.
  std::string _source;

  // A map of which leaf guards are inserted. This is to prevent duplicate
  // guards like TYPE_MATCH.
  std::unordered_set<std::string> _inserted_leaf_guards;

  // Leaf guards are the terminal guards on this object, e.g, type check on a
  // list. These guards have to be run before any children are run.
  //
  // These leaf guards are not shufflable. In almost all cases, these guards
  // will have an order, e,g., type(x) is int guard and x == 5 guard. We also
  // expect very few leaf guards per GuardManager node.
  //
  // NB: Why are leaf guards shared ptr? This is primarily to enable relational
  // guards like `tensor X is not tensor Y`. These guards require multiple
  // values. We handle it by creating one guard object that holds state and this
  // guard is installed in many guard managers, hence a shared ptr.
  std::vector<std::shared_ptr<LeafGuard>> _leaf_guards;

  // GuardAccessors nodes to access the child guards. These guards are
  // shufflable. On a guard failure, they are sorted based on their fail count
  // to enable fail fast for the next check.
  std::vector<std::unique_ptr<GuardAccessor>> _accessors;

  bool _is_dict;
  uint64_t _dict_tag{0};
};

GuardAccessor::GuardAccessor(
    RootGuardManager* root,
    py::object accessor_key,
    std::string source,
    py::handle example_value,
    py::handle guard_manager_enum)
    : _guard_manager(
          make_guard_manager(root, source, example_value, guard_manager_enum)),
      _accessor_key(std::move(accessor_key)),
      _source(std::move(source)) {}

// Cloning related functions
GuardAccessor::GuardAccessor(GuardManager* guard_manager, GuardAccessor* from)
    : _guard_manager(std::unique_ptr<GuardManager>(guard_manager)) {
  from->clone_visitor(this);
}

/**
 Note on [Ownership with cloning] - GuardManagers have the facility to clone
 itself. This is useful for cloning a subset of the guard manager in diff guard
 manager.

 As the ownership goes, the model is exactly same as before. We have unique_ptr
 for GuardAccessor and GuardManagers. So, any state required for the accessors
 and managers is copied over using constructors and clone_visitor functions.
 The main thing to notice is leaf guards. The leaf guards are represented using
 shared_ptr, and they are shared (not cloned) with the cloned managers.

 So for leaf guard state to be released, both the original and cloned managers
 have to be destructed.
*/

/**
 * RootGuardManager is the root of the guard tree. This is primarily
 * constructed to hold the relational guard pointers so that we can reset the
 * state of those guards on guard failure. All the other important
 * implementation is in GuardManager class.
 */

class RootGuardManager : public GuardManager {
 public:
  // This is the root node, set its _root member to nullptr
  RootGuardManager() : GuardManager(this, "L") {}

  // Adds the relational guard resetter
  void add_relational_guard_resetter(
      std::shared_ptr<RelationalGuard> relational_guard) {
    _relational_guard_resetters.emplace_back(std::move(relational_guard));
  }

  // Python visible API to check guard function.
  bool check(py::handle value) {
    return check_nopybind(value.ptr());
  }

  // Python visible API to check_verbose guard function.
  GuardDebugInfo check_verbose(py::handle value) {
    return check_verbose_nopybind(value.ptr());
  }

  // Fast check function.
  template <typename T>
  bool check_nopybind_template(T* value) { // borrowed ref
    // Check [Note on GIL interaction with mutex lock] for details on why we
    // need mutex and its interactions with GIL.
    PyThreadState* _save = nullptr;
    Py_UNBLOCK_THREADS; // ; is added to avoid clang-formatting
    std::lock_guard<std::mutex> lock_guard(_lock);
    Py_BLOCK_THREADS; // ; is added to avoid clang-formatting

    // Get the local state. This will be used for TENSOR_MATCH guards.
    if (_init_local_state) {
      LocalState state;
      _local_state = state;
    }

    if (!GuardManager::check_leaf_guards_nopybind(value)) {
      _reset_relational_guard_state();
      return false;
    }

    // Run accessor guards without TorchFunction enabled
    // Dynamo should only be adding guards on values without
    // torch function at this point, because if there
    // was a torch function, we should've traced through it
    const at::impl::TorchFunctionDisabledState old_state =
        at::impl::PythonTorchFunctionTLS::get_disabled_state();
    at::impl::PythonTorchFunctionTLS::set_disabled_state(
        at::impl::TorchFunctionDisabledState::ALL_DISABLED);

    if (!GuardManager::check_accessors_nopybind(value)) {
      at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
      _reset_relational_guard_state();
      return false;
    }

    // Iterate over epilogue leaf guards.
    for (const auto& guard : _epilogue_lambda_guards) {
      if (!guard->check_nopybind(value)) { // early exit
        at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
        _reset_relational_guard_state();
        return false;
      }
    }

    at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
    _reset_relational_guard_state();
    return true;
  }

  bool check_nopybind(PyObject* value) override {
    return check_nopybind_template(value);
  }

  bool check_nopybind(FrameLocalsMapping* value) override {
    return check_nopybind_template(value);
  }

  // Fast check_verbose function.
  GuardDebugInfo check_verbose_nopybind(
      PyObject* value) override { // borrowed ref
    // Check [Note on GIL interaction with mutex lock] for details on why we
    // need mutex and its interactions with GIL.
    PyThreadState* _save = nullptr;
    Py_UNBLOCK_THREADS; // ; is added to avoid clang-formatting
    std::lock_guard<std::mutex> lock_guard(_lock);
    Py_BLOCK_THREADS; // ; is added to avoid clang-formatting

    // Get the local state. This will be used for TENSOR_MATCH guards.
    if (_init_local_state) {
      LocalState state;
      _local_state = state;
    }

    int num_guards_executed = 0;

    // Run leaf guards
    // This includes the GlobalStateGuard and the Torch Function Mode stack
    // guard, which require Torch Function to be in its unmodified state
    const GuardDebugInfo& debug_info_leaf =
        GuardManager::check_leaf_guards_verbose_nopybind(
            value, num_guards_executed);

    if (!debug_info_leaf.result) {
      _reset_relational_guard_state();
      return debug_info_leaf;
    }

    const at::impl::TorchFunctionDisabledState old_state =
        at::impl::PythonTorchFunctionTLS::get_disabled_state();
    at::impl::PythonTorchFunctionTLS::set_disabled_state(
        at::impl::TorchFunctionDisabledState::ALL_DISABLED);
    const GuardDebugInfo& debug_info_accessors =
        GuardManager::check_accessors_verbose_nopybind(
            value, num_guards_executed);

    if (!debug_info_accessors.result) {
      at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
      _reset_relational_guard_state();
      return debug_info_accessors;
    }

    // Iterate over epilogue leaf guards
    for (const auto& guard : _epilogue_lambda_guards) {
      const GuardDebugInfo& tmp_debug_info =
          guard->check_verbose_nopybind(value);
      num_guards_executed++;
      if (!tmp_debug_info.result) {
        at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
        _reset_relational_guard_state();
        return GuardDebugInfo(
            false, tmp_debug_info.verbose_code_parts, num_guards_executed);
      }
    }
    at::impl::PythonTorchFunctionTLS::set_disabled_state(old_state);
    _reset_relational_guard_state();
    return GuardDebugInfo(true, num_guards_executed);
  }

  void add_epilogue_lambda_guard(std::unique_ptr<LeafGuard> leaf_guard) {
    _epilogue_lambda_guards.emplace_back(std::move(leaf_guard));
  }

  void set_init_local_state_flag() {
    _init_local_state = true;
  }

  // See note on [Ownership with cloning]
  RootGuardManager* clone_manager(const py::function& clone_filter_fn) {
    // Use clone_filter_fn
    if (!py::cast<bool>(clone_filter_fn(this))) {
      return nullptr;
    }
    RootGuardManager* cloned_root = new RootGuardManager();
    clone_common(cloned_root, cloned_root, clone_filter_fn);
    for (const auto& guard : _epilogue_lambda_guards) {
      cloned_root->_epilogue_lambda_guards.emplace_back(guard);
    }
    return cloned_root;
  }

  // DEBUG function - Returning raw pointers because we can't return unique_ptr
  // and pybind does not accept a unique_ptr reference return type.
  std::vector<LeafGuard*> get_epilogue_lambda_guards() const {
    std::vector<LeafGuard*> ret;
    ret.reserve(_epilogue_lambda_guards.size());
    for (const auto& guard : _epilogue_lambda_guards) {
      ret.push_back(guard.get());
    }
    return ret;
  }

 private:
  // Reset the state of all the relational guards on failure.
  void _reset_relational_guard_state() {
    for (auto& guard : _relational_guard_resetters) {
      guard->reset_state();
    }
  }

 public:
  // Local state for TENSOR_MATCH guards.
  LocalState _local_state;

 private:
  // All the relational guards under this guard manager. We only use these
  // when the guard evaluates to False. This ensures that guard state is reset
  // on guard failure so that next invocation is clean.
  std::vector<std::shared_ptr<RelationalGuard>> _relational_guard_resetters;

  // These guards are lambda guards, i.e., the guards that lack C++
  // implementation. For simplicity, we add these guards at the root. They
  // MUST be run after all other guard managers have finished to ensure that
  // the epilogue guards do not step on some nonexistent getattr or getitem.
  // NB - shared_ptr is used to share the epilogue guards with the cloned guard
  // manager.
  std::vector<std::shared_ptr<LeafGuard>> _epilogue_lambda_guards;

  // [Note on GIL interaction with mutex lock]
  // We use std::mutex to prevent multiple threads from running
  // check/check_verbose simultaneously. This is to prevent race condition due
  // to state changes in RelationalGuard.
  //
  // However, we also need to be careful about GIL interaction with mutex. There
  // is a chance of deadlock
  //
  //    Thread 1: has GIL, waiting for lock
  //    Thread 2: has lock, waiting for GIL
  //
  // This can happen when Thread 2 earlier acquired the mutex lock, starting
  // running the critical section of check function and then called some python
  // function (like LAMBDA_GUARD) and reached Cpython codebase that checks if it
  // should release the GIL (typically happens after every few bytecode
  // instructions). Thread 2 here can decide to release the GIL. Thread 1 can
  // acquire GIL and reach the mutex, where it will wait forever.
  //
  // To avoid this, each thread releases the GIL before acquiring the mutex and
  // then acquires the GIL again after acquiring the mutex lock by using
  // Py_BLOCK_THREADS and Py_UNBLOCK_THREADS. This avoids the deadlock.
  std::mutex _lock;

  // We init LocalState only when this flag it set. This flag is set during
  // TENSOR_MATCH guard init.
  bool _init_local_state = false;
};

/*
 * Dicts are common in python code. Therefore, we handle guards for dicts
 * differently and use PyDict_* APIs which are faster than PyObject_* APIs
 * because of no ref count increments/decrements.
 *
 * DictGuardManager relies on the order of dict.keys(). It keeps track of the
 * indices of dict.keys() to access the key, value pair.
 */
typedef std::pair<std::unique_ptr<GuardManager>, std::unique_ptr<GuardManager>>
    KeyValueManager;
class DictGuardManager : public GuardManager {
 public:
  DictGuardManager(
      RootGuardManager* root,
      std::string source,
      py::handle example_value)
      : GuardManager(root, std::move(source)),
        _size(PyDict_Size(example_value.ptr())),
        _expected_type(Py_TYPE(example_value.ptr())),
        _is_exact_dict_type(PyDict_CheckExact(example_value.ptr())) {}

  GuardManager* get_key_manager(
      py::object key_index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum) {
    KeyValueManager& key_value_manager =
        _get_index_manager(std::move(key_index));
    if (!key_value_manager.first) {
      key_value_manager.first = make_guard_manager(
          this->get_root(),
          std::move(source),
          example_value,
          guard_manager_enum);
    };
    return key_value_manager.first.get();
  }

  GuardManager* get_value_manager(
      py::object key_index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum) {
    KeyValueManager& key_value_manager =
        _get_index_manager(std::move(key_index));
    if (!key_value_manager.second) {
      key_value_manager.second = make_guard_manager(
          this->get_root(),
          std::move(source),
          example_value,
          guard_manager_enum);
    };
    return key_value_manager.second.get();
  }

  bool check_nopybind(PyObject* obj) override { // borrowed ref
    // TODO(janimesh) - Implement a fast-path using dict versions.

    if (Py_TYPE(obj) != _expected_type) {
      _fail_count += 1;
      return false;
    }

    if (PyDict_Size(obj) != _size) {
      _fail_count += 1;
      return false;
    }

    // Early return
    if (_size == 0) {
      return true;
    }

    // Invokes the base class's check_nopybind method. We permit a limited set
    // of leaf guards and accessors within the DictGuardManager framework.
    // Integrating certain guards or accessors directly within the
    // DictGuardManager can be challenging. For instance, `type(dict_object)` as
    // an accessor is permissible, which otherwise would be hard to integrate
    // directly into DictGuardManager.  Similarly, incorporating guards such as
    // DICT_CONTAINS and DICT_VERSION as leaf guards offers a simpler solution
    // than embedding these functionalities within the DictGuardManager itself.
    if (!GuardManager::check_nopybind(obj)) {
      _fail_count += 1;
      // No need to shuffle the child guards, just return.
      return false;
    }

    PyObject *key = nullptr, *value = nullptr;
    Py_ssize_t pos = 0;

    // Points to an element in the _indices vector.
    size_t index_pointer = 0;
    // Points to the key index in the dict
    Py_ssize_t dict_pointer = 0;

    while (index_pointer < _indices.size() &&
           PyDict_Next(obj, &pos, &key, &value)) {
      // Skip if dict_pointer is not a saved index.
      if (dict_pointer == _indices[index_pointer]) {
        index_pointer += 1;
        KeyValueManager& key_value_manager = _key_value_managers[dict_pointer];
        std::unique_ptr<GuardManager>& key_manager = key_value_manager.first;
        if (key_manager && !key_manager->check_nopybind(key)) {
          return false;
        }
        std::unique_ptr<GuardManager>& value_manager = key_value_manager.second;
        if (value_manager && !value_manager->check_nopybind(value)) {
          return false;
        }
      }
      dict_pointer += 1;
    }
    return true;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    if (Py_TYPE(obj) != _expected_type) {
      return GuardDebugInfo(false, "TYPE_MISMATCH(" + get_source() + ")", 0);
    }

    if (PyDict_Size(obj) != _size) {
      return GuardDebugInfo(
          false, "len(" + get_source() + ") != " + std::to_string(_size), 0);
    }

    // Early return
    if (_size == 0) {
      return GuardDebugInfo(true, 0);
    }

    // Invokes the base class's check_nopybind method. We permit a limited set
    // of leaf guards and accessors within the DictGuardManager framework.
    // Integrating certain guards or accessors directly within the
    // DictGuardManager can be challenging. For instance, `type(dict_object)` as
    // an accessor is permissible, which otherwise would be hard to integrate
    // directly into DictGuardManager.  Similarly, incorporating guards such as
    // DICT_CONTAINS and DICT_VERSION as leaf guards offers a simpler solution
    // than embedding these functionalities within the DictGuardManager itself.
    GuardDebugInfo debug_info = GuardManager::check_verbose_nopybind(obj);
    if (!debug_info.result) {
      return debug_info;
    }

    PyObject *key = nullptr, *value = nullptr;
    Py_ssize_t pos = 0;

    // Points to an element in the _indices vector.
    size_t index_pointer = 0;
    Py_ssize_t dict_pointer = 0;

    int num_guards_executed = 0;
    while (index_pointer < _indices.size() &&
           PyDict_Next(obj, &pos, &key, &value)) {
      // Skip if pos is not a saved index.
      if (dict_pointer == _indices[index_pointer]) {
        index_pointer += 1;
        KeyValueManager& key_value_manager = _key_value_managers[dict_pointer];
        std::unique_ptr<GuardManager>& key_manager = key_value_manager.first;
        if (key_manager) {
          GuardDebugInfo debug_info = key_manager->check_verbose_nopybind(key);
          num_guards_executed += debug_info.num_guards_executed;
          if (!debug_info.result) {
            return GuardDebugInfo(
                false, debug_info.verbose_code_parts, num_guards_executed);
          }
        }
        std::unique_ptr<GuardManager>& value_manager = key_value_manager.second;
        if (value_manager) {
          GuardDebugInfo debug_info =
              value_manager->check_verbose_nopybind(value);
          num_guards_executed += debug_info.num_guards_executed;
          if (!debug_info.result) {
            return GuardDebugInfo(
                false, debug_info.verbose_code_parts, num_guards_executed);
          }
        }
      }
      dict_pointer += 1;
    }
    return GuardDebugInfo(true, num_guards_executed);
  }

  void skip_adding_guard(const py::object& a, const py::object& b) {
    // The `add_leaf_guard` method in `DictGuardManager` is overridden to block
    // the addition of leaf guards. However, this is too strict. Python side of
    // guard management frequently adds TYPE_MATCH and DICT_LENGTH on
    // DictGuardManager. We could refactor Python side to never call these
    // guards on dict objects, but that results in messy code. Instead, we just
    // override these two guards to not go through add_leaf_guard code path and
    // skip adding guards. This makes the python side easy.
  }

  void fail_on_get_child_manager(
      const py::object& a,
      const std::string& source,
      const py::object& b) {
    throw std::runtime_error("Can not add an accessor to DictGuardManager");
  }

  void add_leaf_guard(std::shared_ptr<LeafGuard> leaf_guard) override {
    // If you are calling this, you probably want to go through a key, value
    // child manager and then add a leaf guard on them. DictGuardManager already
    // has TYPE_MATCH and LENGTH_CHECK built in.
    throw std::runtime_error("DictGuardManager does not support a leaf_guard");
  }

  // Debug helper - Returning raw pointers because we can't return unique_ptr
  // and pybind does not accept a unique_ptr reference return type.
  std::unordered_map<Py_ssize_t, std::pair<GuardManager*, GuardManager*>>
  get_key_value_managers() {
    std::unordered_map<Py_ssize_t, std::pair<GuardManager*, GuardManager*>> ret;
    for (auto index : _indices) {
      ret[index] = std::make_pair(
          _key_value_managers[index].first.get(),
          _key_value_managers[index].second.get());
    }
    return ret;
  }

  bool is_exact_dict_type() {
    return _is_exact_dict_type;
  }

 public: // cloning functions
  DictGuardManager(
      RootGuardManager* cloned_root,
      std::string source,
      Py_ssize_t size,
      PyTypeObject* expected_type,
      bool is_exact_dict_type,
      std::vector<Py_ssize_t> indices)
      : GuardManager(cloned_root, std::move(source), true),
        _size(size),
        _expected_type(expected_type),
        _is_exact_dict_type(is_exact_dict_type),
        _indices(std::move(indices)) {}

  template <typename T>
  GuardManager* clone_dict_guard_manager(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) {
    if (!py::cast<bool>(clone_filter_fn(this))) {
      return nullptr;
    }
    T* cloned_mgr = new T(
        cloned_root,
        get_source(),
        _size,
        _expected_type,
        _is_exact_dict_type,
        _indices);

    clone_common(cloned_root, cloned_mgr, clone_filter_fn);
    for (auto index : _indices) {
      KeyValueManager& key_value_manager = _key_value_managers[index];
      std::unique_ptr<GuardManager>& key_manager = key_value_manager.first;
      std::unique_ptr<GuardManager>& value_manager = key_value_manager.second;

      cloned_mgr->_key_value_managers[index] = std::make_pair(nullptr, nullptr);

      if (key_manager) {
        GuardManager* cloned_key_manager =
            key_manager->clone(cloned_root, clone_filter_fn);
        if (cloned_key_manager) {
          cloned_mgr->_key_value_managers[index].first =
              std::unique_ptr<GuardManager>(cloned_key_manager);
        }
      }

      if (value_manager) {
        GuardManager* cloned_value_manager =
            value_manager->clone(cloned_root, clone_filter_fn);
        if (cloned_value_manager) {
          cloned_mgr->_key_value_managers[index].second =
              std::unique_ptr<GuardManager>(cloned_value_manager);
        }
      }
    }
    return cloned_mgr;
  }

  GuardManager* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_dict_guard_manager<DictGuardManager>(
        cloned_root, clone_filter_fn);
  }

 private:
  /**
   * Adds a new KeyDictGuardAccessor. If the accessor is already present, we
   * just return the guard manager.
   */
  KeyValueManager& _get_index_manager(py::object key_index) {
    // Check if the accessor is already present.
    Py_ssize_t index = py::cast<Py_ssize_t>(std::move(key_index));
    auto it = _key_value_managers.find(index);
    if (it != _key_value_managers.end()) {
      return it->second;
    }
    _indices.push_back(index);
    // Always keep the _indices array sorted
    std::sort(_indices.begin(), _indices.end());
    _key_value_managers[index] = std::make_pair(nullptr, nullptr);
    return _key_value_managers[index];
  }

 protected:
  Py_ssize_t _size;
  // DictGuardManager supports both exact dict type and non-exact dict type.
  // Therefore, we have to compare the type to early exit.
  PyTypeObject* _expected_type;
  bool _is_exact_dict_type; // Useful to check getattr_manager validity.
  std::vector<Py_ssize_t> _indices;
  std::unordered_map<Py_ssize_t, KeyValueManager> _key_value_managers;
};

GuardManager* clone_guard_manager(
    GuardManager* from,
    RootGuardManager* cloned_root,
    const py::function& clone_filter_fn) {
  return from->clone(cloned_root, clone_filter_fn);
}

void add_relational_guard_resetter_to_cloned_root(
    RootGuardManager* root,
    std::shared_ptr<RelationalGuard> guard) {
  root->add_relational_guard_resetter(std::move(guard));
}

std::unique_ptr<GuardManager> make_guard_manager(
    RootGuardManager* root,
    std::string source,
    py::handle example_value,
    py::handle guard_manager_enum) {
#if IS_PYBIND_2_13_PLUS
  using threeobjects = std::tuple<py::object, py::object, py::object>;
  PYBIND11_CONSTINIT static py::gil_safe_call_once_and_store<threeobjects>
      storage;

  auto& [guard_manager_enum_class, base_guard_manager_enum, dict_guard_manager_enum] =
      storage
          .call_once_and_store_result([]() -> threeobjects {
            py::object guard_manager_enum_class =
                py::module_::import("torch._dynamo.guards")
                    .attr("GuardManagerType");
            return {
                guard_manager_enum_class,
                guard_manager_enum_class.attr("GUARD_MANAGER"),
                guard_manager_enum_class.attr("DICT_GUARD_MANAGER")};
          })
          .get_stored();
#else
  static py::object guard_manager_enum_class =
      py::module_::import("torch._dynamo.guards").attr("GuardManagerType");
  static py::object base_guard_manager_enum =
      guard_manager_enum_class.attr("GUARD_MANAGER");
  static py::object dict_guard_manager_enum =
      guard_manager_enum_class.attr("DICT_GUARD_MANAGER");
#endif
  if (py::isinstance<py::dict>(example_value)) {
    if (guard_manager_enum.is(base_guard_manager_enum)) {
      // For dicts that don't need to guard on keys, we can just rely on the
      // base GuardManager.
      return std::make_unique<GuardManager>(
          root, std::move(source), example_value);
    } else if (guard_manager_enum.is(dict_guard_manager_enum)) {
      return std::make_unique<DictGuardManager>(
          root, std::move(source), example_value);
    } else {
      throw py::type_error("Invalid guard manager enum");
    }
  }
  return std::make_unique<GuardManager>(root, std::move(source));
}

class TORCH_FUNCTION_MODE_STACK : public LeafGuard {
 public:
  TORCH_FUNCTION_MODE_STACK(
      const py::list& initial_stack,
      py::object verbose_code_parts)
      : LeafGuard(std::move(verbose_code_parts)) {
    Py_ssize_t len = PyList_Size(initial_stack.ptr());
    for (Py_ssize_t idx = 0; idx < len; idx++) {
      PyObject* mode = PyList_GetItem(initial_stack.ptr(), idx); // borrowed ref
      auto type = Py_TYPE(mode);
      this->_ref_stack.push_back(type);
    }
  }

  template <typename T>
  bool check_nopybind_template(T* value) {
    // Ignore value arg, only used to satisfy the interface
    const size_t len = (size_t)at::impl::PythonTorchFunctionTLS::stack_len();
    const size_t ref_stack_size = this->_ref_stack.size();

    if (len != ref_stack_size) {
      return false;
    }

    for (int64_t idx = 0; (size_t)idx < len; idx++) {
      std::shared_ptr<c10::SafePyObject> mode =
          at::impl::PythonTorchFunctionTLS::get_stack_at(idx);

      PyTypeObject* mode_type = Py_TYPE(mode->ptr(getPyInterpreter()));
      if (mode_type != _ref_stack.at(idx)) {
        return false;
      }
    }

    return true;
  }

  bool check_nopybind(PyObject* value) override {
    return check_nopybind_template(value);
  }

  bool check_nopybind(FrameLocalsMapping* value) override {
    return check_nopybind_template(value);
  }

 private:
  std::vector<PyTypeObject*> _ref_stack;
};

class DISPATCH_KEY_SET_MATCH : public LeafGuard {
 public:
  DISPATCH_KEY_SET_MATCH(
      RootGuardManager* root_guard_manager,
      py::object value,
      py::object verbose_code_parts)
      : LeafGuard(root_guard_manager, std::move(verbose_code_parts)) {
    root_guard_manager->set_init_local_state_flag();
    c10::DispatchKeySet value_ = value.cast<c10::DispatchKeySet>();
    raw_repr = _root_guard_manager->_local_state.apply(value_).raw_repr();
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    py::handle handle = py::handle(value);
    c10::DispatchKeySet value_ = handle.cast<c10::DispatchKeySet>();
    return raw_repr ==
        _root_guard_manager->_local_state.apply(value_).raw_repr();
  }

 private:
  uint64_t raw_repr;
};

class TENSOR_MATCH : public LeafGuard {
 public:
  TENSOR_MATCH(
      RootGuardManager* root_guard_manager,
      py::object value,
      py::object dynamic_dims_sizes_py,
      py::object dynamic_dims_strides_py,
      py::object tensor_name,
      py::object verbose_code_parts,
      py::object pytype,
      py::object dispatch_keys)
      : LeafGuard(root_guard_manager, std::move(verbose_code_parts)),
        _tensor_name(py::cast<std::string>(std::move(tensor_name))) {
    root_guard_manager->set_init_local_state_flag();
    PyObject* item = value.ptr();
    if (!THPVariable_CheckExact(item) && !THPVariable_Check(item)) {
      PyErr_SetString(PyExc_TypeError, "expected Tensor()");
      return;
    }
    if (!PyType_Check(pytype.ptr())) {
      PyErr_SetString(PyExc_TypeError, "expected type object");
      return;
    }
    auto tensor = THPVariable_Unpack(item);

    std::vector<std::optional<c10::SymInt>> tensor_dims_size =
        pyListToVecOptInt(dynamic_dims_sizes_py.ptr());
    std::vector<std::optional<c10::SymInt>> tensor_dims_stride =
        pyListToVecOptInt(dynamic_dims_strides_py.ptr());

    tensor_dims_size = tensor_dims_size.empty()
        ? wrapIntegersInOptional(tensor.sym_sizes())
        : tensor_dims_size;
    tensor_dims_stride = tensor_dims_stride.empty()
        ? wrapIntegersInOptional(tensor.sym_strides())
        : tensor_dims_stride;
    LocalState state;
    _tensor_check = std::make_unique<TensorCheck>(
        state,
        (PyTypeObject*)pytype.ptr(),
        std::move(tensor),
        dispatch_keys.cast<c10::DispatchKeySet>(),
        std::move(tensor_dims_size),
        std::move(tensor_dims_stride));
  }

  bool check_nopybind(PyObject* value) override { // borrowed ref
    if (Py_TYPE(value) != _tensor_check->pytype) {
      return false;
    }
    return _tensor_check->check(
        _root_guard_manager->_local_state, THPVariable_Unpack(value));
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* value) override { // borrowed ref

    if (Py_TYPE(value) != _tensor_check->pytype) {
      std::stringstream fail_reason;
      PyObject* type_str = PyObject_Str(PyObject_Type(value));
      fail_reason << "expected type of '" << _tensor_name
                  << "' to be a tensor type, ";
      if (!type_str) {
        fail_reason << "but found a different type";
      } else {
        fail_reason << "' but found " << PyUnicode_AsUTF8(type_str);
      }
      return GuardDebugInfo(false, fail_reason.str(), 0);
    }

    std::string fail_reason = _tensor_check->check_verbose(
        _root_guard_manager->_local_state,
        THPVariable_Unpack(value),
        _tensor_name);

    if (!fail_reason.empty()) {
      if (is_parameter(py::handle(value))) {
        fail_reason += ". Guard failed on a parameter, consider using ";
        fail_reason +=
            "torch._dynamo.config.force_parameter_static_shapes = False ";
        fail_reason += "to allow dynamism on parameters.";
      }
      return GuardDebugInfo(false, fail_reason, 0);
    }
    return GuardDebugInfo(true, 1);
  }

 private:
  std::string _tensor_name;
  std::unique_ptr<TensorCheck> _tensor_check;
};

/**
 * Represents __getattr__ accessor.
 */
class GetAttrGuardAccessor : public GuardAccessor {
 public:
  GetAttrGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            name,
            std::move(source),
            example_value,
            guard_manager_enum),
        _attr_name(name.ptr()) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyObject_GetAttr(obj, _attr_name); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyObject_GetAttr(obj, _attr_name); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return GuardDebugInfo(
          false, "getattr failed on source " + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "GetAttrGuardAccessor(" + py::str(_attr_name).cast<std::string>() +
        ")";
  }

 public: // cloning functions
  GetAttrGuardAccessor(GuardManager* guard_manager, GetAttrGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GetAttrGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(GetAttrGuardAccessor* to) {
    to->_attr_name = _attr_name;
  }

 private:
  // no need of py::object here because the attr_name is already passed on to
  // the base class as accessor_key which is a py::object.
  PyObject* _attr_name{nullptr};
};

/**
 * Represents object.__getattribute__(obj, attr_name) accessor.
 */
class GenericGetAttrGuardAccessor : public GuardAccessor {
 public:
  GenericGetAttrGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            name,
            std::move(source),
            example_value,
            guard_manager_enum),
        _attr_name(name.ptr()) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyObject_GenericGetAttr(obj, _attr_name); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyObject_GenericGetAttr(obj, _attr_name); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return GuardDebugInfo(
          false, "getattr failed on source " + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "GenericGetAttrGuardAccessor(" +
        py::str(_attr_name).cast<std::string>() + ")";
  }

 public: // cloning functions
  GenericGetAttrGuardAccessor(
      GuardManager* guard_manager,
      GenericGetAttrGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GenericGetAttrGuardAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(GenericGetAttrGuardAccessor* to) {
    to->_attr_name = _attr_name;
  }

 private:
  // no need of py::object here because the attr_name is already passed on to
  // the base class as accessor_key which is a py::object.
  PyObject* _attr_name{nullptr};
};

/**
 * Represents x.__dict__ accessor.
 */
class GetGenericDictGuardAccessor : public GuardAccessor {
 public:
  GetGenericDictGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyObject_GenericGetDict(obj, nullptr); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyObject_GenericGetDict(obj, nullptr); // new ref
    if (x == nullptr) {
      // Attribute absent, clear the exception and return false.
      PyErr_Clear();
      return GuardDebugInfo(
          false, "getattr failed on source " + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "GetGenericDictGuardAccessor";
  }

 public: // cloning functions
  GetGenericDictGuardAccessor(
      GuardManager* guard_manager,
      GetGenericDictGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GetGenericDictGuardAccessor>(
        cloned_root, clone_filter_fn);
  }
};

/**
 * Represents __getitem__ accessor.
 */
class GetItemGuardAccessor : public GuardAccessor {
 public:
  GetItemGuardAccessor(
      RootGuardManager* root,
      py::object name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            name,
            std::move(source),
            example_value,
            guard_manager_enum),
        _attr_name(name.ptr()) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyObject_GetItem(obj, _attr_name); // new ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyObject_GetItem(obj, _attr_name); // new ref
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("KeyError on ") + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    return "GetItemGuardAccessor(" + py::str(_attr_name).cast<std::string>() +
        ")";
  }

 public: // cloning functions
  GetItemGuardAccessor(GuardManager* guard_manager, GetItemGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GetItemGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(GetItemGuardAccessor* to) {
    to->_attr_name = _attr_name;
  }

 private:
  // no need of py::object here because the attr_name is already passed on to
  // the base class as accessor_key which is a py::object.
  PyObject* _attr_name{nullptr};
};

/**
 * Represents f_locals[name] accessor. Special handling for frame locals since
 * we avoid converting it to Python as much as possible.
 * NB: We don't check for name order in frame locals since it is constant
 * across frames corresponding to the same code object.
 */
class FrameLocalsGuardAccessor : public GuardAccessor {
 public:
  FrameLocalsGuardAccessor(
      RootGuardManager* root,
      const py::tuple& key,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            key[0],
            std::move(source),
            example_value,
            guard_manager_enum),
        _key(key[0].ptr()),
        _framelocals_idx(key[1].cast<int>()),
        _is_immutable_object(is_immutable_object(example_value)) {}

  // Run as a result of calling run_root_guard_manager/check_nopybind
  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(
      FrameLocalsMapping* obj,
      bool matches_dict_tag = false) override { // borrowed ref
    if (matches_dict_tag && _is_immutable_object) {
      // immutable object and dict tag matches, we can skip the guard subtree.
      return true;
    }

    PyObject* x = obj->get(_framelocals_idx);
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    return _guard_manager->check_nopybind(x);
  }

  // Run as a result of calling check(), e.g. from Python
  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false) override {
    if (!PyDict_Check(obj)) {
      // This should not cause guard failure.
      // If this error is encountered, it probably means
      // we did not convert FrameLocalsMapping to dict (using to_dict()).
      throw std::runtime_error(
          "FrameLocalsGuardAccessor check expected dict() input");
    }

    if (matches_dict_tag && _is_immutable_object) {
      // immutable object and dict tag matches, we can skip the guard subtree.
      return true;
    }

    PyObject* x = PyDict_GetItem(obj, _key); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    return result;
  }

  // If we've reached here, it means the guard failed - `obj` should be the
  // FrameLocalsMapping converted into a Python dict and we should
  // behave like DictGetItemGuardAccessor.
  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    if (!PyDict_Check(obj)) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, "FrameLocalsGuardAccessor check expected dict() input", 0);
    }
    PyObject* x = PyDict_GetItem(obj, _key); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("KeyError on ") + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    return result;
  }

  std::string repr() const override {
    return "FrameLocalsGuardAccessor(key=" +
        py::repr(_key).cast<std::string>() +
        ", framelocals_idx=" + std::to_string(_framelocals_idx) + ")";
  }

 public: // cloning functions
  FrameLocalsGuardAccessor(
      GuardManager* guard_manager,
      FrameLocalsGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<FrameLocalsGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(FrameLocalsGuardAccessor* to) {
    to->_key = _key;
    to->_framelocals_idx = _framelocals_idx;
    to->_is_immutable_object = _is_immutable_object;
  }

 private:
  PyObject* _key{nullptr};
  int _framelocals_idx{-1};

  // If immutable object and dict tag matches, we can skip the guard subtree and
  // return true.
  bool _is_immutable_object{false};
};

/**
 * Represents dict[name] accessor. Needed since DictGuardManager does not
 * support sorting. We differentiate it from GetItemGuardAccessor because
 * PyDict_GetItem should be faster than PyObject_GetItem.
 */
class DictGetItemGuardAccessor : public GuardAccessor {
 public:
  DictGetItemGuardAccessor(
      RootGuardManager* root,
      py::object key,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            key,
            std::move(source),
            example_value,
            guard_manager_enum),
        _key(key.ptr()),
        _is_immutable_object(is_immutable_object(example_value)) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false) override {
    if (matches_dict_tag && _is_immutable_object &&
        _guard_manager->has_no_accessors()) {
      // immutable object and dict tag matches, we can skip the guard subtree.
      // NB: We only skip the subtree if there are no accessors in the subtree.
      // This is specifically for tensors which are used in symbolic shape C++
      // guards, and therefore have accessors on the tensor GuardManager itself.
      return true;
    }

    PyObject* x = PyDict_GetItem(obj, _key); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyDict_GetItem(obj, _key); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("KeyError on ") + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    return result;
  }

  std::string repr() const override {
    return "DictGetItemGuardAccessor(" + py::repr(_key).cast<std::string>() +
        ")";
  }

 public: // cloning functions
  DictGetItemGuardAccessor(
      GuardManager* guard_manager,
      DictGetItemGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<DictGetItemGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(DictGetItemGuardAccessor* to) {
    to->_key = _key;
    to->_is_immutable_object = _is_immutable_object;
  }

 private:
  PyObject* _key{nullptr};

  // If immutable object and dict tag matches, we can skip the guard subtree and
  // return true.
  bool _is_immutable_object{false};
};

/**
 * Represents list[index] accessor. It is faster than generic
 * GetItemGuardAccessor.
 */
class ListGetItemGuardAccessor : public GuardAccessor {
 public:
  ListGetItemGuardAccessor(
      RootGuardManager* root,
      const py::object& index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            index,
            std::move(source),
            example_value,
            guard_manager_enum),
        _index(py::cast<Py_ssize_t>(index)) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyList_GetItem(obj, _index); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyList_GetItem(obj, _index); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("IndexError on ") + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    return result;
  }

  std::string repr() const override {
    return "ListGetItemGuardAccessor(" + std::to_string(_index) + ")";
  }

 public: // cloning functions
  ListGetItemGuardAccessor(
      GuardManager* guard_manager,
      ListGetItemGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<ListGetItemGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(ListGetItemGuardAccessor* to) {
    to->_index = _index;
  }

 private:
  Py_ssize_t _index{-1};
};

/**
 * Represents tuple[index] accessor. It is faster than generic
 * GetItemGuardAccessor.
 */
class TupleGetItemGuardAccessor : public GuardAccessor {
 public:
  TupleGetItemGuardAccessor(
      RootGuardManager* root,
      const py::object& index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            index,
            std::move(source),
            example_value,
            guard_manager_enum),
        _index(py::cast<Py_ssize_t>(index)) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyTuple_GetItem(obj, _index); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyTuple_GetItem(obj, _index); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("IndexError on ") + get_source(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    return result;
  }

  std::string repr() const override {
    return "TupleGetItemGuardAccessor(" + std::to_string(_index) + ")";
  }

 public: // cloning functions
  TupleGetItemGuardAccessor(
      GuardManager* guard_manager,
      TupleGetItemGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<TupleGetItemGuardAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(TupleGetItemGuardAccessor* to) {
    to->_index = _index;
  }

 private:
  Py_ssize_t _index{-1};
};

enum class TensorProperty {
  SIZE = 0,
  STRIDE = 1,
  STORAGE_OFFSET = 2,
};

std::string to_string(TensorProperty prop) {
  switch (prop) {
    case TensorProperty::SIZE:
      return "TensorProperty::SIZE";
    case TensorProperty::STRIDE:
      return "TensorProperty::STRIDE";
    case TensorProperty::STORAGE_OFFSET:
      return "TensorProperty::STORAGE_OFFSET";
    default:
      return "TensorProperty::Unknown";
  }
}

/**
 * Represents tensor.size/shape/storage_offset accessor.
 */
template <TensorProperty _prop>
class TensorPropertyGuardAccessor : public GuardAccessor {
 public:
  TensorPropertyGuardAccessor(
      RootGuardManager* root,
      const py::object& index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            index,
            std::move(source),
            example_value,
            guard_manager_enum) {
    if (_prop != TensorProperty::STORAGE_OFFSET) {
      _index = py::cast<Py_ssize_t>(index);
    }
  }
  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    // We need to check here to ensure that `obj` is a tensor
    // Usually we have a TENSOR_MATCH in the parent guard manager
    // but in the case of ``tensor._base` we don't. When the tensor
    // is not a view `tensor._base` is None and we have to check
    // that here.
    if (!THPVariable_CheckExact(obj) && !THPVariable_Check(obj)) {
      return false;
    }
    at::Tensor tensor = THPVariable_Unpack(obj);
    std::optional<int64_t> opt_value;
    if (_prop == TensorProperty::SIZE) {
      if (_index >= tensor.dim()) {
        return false;
      }
      opt_value = tensor.sym_size(_index).maybe_as_int();
    } else if (_prop == TensorProperty::STRIDE) {
      if (_index >= tensor.dim()) {
        return false;
      }
      opt_value = tensor.sym_stride(_index).maybe_as_int();
    } else if (_prop == TensorProperty::STORAGE_OFFSET) {
      opt_value = tensor.sym_storage_offset().maybe_as_int();
    } else {
      throw std::runtime_error("Unknown property");
    }

    if (!opt_value.has_value()) {
      return false;
    }

    PyObject* py_value =
        PyLong_FromLongLong(opt_value.value()); // New reference
    bool result = _guard_manager->check_nopybind(py_value);
    Py_DECREF(py_value);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    // check that its a tensor
    if (!THPVariable_CheckExact(obj) && !THPVariable_Check(obj)) {
      return GuardDebugInfo(false, "not a tensor" + get_source(), 0);
    }
    at::Tensor tensor = THPVariable_Unpack(obj);
    std::optional<int64_t> opt_value;
    if (_prop == TensorProperty::SIZE) {
      if (_index >= tensor.dim()) {
        return GuardDebugInfo(false, "tensor has too few dimensions", 0);
      }
      opt_value = tensor.sym_size(_index).maybe_as_int();
    } else if (_prop == TensorProperty::STRIDE) {
      if (_index >= tensor.dim()) {
        return GuardDebugInfo(false, "tensor has too few dimensions", 0);
      }
      opt_value = tensor.sym_stride(_index).maybe_as_int();
    } else if (_prop == TensorProperty::STORAGE_OFFSET) {
      opt_value = tensor.sym_storage_offset().maybe_as_int();
    } else {
      return GuardDebugInfo(false, "unknown property", 0);
    }

    if (!opt_value.has_value()) {
      return GuardDebugInfo(false, "symbolic values found", 0);
    }

    PyObject* py_value =
        PyLong_FromLongLong(opt_value.value()); // New reference
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(py_value);
    Py_DECREF(py_value);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "TensorPropertyGuardAccessor<" + to_string(_prop) + +">(" +
        std::to_string(_index) + ")";
  }

 public: // cloning functions
  TensorPropertyGuardAccessor(
      GuardManager* guard_manager,
      TensorPropertyGuardAccessor<_prop>* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<TensorPropertyGuardAccessor<_prop>>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(TensorPropertyGuardAccessor<_prop>* to) {
    to->_index = _index;
  }

 private:
  Py_ssize_t _index{-1};
};

/**
 * Indexed Guard Accessor that retrieves a value from the child
 * and sends a (index, source) to the parent.
 */
class IndexedGuardAccessor : public GuardAccessor {
 public:
  IndexedGuardAccessor(
      RootGuardManager* root,
      py::int_ index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            index,
            std::move(source),
            example_value,
            guard_manager_enum),
        _index(index) {}
  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* tuple = PyTuple_Pack(2, _index.ptr(), obj); // New reference
    bool result = _guard_manager->check_nopybind(tuple);
    Py_DECREF(tuple);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* tuple = PyTuple_Pack(2, _index.ptr(), obj); // New reference
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(tuple);
    Py_DECREF(tuple);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "IndexedGuardAccesor(" +
        std::to_string(py::cast<Py_ssize_t>(_index)) + ")";
  }

 public: // cloning functions
  IndexedGuardAccessor(GuardManager* guard_manager, IndexedGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<IndexedGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(IndexedGuardAccessor* to) {
    to->_index = _index;
  }

 private:
  py::int_ _index{-1};
};

/**
 * Represents tensor.grad accessor.
 */
class GradGuardAccessor : public GuardAccessor {
 public:
  GradGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    // check that its a tensor
    if (!THPVariable_CheckExact(obj) && !THPVariable_Check(obj)) {
      return false;
    }
    PyObject* grad =
        THPVariable_Wrap(THPVariable_Unpack(obj).grad()); // New reference
    bool result = _guard_manager->check_nopybind(grad);
    // For undefined tensor, THPVariable_Wrap returns Py_RETURN_NONE. So, no
    // need of Py_XDECREF.
    Py_DECREF(grad);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    // check that its a tensor
    if (!THPVariable_CheckExact(obj) && !THPVariable_Check(obj)) {
      return GuardDebugInfo(
          false, "not a tensor - grad field is accessed " + get_source(), 0);
    }
    PyObject* grad =
        THPVariable_Wrap(THPVariable_Unpack(obj).grad()); // New reference
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(grad);
    // For undefined tensor, THPVariable_Wrap returns Py_RETURN_NONE. So, no
    // need of Py_XDECREF.
    Py_DECREF(grad);
    return result;
  }

  std::string repr() const override {
    // Helpful when printing GuardManager tree structure.
    return "GradGuardAccessor(grad)";
  }

 public: // cloning functions
  GradGuardAccessor(GuardManager* guard_manager, GradGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GradGuardAccessor>(cloned_root, clone_filter_fn);
  }
};

/**
 * Represents func.__defaults__ accessor.
 */
class FuncDefaultsGuardAccessor : public GuardAccessor {
 public:
  FuncDefaultsGuardAccessor(
      RootGuardManager* root,
      py::object name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* func = obj;
    if (PyMethod_Check(obj)) {
      func = PyMethod_GET_FUNCTION(obj); // borrowed ref
    } else if (PyInstanceMethod_Check(obj)) {
      func = PyInstanceMethod_GET_FUNCTION(obj); // borrowed ref
    }
    PyObject* x = PyFunction_GetDefaults(func); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    return _guard_manager->check_nopybind(x);
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* func = obj;
    if (PyMethod_Check(obj)) {
      func = PyMethod_GET_FUNCTION(obj); // borrowed ref
    } else if (PyInstanceMethod_Check(obj)) {
      func = PyInstanceMethod_GET_FUNCTION(obj); // borrowed ref
    }
    PyObject* x = PyFunction_GetDefaults(func);
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false,
          std::string(repr() + ": Not a function on ") + get_source(),
          0);
    }

    return _guard_manager->check_verbose_nopybind(x);
  }

  std::string repr() const override {
    return "FuncDefaultsGuardAccessor";
  }

 public: // cloning functions
  FuncDefaultsGuardAccessor(
      GuardManager* guard_manager,
      FuncDefaultsGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }
  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<FuncDefaultsGuardAccessor>(
        cloned_root, clone_filter_fn);
  }
};

/**
 * Represents func.__kwdefaults__ accessor.
 */
class FuncKwDefaultsGuardAccessor : public GuardAccessor {
 public:
  FuncKwDefaultsGuardAccessor(
      RootGuardManager* root,
      py::object name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* func = obj;
    if (PyMethod_Check(obj)) {
      func = PyMethod_GET_FUNCTION(obj); // borrowed ref
    } else if (PyInstanceMethod_Check(obj)) {
      func = PyInstanceMethod_GET_FUNCTION(obj); // borrowed ref
    }
    PyObject* x = PyFunction_GetKwDefaults(func); // borrowed ref
    if (x == nullptr) {
      PyErr_Clear();
      return false;
    }
    return _guard_manager->check_nopybind(x);
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* func = obj;
    if (PyMethod_Check(obj)) {
      func = PyMethod_GET_FUNCTION(obj); // borrowed ref
    } else if (PyInstanceMethod_Check(obj)) {
      func = PyInstanceMethod_GET_FUNCTION(obj); // borrowed ref
    }
    PyObject* x = PyFunction_GetKwDefaults(func);
    if (x == nullptr) {
      PyErr_Clear();
      return GuardDebugInfo(
          false,
          std::string(repr() + ": Not a function on ") + get_source(),
          0);
    }

    return _guard_manager->check_verbose_nopybind(x);
  }

  std::string repr() const override {
    return "FuncKwDefaultsGuardAccessor";
  }

 public: // cloning functions
  FuncKwDefaultsGuardAccessor(
      GuardManager* guard_manager,
      FuncKwDefaultsGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<FuncKwDefaultsGuardAccessor>(
        cloned_root, clone_filter_fn);
  }
};

/**
 * Represents f_globals accessor. This sits as a child accessor of the
 * RootGuardManager.
 */
class GlobalsGuardAccessor : public GuardAccessor {
 public:
  GlobalsGuardAccessor(
      RootGuardManager* root,
      py::dict globals_dict,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            globals_dict,
            std::move(source),
            example_value,
            guard_manager_enum),
        _globals_dict(globals_dict.ptr()) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    // Ignore the obj arg. This is required to satisfy the function signature.
    // Just pass on the globals dict to the child manager.
    return _guard_manager->check_nopybind(_globals_dict);
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    // Ignore the obj arg. This is required to satisfy the function signature.
    // Just pass on the globals dict to the child manager.
    return _guard_manager->check_verbose_nopybind(_globals_dict);
  }

  std::string repr() const override {
    return "GlobalsGuardAccessor";
  }

 public: // cloning functions
  GlobalsGuardAccessor(GuardManager* guard_manager, GlobalsGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GlobalsGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(GlobalsGuardAccessor* to) {
    to->_globals_dict = _globals_dict;
  }

 private:
  // no need of py::object here because the globals_dict is already passed on to
  // the base class as accessor_key which is a py::object.
  PyObject* _globals_dict{nullptr};
};

/**
 * Represent type(...) accessor.
 */
class TypeGuardAccessor : public GuardAccessor {
 public:
  // name = __type_accessor__, a unique string used as attribute name.
  TypeGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = (PyObject*)Py_TYPE(obj); // borrowed ref
    return _guard_manager->check_nopybind(x);
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = (PyObject*)Py_TYPE(obj); // borrowed ref
    return _guard_manager->check_verbose_nopybind(x);
  }

  std::string repr() const override {
    return "TypeGuardAccessor";
  }

 public: // cloning functions
  TypeGuardAccessor(GuardManager* guard_manager, TypeGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<TypeGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(TypeGuardAccessor* to) {}
};

/**
 * Getitem tuple_iterator accessor.
 */
class TupleIteratorGetItemAccessor : public GuardAccessor {
 public:
  TupleIteratorGetItemAccessor(
      RootGuardManager* root,
      py::object index,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            index,
            std::move(source),
            example_value,
            guard_manager_enum),
        _index(py::cast<Py_ssize_t>(std::move(index))) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    _PyTupleIterObject* it = (_PyTupleIterObject*)obj;
    PyObject* x =
        PyTuple_GET_ITEM(it->it_seq, it->it_index + _index); // borrowed ref
    if (x == nullptr) {
      // Out of range.
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    _PyTupleIterObject* it = (_PyTupleIterObject*)obj;
    PyObject* x =
        PyTuple_GET_ITEM(it->it_seq, it->it_index + _index); // borrowed ref
    if (x == nullptr) {
      // Out of range.
      PyErr_Clear();
      return GuardDebugInfo(false, std::string("IndexError ") + repr(), 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    return result;
  }

  std::string repr() const override {
    return "TupleIteratorGetItemAccessor(" + std::to_string(_index) + ")";
  }

 public: // cloning functions
  TupleIteratorGetItemAccessor(
      GuardManager* guard_manager,
      TupleIteratorGetItemAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<TupleIteratorGetItemAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(TupleIteratorGetItemAccessor* to) {
    to->_index = _index;
  }

 private:
  Py_ssize_t _index{-1};
};

/**
 * GlobalWeakRef accessor. Dynamo can insert a weakref object into the frame
 * globals. This accessor reads the globals and then calls the weakref object
 * to get the underlying object. This is a child of GlobalsGuardAccessor.
 * Therefore, we will get the globals dict while calling check_nopybind.
 */
class GlobalWeakRefGuardAccessor : public GuardAccessor {
 public:
  GlobalWeakRefGuardAccessor(
      RootGuardManager* root,
      py::object global_name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            global_name,
            std::move(source),
            example_value,
            guard_manager_enum),
        _global_name(global_name.ptr()) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    // obj is globals dict because GlobalWeakRefGuardAccessor has to be a
    // child of GlobalsGuardAccessor.
    PyObject* weakref = PyDict_GetItem(obj, _global_name); // borrowed ref
    if (weakref == nullptr) {
      // The weakref is not in the globals dict.
      PyErr_Clear();
      return false;
    }

    if (!PyWeakref_Check(weakref)) {
      return false;
    }

    PyObject* x = nullptr;
    if (PyWeakref_GetRef(weakref, &x) == -1) { // strong reference
      // error when attempting to call ref
      PyErr_Clear();
      return false;
    }
    if (x == nullptr) {
      // weakref is dead
      x = Py_NewRef(Py_None);
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    // obj is globals dict because GlobalWeakRefGuardAccessor has to be a
    // child of GlobalsGuardAccessor.
    PyObject* weakref = PyDict_GetItem(obj, _global_name); // borrowed ref
    if (weakref == nullptr) {
      // The weakref is not in the globals dict.
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("KeyError on ") + get_source(), 0);
    }

    if (!PyWeakref_Check(weakref)) {
      return GuardDebugInfo(
          false, std::string("Not a weakref ") + get_source(), 0);
    }

    PyObject* x = nullptr;
    if (PyWeakref_GetRef(weakref, &x) == -1) { // strong reference
      // error when attempting to call ref
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("Weakref_GetRef failed ") + get_source(), 0);
    }
    if (x == nullptr) {
      // weakref is dead
      x = Py_NewRef(Py_None);
    }
    auto result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    return "GlobalWeakRefGuardAccessor(" +
        py::str(_global_name).cast<std::string>() + ")";
  }

 public: // cloning functions
  GlobalWeakRefGuardAccessor(
      GuardManager* guard_manager,
      GlobalWeakRefGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }
  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<GlobalWeakRefGuardAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(GlobalWeakRefGuardAccessor* to) {
    to->_global_name = _global_name;
  }

 private:
  PyObject* _global_name{nullptr};
};

/**
 * Implements weakref call - x_weak()
 */
class WeakRefCallGuardAccessor : public GuardAccessor {
 public:
  WeakRefCallGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    if (!PyWeakref_Check(obj)) {
      return false;
    }

    PyObject* x = nullptr;
    if (PyWeakref_GetRef(obj, &x) == -1) { // strong reference
      // error when attempting to call ref
      PyErr_Clear();
      return false;
    }
    if (x == nullptr) {
      // weakref is dead
      x = Py_NewRef(Py_None);
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    if (!PyWeakref_Check(obj)) {
      return GuardDebugInfo(
          false, std::string("Not a weakref obj ") + get_source(), 0);
    }

    PyObject* x = nullptr;
    if (PyWeakref_GetRef(obj, &x) == -1) { // strong reference
      // error when attempting to call ref
      PyErr_Clear();
      return GuardDebugInfo(
          false, std::string("Weakref_GetRef failed ") + get_source(), 0);
    }
    if (x == nullptr) {
      // weakref is dead
      x = Py_NewRef(Py_None);
    }
    auto result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    return "WeakRefCallGuardAccessor()";
  }

 public: // cloning functions
  WeakRefCallGuardAccessor(
      GuardManager* guard_manager,
      WeakRefCallGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<WeakRefCallGuardAccessor>(cloned_root, clone_filter_fn);
  }

  void clone_visitor(WeakRefCallGuardAccessor* to) {}
};

/**
 * Implements function call no args - e.g, torch.cuda.current_device()
 */
class CallFunctionNoArgsGuardAccessor : public GuardAccessor {
 public:
  CallFunctionNoArgsGuardAccessor(
      RootGuardManager* root,
      py::str name,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            std::move(name),
            std::move(source),
            example_value,
            guard_manager_enum) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    if (!PyCallable_Check(obj)) {
      return false;
    }

    PyObject* x = PyObject_CallNoArgs(obj);
    if (x == nullptr) {
      // Call failed, clear the exception and return false.
      PyErr_Clear();
      return false;
    }

    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    if (!PyCallable_Check(obj)) {
      return GuardDebugInfo(
          false, std::string("Not a callable obj ") + get_source(), 0);
    }

    PyObject* x = PyObject_CallNoArgs(obj);
    if (x == nullptr) {
      // Call failed, clear the exception and return debug info.
      std::string exc_message = get_exception_message();
      PyErr_Clear();
      return GuardDebugInfo(false, exc_message, 0);
    }

    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    return "CallFunctionNoArgsGuardAccessor()";
  }

 public: // cloning functions
  CallFunctionNoArgsGuardAccessor(
      GuardManager* guard_manager,
      CallFunctionNoArgsGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<CallFunctionNoArgsGuardAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(CallFunctionNoArgsGuardAccessor* to) {}
};

/**
 * Similar to PythonLambdaLeafGuard, this class is a way to allow developers to
 * supply accessor as a python function. This is useful for from_numpy source.
 */
class PythonLambdaGuardAccessor : public GuardAccessor {
 public:
  PythonLambdaGuardAccessor(
      RootGuardManager* root,
      py::function accessor_fn,
      std::string source,
      py::handle example_value,
      py::handle guard_manager_enum)
      : GuardAccessor(
            root,
            accessor_fn,
            std::move(source),
            example_value,
            guard_manager_enum),
        _accessor_fn(std::move(accessor_fn)) {}

  // NB: Intentional duplication between check_nopybind and
  // check_verbose_nopybind.
  bool check_nopybind(PyObject* obj, bool matches_dict_tag = false)
      override { // borrowed ref
    PyObject* x = PyObject_CallOneArg(_accessor_fn.ptr(), obj); // new ref
    if (x == nullptr) {
      // The accessor function failed.
      PyErr_Clear();
      return false;
    }
    bool result = _guard_manager->check_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  GuardDebugInfo check_verbose_nopybind(
      PyObject* obj) override { // borrowed ref
    PyObject* x = PyObject_CallOneArg(_accessor_fn.ptr(), obj); // new ref
    if (x == nullptr) {
      // The accessor function failed.
      std::string exc_message = get_exception_message();
      PyErr_Clear();
      return GuardDebugInfo(false, exc_message, 0);
    }
    GuardDebugInfo result = _guard_manager->check_verbose_nopybind(x);
    Py_DECREF(x);
    return result;
  }

  std::string repr() const override {
    return "PythonLambdaGuardAccessor";
  }

 public: // cloning functions
  PythonLambdaGuardAccessor(
      GuardManager* guard_manager,
      PythonLambdaGuardAccessor* from)
      : GuardAccessor(guard_manager, from) {
    from->clone_visitor(this);
  }

  GuardAccessor* clone(
      RootGuardManager* cloned_root,
      const py::function& clone_filter_fn) override {
    return clone_common<PythonLambdaGuardAccessor>(
        cloned_root, clone_filter_fn);
  }

  void clone_visitor(PythonLambdaGuardAccessor* to) {
    to->_accessor_fn = _accessor_fn;
  }

 private:
  py::object _accessor_fn;
};

void install_object_aliasing_guard(
    GuardManager* x,
    GuardManager* y,
    py::object verbose_code_parts) {
  // Adds tensor X is tensor Y guard. This is a an example of relational guard.
  // There is one guard object that is shared between two guard managers.
  std::shared_ptr<RelationalGuard> guard =
      std::make_shared<OBJECT_ALIASING>(std::move(verbose_code_parts));

  // Register the resetter on the root guard manager, so that it can reset
  // the newly added relational guard when the guard eval fails.
  x->get_root()->add_relational_guard_resetter(guard);

  // In case the guard is a DictGuardManager, OBJECT_ALIASING guard is a
  // permitted guard.
  x->add_permitted_leaf_guard(guard);
  y->add_permitted_leaf_guard(guard);
}

void install_no_tensor_aliasing_guard(
    const py::list& guard_managers,
    const py::list& tensor_names,
    py::object verbose_code_parts) {
  // Adds a guard that checks none of tensors alias. This is a an example of
  // relational guard. There is one guard object that is shared between multiple
  // guard managers.
  std::shared_ptr<RelationalGuard> guard = std::make_shared<NO_TENSOR_ALIASING>(
      tensor_names, std::move(verbose_code_parts));

  // Register the resetter on the root guard manager, so that it can reset
  // the newly added relational guard when the guard eval fails.
  py::cast<GuardManager*>(guard_managers[0])
      ->get_root()
      ->add_relational_guard_resetter(guard);
  for (const auto& guard_manager : guard_managers) {
    py::cast<GuardManager*>(guard_manager)->add_leaf_guard(guard);
  }
}

void install_symbolic_shape_guard(
    const py::list& guard_managers,
    py::int_ nargs_int,
    py::int_ nargs_float,
    py::int_ py_addr,
    py::object py_addr_keep_alive,
    py::object verbose_code_parts) {
  // Adds a guard that checks symbolic shapes. This is a an example of
  // relational guard. There is one guard object that is shared between
  // multiple guard managers.
  std::shared_ptr<RelationalGuard> guard =
      std::make_shared<SYMBOLIC_SHAPE_GUARD>(
          std::move(nargs_int),
          std::move(nargs_float),
          std::move(py_addr),
          std::move(py_addr_keep_alive),
          std::move(verbose_code_parts));

  // Register the resetter on the root guard manager, so that it can reset
  // the newly added relational guard when the guard eval fails.
  py::cast<GuardManager*>(guard_managers[0])
      ->get_root()
      ->add_relational_guard_resetter(guard);
  for (const auto& guard_manager : guard_managers) {
    py::cast<GuardManager*>(guard_manager)->add_leaf_guard(guard);
  }
}

void install_storage_overlapping_guard_with_checker(
    const std::shared_ptr<StorageOverlapChecker>& checker,
    const py::list& guard_managers,
    const py::object& verbose_code_parts,
    bool overlapping) {
  if (guard_managers.empty()) {
    // If there are no GuardManagers, there's no need to create a
    // STORAGE_OVERLAPPING guard.
    return;
  }

  std::shared_ptr<RelationalGuard> guard =
      std::make_shared<STORAGE_OVERLAPPING>(
          overlapping, checker, verbose_code_parts);
  py::cast<GuardManager*>(guard_managers[0])
      ->get_root()
      ->add_relational_guard_resetter(guard);
  for (const auto& guard_manager : guard_managers) {
    py::cast<GuardManager*>(guard_manager)->add_leaf_guard(guard);
  }
}

void install_storage_overlapping_guard(
    const py::list& overlapping_guard_managers,
    const py::list& non_overlapping_guard_managers,
    const py::object& verbose_code_parts) {
  // Create a single StorageOverlapChecker that will be shared amongst
  // the 2 STORAGE_OVERLAPPING guards below.
  std::shared_ptr<StorageOverlapChecker> checker =
      std::make_shared<StorageOverlapChecker>(
          overlapping_guard_managers.size(),
          non_overlapping_guard_managers.size());
  // Create the possibly overlapping storage guard.
  install_storage_overlapping_guard_with_checker(
      checker,
      overlapping_guard_managers,
      verbose_code_parts,
      /* overlapping= */ true);
  // Create the non-overlapping storage guard.
  install_storage_overlapping_guard_with_checker(
      checker,
      non_overlapping_guard_managers,
      verbose_code_parts,
      /* overlapping= */ false);
}

char flush_cache_by_eviction() {
  constexpr size_t evict_size = 32 * 1024 * 1024;
  std::vector<char> buffer(evict_size, 1);

  volatile char sink = 0;
  for (size_t i = 0; i < buffer.size(); i += 64) {
    sink ^= buffer[i];
  }
  return sink;
}

double profile_guard_manager(
    RootGuardManager* root,
    py::object f_locals,
    int n_iters) {
  PyObject* locals = f_locals.ptr();

  // Warmup to setup fast paths (like dict_tags) for the actual profiling
  for (int i = 0; i < 5; i++) {
    root->check_nopybind(locals);
  }

  std::chrono::duration<double> total_elapsed{0.0};
  for (int i = 0; i < n_iters; i++) {
    // Flush the caches to accurately measure the overhead
    // store into a volatile to prevent optimization
    volatile char dummy = flush_cache_by_eviction();
    (void)dummy;

    auto start = std::chrono::high_resolution_clock::now();
    root->check_nopybind(locals);
    auto end = std::chrono::high_resolution_clock::now();
    total_elapsed += end - start;
  }

  // Calculate the average time per iteration in microseconds
  return (total_elapsed.count() * 1e6) / n_iters;
}

} // namespace

static void* _torchinductor_pyobject_tensor_data_ptr(PyObject* obj) {
  if (C10_UNLIKELY(
          obj == nullptr ||
          (!THPVariable_CheckExact(obj) && !THPVariable_Check(obj)))) {
    throw std::runtime_error(
        "_torchinductor_pyobject_tensor_data_ptr: non-tensor input");
  }
  return THPVariable_Unpack(obj).data_ptr();
}

void* convert_to_root_guard_manager(py::object root) {
  // For invalidated guards, return nullptr
  if (root.is(py::none())) {
    return nullptr;
  }
  RootGuardManager* root_mgr = std::move(root).cast<RootGuardManager*>();
  return (void*)root_mgr;
}

bool run_root_guard_manager(void* root, FrameLocalsMapping* f_locals) {
  // for invalidated guards, return false
  if (root == nullptr) {
    return false;
  }
  return ((RootGuardManager*)root)->check_nopybind(f_locals);
}

PyObject* torch_c_dynamo_guards_init() {
  // initialize TensorGuardsType
  TensorGuardsType.tp_name = "torch._C._dynamo.guards.TensorGuards";
  TensorGuardsType.tp_basicsize = sizeof(TensorGuards);
  TensorGuardsType.tp_itemsize = 0;
  TensorGuardsType.tp_dealloc = (destructor)TensorGuards_dealloc;
  TensorGuardsType.tp_flags = Py_TPFLAGS_DEFAULT;
  TensorGuardsType.tp_doc = "Check properties of a torch.Tensor";
  TensorGuardsType.tp_methods = TensorGuards_methods;
  TensorGuardsType.tp_init = (initproc)TensorGuards_init;
  TensorGuardsType.tp_new = TensorGuards_new;

  if (PyType_Ready(&TensorGuardsType) < 0)
    return nullptr;

  GlobalStateGuardType.tp_name = "torch._C._dynamo.guards.GlobalStateGuard";
  GlobalStateGuardType.tp_basicsize = sizeof(GlobalStateGuard);
  GlobalStateGuardType.tp_itemsize = 0;
  GlobalStateGuardType.tp_flags = Py_TPFLAGS_DEFAULT;
  GlobalStateGuardType.tp_doc = "Guard on PyTorch global flags such as no_grad";
  GlobalStateGuardType.tp_methods = GlobalStateGuard_methods;
  GlobalStateGuardType.tp_init = (initproc)GlobalStateGuard_init;
  GlobalStateGuardType.tp_new = PyType_GenericNew;

  if (PyType_Ready(&GlobalStateGuardType) < 0)
    return nullptr;

  auto m = PyModule_Create(&_module);
  if (m == nullptr)
    return nullptr;

#ifdef Py_GIL_DISABLED
  PyUnstable_Module_SetGIL(m, Py_MOD_GIL_NOT_USED);
#endif

  Py_INCREF(&TensorGuardsType);
  if (PyModule_AddObject(m, "TensorGuards", (PyObject*)&TensorGuardsType) < 0) {
    Py_DECREF(&TensorGuardsType);
    Py_DECREF(m);
    return nullptr;
  }

  Py_INCREF(&GlobalStateGuardType);
  if (PyModule_AddObject(
          m, "GlobalStateGuard", (PyObject*)&GlobalStateGuardType) < 0) {
    Py_DECREF(&GlobalStateGuardType);
    Py_DECREF(m);
    return nullptr;
  }

  // We expose the address of _torchinductor_pyobject_tensor_data_ptr in order
  // to allow manual linking in our generated TorchInductor Python bindings.
  // While regular linking works in most cases, it does not work properly in
  // fbcode due to janky build setup there.
  if (PyModule_AddObject(
          m,
          "_torchinductor_pyobject_tensor_data_ptr",
          PyLong_FromVoidPtr(reinterpret_cast<void*>(
              &_torchinductor_pyobject_tensor_data_ptr))) < 0) {
    return nullptr;
  }

  auto py_m = py::handle(m).cast<py::module>();
  py::class_<GuardDebugInfo, std::unique_ptr<GuardDebugInfo>>(
      py_m, "GuardDebugInfo")
      .def(py::init<bool, py::list, int>())
      .def("__str__", &GuardDebugInfo::to_string)
      .def_readonly("result", &GuardDebugInfo::result)
      .def_readonly("verbose_code_parts", &GuardDebugInfo::verbose_code_parts)
      .def_readonly(
          "num_guards_executed", &GuardDebugInfo::num_guards_executed);

  // Leaf Guards
  py::class_<LeafGuard, std::shared_ptr<LeafGuard>>(py_m, "LeafGuard")
      .def("verbose_code_parts", &LeafGuard::verbose_code_parts);
  py::class_<LAMBDA_GUARD, LeafGuard, std::shared_ptr<LAMBDA_GUARD>>(
      py_m, "LAMBDA_GUARD")
      .def(py::init<py::function, py::list>())
      .def("__call__", &LAMBDA_GUARD::check);
  py::class_<TYPE_MATCH, LeafGuard, std::shared_ptr<TYPE_MATCH>>(
      py_m, "TYPE_MATCH")
      .def(py::init<py::object, py::list>())
      .def("__call__", &TYPE_MATCH::check);
  py::class_<ID_MATCH, LeafGuard, std::shared_ptr<ID_MATCH>>(py_m, "ID_MATCH")
      .def(py::init<py::object, py::list>())
      .def("__call__", &ID_MATCH::check);
  py::class_<NONE_MATCH, LeafGuard, std::shared_ptr<NONE_MATCH>>(
      py_m, "NONE_MATCH")
      .def(py::init<py::list>())
      .def("__call__", &NONE_MATCH::check);
  py::class_<TRUE_MATCH, LeafGuard, std::shared_ptr<TRUE_MATCH>>(
      py_m, "TRUE_MATCH")
      .def(py::init<py::list>())
      .def("__call__", &TRUE_MATCH::check);
  py::class_<FALSE_MATCH, LeafGuard, std::shared_ptr<FALSE_MATCH>>(
      py_m, "FALSE_MATCH")
      .def(py::init<py::list>())
      .def("__call__", &FALSE_MATCH::check);
  py::class_<EQUALS_MATCH, LeafGuard, std::shared_ptr<EQUALS_MATCH>>(
      py_m, "EQUALS_MATCH")
      .def(py::init<py::object, py::list>())
      .def("__call__", &EQUALS_MATCH::check);
  py::class_<LENGTH_CHECK, LeafGuard, std::shared_ptr<LENGTH_CHECK>>(
      py_m, "LENGTH_CHECK")
      .def(py::init<py::object, py::list>())
      .def("__call__", &LENGTH_CHECK::check);
  py::class_<DICT_LENGTH, LeafGuard, std::shared_ptr<DICT_LENGTH>>(
      py_m, "DICT_LENGTH")
      .def(py::init<py::object, py::list>())
      .def("__call__", &DICT_LENGTH::check);
  py::class_<DEFAULT_DEVICE, LeafGuard, std::shared_ptr<DEFAULT_DEVICE>>(
      py_m, "DEFAULT_DEVICE")
      .def(py::init<py::list>())
      .def("__call__", &DEFAULT_DEVICE::check);
  py::class_<NOT_NONE, LeafGuard, std::shared_ptr<NOT_NONE>>(py_m, "NOT_NONE")
      .def(py::init<py::list>())
      .def("__call__", &NOT_NONE::check);
  py::class_<
      MAPPING_KEYS_MATCH,
      LeafGuard,
      std::shared_ptr<MAPPING_KEYS_MATCH>>(py_m, "MAPPING_KEYS_MATCH")
      .def(py::init<py::object, py::list>())
      .def("__call__", &MAPPING_KEYS_MATCH::check);
  py::class_<
      TUPLE_ITERATOR_LEN,
      LeafGuard,
      std::shared_ptr<TUPLE_ITERATOR_LEN>>(py_m, "TUPLE_ITERATOR_LEN")
      .def(py::init<py::object, py::object, py::list>())
      .def("__call__", &TUPLE_ITERATOR_LEN::check);
  py::class_<
      RANGE_ITERATOR_MATCH,
      LeafGuard,
      std::shared_ptr<RANGE_ITERATOR_MATCH>>(py_m, "RANGE_ITERATOR_MATCH")
      .def(py::init<py::object, py::object, py::object, py::object, py::list>())
      .def("__call__", &RANGE_ITERATOR_MATCH::check);
  py::class_<GLOBAL_STATE, LeafGuard, std::shared_ptr<GLOBAL_STATE>>(
      py_m, "GLOBAL_STATE")
      .def(py::init<py::list>())
      .def("check_verbose", &GLOBAL_STATE::check_verbose)
      .def("__call__", &GLOBAL_STATE::check);
  py::class_<
      TORCH_FUNCTION_MODE_STACK,
      LeafGuard,
      std::shared_ptr<TORCH_FUNCTION_MODE_STACK>>(
      py_m, "TORCH_FUNCTION_MODE_STACK")
      .def(py::init<py::list, py::list>())
      .def("__call__", &TORCH_FUNCTION_MODE_STACK::check);
  py::class_<NO_HASATTR, LeafGuard, std::shared_ptr<NO_HASATTR>>(
      py_m, "NO_HASATTR")
      .def(py::init<py::object, py::list>())
      .def("__call__", &NO_HASATTR::check);
  py::class_<DICT_CONTAINS, LeafGuard, std::shared_ptr<DICT_CONTAINS>>(
      py_m, "DICT_CONTAINS")
      .def(py::init<bool, py::object, py::list>())
      .def("__call__", &DICT_CONTAINS::check);
  py::class_<DYNAMIC_INDICES, LeafGuard, std::shared_ptr<DYNAMIC_INDICES>>(
      py_m, "DYNAMIC_INDICES")
      .def(py::init<py::set, py::list>())
      .def("__call__", &DYNAMIC_INDICES::check);
  py::class_<DICT_VERSION, LeafGuard, std::shared_ptr<DICT_VERSION>>(
      py_m, "DICT_VERSION")
      .def(py::init<py::object, py::list>())
      .def("__call__", &DICT_VERSION::check);
  py::class_<
      DISPATCH_KEY_SET_MATCH,
      LeafGuard,
      std::shared_ptr<DISPATCH_KEY_SET_MATCH>>(py_m, "DISPATCH_KEY_SET_MATCH")
      .def(py::init<RootGuardManager*, py::object, py::list>())
      .def("__call__", &DISPATCH_KEY_SET_MATCH::check);
  py::class_<TENSOR_MATCH, LeafGuard, std::shared_ptr<TENSOR_MATCH>>(
      py_m, "TENSOR_MATCH")
      .def(py::init<
           RootGuardManager*,
           py::object,
           py::object,
           py::object,
           py::str,
           py::list,
           py::type,
           py::object>())
      .def("__call__", &TENSOR_MATCH::check);
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<RelationalGuard, LeafGuard, std::shared_ptr<RelationalGuard>>(
      py_m, "RelationalGuard");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      OBJECT_ALIASING,
      RelationalGuard,
      std::shared_ptr<OBJECT_ALIASING>>(py_m, "OBJECT_ALIASING");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      NO_TENSOR_ALIASING,
      RelationalGuard,
      std::shared_ptr<NO_TENSOR_ALIASING>>(py_m, "NO_TENSOR_ALIASING");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      STORAGE_OVERLAPPING,
      RelationalGuard,
      std::shared_ptr<STORAGE_OVERLAPPING>>(py_m, "STORAGE_OVERLAPPING");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      SYMBOLIC_SHAPE_GUARD,
      RelationalGuard,
      std::shared_ptr<SYMBOLIC_SHAPE_GUARD>>(py_m, "SYMBOLIC_SHAPE_GUARD");

  // Guard Accessors - These are present so that we can iterate over the
  // GuardManager hierarchy. We intentionally do not provide even an init
  // function on these, because these should be constructed from within C++.
  py::class_<GuardAccessor, std::unique_ptr<GuardAccessor>>(
      py_m, "GuardAccessor")
      .def("repr", &GuardAccessor::repr);
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GetAttrGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GetAttrGuardAccessor>>(py_m, "GetAttrGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GenericGetAttrGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GenericGetAttrGuardAccessor>>(
      py_m, "GenericGetAttrGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GetGenericDictGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GetGenericDictGuardAccessor>>(
      py_m, "GetGenericDictGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GetItemGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GetItemGuardAccessor>>(py_m, "GetItemGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      FrameLocalsGuardAccessor,
      GuardAccessor,
      std::unique_ptr<FrameLocalsGuardAccessor>>(
      py_m, "FrameLocalsGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      DictGetItemGuardAccessor,
      GuardAccessor,
      std::unique_ptr<DictGetItemGuardAccessor>>(
      py_m, "DictGetItemGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      ListGetItemGuardAccessor,
      GuardAccessor,
      std::unique_ptr<ListGetItemGuardAccessor>>(
      py_m, "ListGetItemGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      TupleGetItemGuardAccessor,
      GuardAccessor,
      std::unique_ptr<TupleGetItemGuardAccessor>>(
      py_m, "TupleGetItemGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      FuncDefaultsGuardAccessor,
      GuardAccessor,
      std::unique_ptr<FuncDefaultsGuardAccessor>>(
      py_m, "FuncDefaultsGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      FuncKwDefaultsGuardAccessor,
      GuardAccessor,
      std::unique_ptr<FuncKwDefaultsGuardAccessor>>(
      py_m, "FuncKwDefaultsGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GlobalsGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GlobalsGuardAccessor>>(py_m, "GlobalsGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      TypeGuardAccessor,
      GuardAccessor,
      std::unique_ptr<TypeGuardAccessor>>(py_m, "TypeGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      WeakRefCallGuardAccessor,
      GuardAccessor,
      std::unique_ptr<WeakRefCallGuardAccessor>>(
      py_m, "WeakRefCallGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      CallFunctionNoArgsGuardAccessor,
      GuardAccessor,
      std::unique_ptr<CallFunctionNoArgsGuardAccessor>>(
      py_m, "CallFunctionNoArgsGuardAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      TupleIteratorGetItemAccessor,
      GuardAccessor,
      std::unique_ptr<TupleIteratorGetItemAccessor>>(
      py_m, "TupleIteratorGetItemAccessor");
  // NOLINTNEXTLINE(bugprone-unused-raii)
  py::class_<
      GlobalWeakRefGuardAccessor,
      GuardAccessor,
      std::unique_ptr<GlobalWeakRefGuardAccessor>>(
      py_m, "GlobalWeakRefGuardAccessor");

  // Guard Manager - No constructor in python, python should use
  // RootGuardManager.
  py::class_<GuardManager, std::unique_ptr<GuardManager>>(py_m, "GuardManager")
      // return by reference because GuardManager has the ownership of accessors
      .def("get_source", &GuardManager::get_source)
      .def("fail_count", &GuardManager::fail_count)
      .def(
          "get_accessors",
          &GuardManager::get_accessors,
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of child
      // managers
      .def(
          "get_child_managers",
          &GuardManager::get_child_managers,
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of leaf
      // guards
      .def(
          "get_leaf_guards",
          &GuardManager::get_leaf_guards,
          py::return_value_policy::reference)
      .def(
          "add_lambda_guard",
          [](GuardManager& self,
             py::object lambda,
             py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<LAMBDA_GUARD>(
                std::move(lambda), std::move(verbose_code_parts)));
          })
      .def(
          "add_type_match_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("TYPE_MATCH");
            self.add_leaf_guard(std::make_shared<TYPE_MATCH>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_id_match_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("ID_MATCH");
            self.add_leaf_guard(std::make_shared<ID_MATCH>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_none_match_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("NONE_MATCH");
            self.add_leaf_guard(
                std::make_shared<NONE_MATCH>(std::move(verbose_code_parts)));
          })
      .def(
          "add_true_match_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("TRUE_MATCH");
            self.add_leaf_guard(
                std::make_shared<TRUE_MATCH>(std::move(verbose_code_parts)));
          })
      .def(
          "add_false_match_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("FALSE_MATCH");
            self.add_leaf_guard(
                std::make_shared<FALSE_MATCH>(std::move(verbose_code_parts)));
          })
      .def(
          "add_equals_match_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("EQUALS_MATCH");
            self.add_leaf_guard(std::make_shared<EQUALS_MATCH>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_length_check_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("LENGTH_CHECK");
            self.add_leaf_guard(std::make_shared<LENGTH_CHECK>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_dict_length_check_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("DICT_LENGTH");
            self.add_leaf_guard(std::make_shared<DICT_LENGTH>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_tuple_iterator_length_guard",
          [](GuardManager& self,
             py::object length,
             py::object type_id,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("TUPLE_ITERATOR_LEN");
            self.add_leaf_guard(std::make_shared<TUPLE_ITERATOR_LEN>(
                std::move(length),
                std::move(type_id),
                std::move(verbose_code_parts)));
          })
      .def(
          "add_range_iterator_match_guard",
          [](GuardManager& self,
             py::object start,
             py::object stop,
             py::object step,
             py::object type_id,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("RANGE_ITERATOR_MATCH");
            self.add_leaf_guard(std::make_shared<RANGE_ITERATOR_MATCH>(
                std::move(start),
                std::move(stop),
                std::move(step),
                std::move(type_id),
                std::move(verbose_code_parts)));
          })
      .def(
          "add_default_device_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<DEFAULT_DEVICE>(
                std::move(verbose_code_parts)));
          })
      .def(
          "add_not_none_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("NOT_NONE");
            self.add_leaf_guard(
                std::make_shared<NOT_NONE>(std::move(verbose_code_parts)));
          })
      .def(
          "add_mapping_keys_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("MAPPING_KEYS_MATCH");
            self.add_leaf_guard(std::make_shared<MAPPING_KEYS_MATCH>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_dispatch_key_set_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("DISPATCH_KEY_SET_MATCH");
            self.add_leaf_guard(std::make_shared<DISPATCH_KEY_SET_MATCH>(
                self.get_root(),
                std::move(value),
                std::move(verbose_code_parts)));
          })
      .def(
          "add_global_state_guard",
          [](GuardManager& self, py::object verbose_code_parts) -> void {
            self.add_leaf_guard(
                std::make_shared<GLOBAL_STATE>(std::move(verbose_code_parts)));
          })
      .def(
          "add_torch_function_mode_stack_guard",
          [](GuardManager& self,
             const py::list& initial_stack,
             py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<TORCH_FUNCTION_MODE_STACK>(
                initial_stack, std::move(verbose_code_parts)));
          })
      .def(
          "add_no_hasattr_guard",
          [](GuardManager& self,
             py::object attr_name,
             py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<NO_HASATTR>(
                std::move(attr_name), std::move(verbose_code_parts)));
          })
      .def(
          "add_dict_contains_guard",
          [](GuardManager& self,
             bool contains,
             py::object key,
             py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<DICT_CONTAINS>(
                contains, std::move(key), std::move(verbose_code_parts)));
          })
      .def(
          "add_dynamic_indices_guard",
          [](GuardManager& self,
             py::set value,
             py::object verbose_code_parts) -> void {
            self.add_leaf_guard(std::make_shared<DYNAMIC_INDICES>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_dict_version_guard",
          [](GuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("DICT_VERSION");
            self.add_leaf_guard(std::make_shared<DICT_VERSION>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_tensor_match_guard",
          [](GuardManager& self,
             py::object value,
             py::object sizes,
             py::object strides,
             py::object tensor_name,
             py::object verbose_code_parts,
             py::object pytype,
             py::object dispatch_keys) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("TENSOR_MATCH");
            self.add_leaf_guard(std::make_shared<TENSOR_MATCH>(
                self.get_root(),
                std::move(value),
                std::move(sizes),
                std::move(strides),
                std::move(tensor_name),
                std::move(verbose_code_parts),
                std::move(pytype),
                std::move(dispatch_keys)));
          })

      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "getitem_manager",
          &GuardManager::get_child_manager<GetItemGuardAccessor>,
          py::arg("key"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "framelocals_manager",
          &GuardManager::get_child_manager<FrameLocalsGuardAccessor>,
          py::arg("key"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "dict_getitem_manager",
          &GuardManager::get_child_manager<DictGetItemGuardAccessor>,
          py::arg("key"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "list_getitem_manager",
          &GuardManager::get_child_manager<ListGetItemGuardAccessor>,
          py::arg("key"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "indexed_manager",
          &GuardManager::get_child_manager<IndexedGuardAccessor>,
          py::arg("idx"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "tensor_property_size_manager",
          &GuardManager::get_child_manager<
              TensorPropertyGuardAccessor<TensorProperty::SIZE>>,
          py::arg("idx"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "tensor_property_stride_manager",
          &GuardManager::get_child_manager<
              TensorPropertyGuardAccessor<TensorProperty::STRIDE>>,
          py::arg("idx"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "tensor_property_storage_offset_manager",
          &GuardManager::get_child_manager<
              TensorPropertyGuardAccessor<TensorProperty::STORAGE_OFFSET>>,
          py::arg("idx"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "tuple_getitem_manager",
          &GuardManager::get_child_manager<TupleGetItemGuardAccessor>,
          py::arg("key"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "func_defaults_manager",
          [](GuardManager& self,
             std::string source,
             py::object example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__defaults_accessor__");
            return self.get_child_manager<FuncDefaultsGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                std::move(example_value),
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)

      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "func_kwdefaults_manager",
          [](GuardManager& self,
             std::string source,
             py::object example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__kwdefaults_accessor__");
            return self.get_child_manager<FuncKwDefaultsGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                std::move(example_value),
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "globals_dict_manager",
          &GuardManager::get_child_manager<GlobalsGuardAccessor>,
          py::arg("f_globals"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "type_manager",
          [](GuardManager& self,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__type_accessor__");
            return self.get_child_manager<TypeGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "weakref_call_manager",
          [](GuardManager& self,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__weakref_call_accessor__");
            return self.get_child_manager<WeakRefCallGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "call_function_no_args_manager",
          [](GuardManager& self,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__call_function_no_args_accessor__");
            return self.get_child_manager<CallFunctionNoArgsGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "tuple_iterator_getitem_manager",
          &GuardManager::get_child_manager<TupleIteratorGetItemAccessor>,
          py::arg("index"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "global_weakref_manager",
          &GuardManager::get_child_manager<GlobalWeakRefGuardAccessor>,
          py::arg("global_name"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "lambda_manager",
          &GuardManager::get_child_manager<PythonLambdaGuardAccessor>,
          py::arg("python_lambda"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "grad_manager",
          [](GuardManager& self,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__grad_accessor__");
            return self.get_child_manager<GradGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "get_generic_dict_manager",
          [](GuardManager& self,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            // A unique key is used to save as the accessor key.
            py::str unique_key("__generic_dict_accessor__");
            return self.get_child_manager<GetGenericDictGuardAccessor>(
                std::move(unique_key),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because C++ GuardManager has the ownership of
      // accessors and guard managers
      .def(
          "generic_getattr_manager",
          &GuardManager::get_child_manager<GenericGetAttrGuardAccessor>,
          py::arg("attr"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because C++ GuardManager has the ownership of
      // accessors and guard managers
      .def(
          "getattr_manager",
          &GuardManager::get_child_manager<GetAttrGuardAccessor>,
          py::arg("attr"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference);

  // Root Guard Manager
  py::class_<RootGuardManager, GuardManager, std::unique_ptr<RootGuardManager>>(
      py_m, "RootGuardManager")
      .def(py::init<>())
      .def("check", &RootGuardManager::check)
      .def("check_verbose", &RootGuardManager::check_verbose)
      .def(
          "clone_manager",
          &RootGuardManager::clone_manager,
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of leaf
      // guards
      .def(
          "get_epilogue_lambda_guards",
          &RootGuardManager::get_epilogue_lambda_guards,
          py::return_value_policy::reference)
      .def(
          "add_epilogue_lambda_guard",
          [](RootGuardManager& self,
             py::object lambda,
             py::object verbose_code_parts) -> void {
            self.add_epilogue_lambda_guard(std::make_unique<LAMBDA_GUARD>(
                std::move(lambda), std::move(verbose_code_parts)));
          });

  // Dict Guard Manager
  py::class_<DictGuardManager, GuardManager, std::unique_ptr<DictGuardManager>>(
      py_m, "DictGuardManager")
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "get_key_manager",
          [](DictGuardManager& self,
             py::object index,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            return self.get_key_manager(
                std::move(index),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("index"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "get_value_manager",
          [](DictGuardManager& self,
             py::object index,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            return self.get_value_manager(
                std::move(index),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("index"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference)
      // return by reference because GuardManager has the ownership of leaf
      // guards
      .def(
          "get_key_value_managers",
          &DictGuardManager::get_key_value_managers,
          py::return_value_policy::reference)
      // Skipped leaf guards
      .def("add_type_match_guard", &DictGuardManager::skip_adding_guard)
      .def("add_dict_length_check_guard", &DictGuardManager::skip_adding_guard)
      // Permitted leaf guards
      .def(
          "add_dict_contains_guard",
          [](DictGuardManager& self,
             bool contains,
             py::object key,
             py::object verbose_code_parts) -> void {
            self.add_permitted_leaf_guard(std::make_shared<DICT_CONTAINS>(
                contains, std::move(key), std::move(verbose_code_parts)));
          })
      .def(
          "add_dict_version_guard",
          [](DictGuardManager& self,
             py::object value,
             py::object verbose_code_parts) -> void {
            SKIP_IF_GUARD_ALREADY_PRESENT("DICT_VERSION");
            self.add_permitted_leaf_guard(std::make_shared<DICT_VERSION>(
                std::move(value), std::move(verbose_code_parts)));
          })
      .def(
          "add_no_hasattr_guard",
          [](DictGuardManager& self,
             py::object attr_name,
             py::object verbose_code_parts) -> void {
            self.add_permitted_leaf_guard(std::make_shared<NO_HASATTR>(
                std::move(attr_name), std::move(verbose_code_parts)));
          })
      // Not permitted accessors
      .def("lambda_manager", &DictGuardManager::fail_on_get_child_manager)
      .def("getitem_manager", &DictGuardManager::fail_on_get_child_manager)
      .def("dict_getitem_manager", &DictGuardManager::fail_on_get_child_manager)
      .def("globals_dict_manager", &DictGuardManager::fail_on_get_child_manager)
      .def(
          "tuple_iterator_getitem_manager",
          &DictGuardManager::fail_on_get_child_manager)
      .def(
          "global_weakref_manager",
          &DictGuardManager::fail_on_get_child_manager)
      .def("lambda_manager", &DictGuardManager::fail_on_get_child_manager)
      // Permitted accessors (and also type_manager)
      // return by reference because GuardManager has the ownership of accessors
      // and guard managers
      .def(
          "getattr_manager",
          [](DictGuardManager& self,
             py::object attr_name,
             std::string source,
             py::handle example_value,
             py::handle guard_manager_enum) -> GuardManager* {
            if (self.is_exact_dict_type()) {
              throw std::runtime_error(
                  "getattr_manager on a DictGuardManager is supported only for dict subclasses");
            }
            return self.get_child_manager<GetAttrGuardAccessor>(
                std::move(attr_name),
                std::move(source),
                example_value,
                guard_manager_enum);
          },
          py::arg("attr"),
          py::arg("source"),
          py::arg("example_value"),
          py::arg("guard_manager_enum"),
          py::return_value_policy::reference);

  py_m.def("install_object_aliasing_guard", install_object_aliasing_guard);
  py_m.def(
      "install_no_tensor_aliasing_guard", install_no_tensor_aliasing_guard);
  py_m.def(
      "install_storage_overlapping_guard", install_storage_overlapping_guard);
  py_m.def(
      "compute_overlapping_tensors",
      [](const std::vector<Tensor> tensors, bool symbolic) {
        // Pick the correct Meta class, depending on whether we are
        // dealing with symbolic values or not.
        if (symbolic) {
          return compute_overlapping_tensors<DynamicMeta>(tensors);
        } else {
          return compute_overlapping_tensors<StaticMeta>(tensors);
        }
      },
      py::arg("tensors"),
      py::arg("symbolic") = true);
  py_m.def("install_symbolic_shape_guard", install_symbolic_shape_guard);
  py_m.def("profile_guard_manager", profile_guard_manager);

// initialize dict_version_map watcher for 3.12
#if IS_PYTHON_3_12_PLUS

  dict_version_watcher_id = PyDict_AddWatcher(dict_version_watch_callback);
  if (dict_version_watcher_id == -1) {
    throw std::runtime_error("Failed to install dict_version_watch_callback");
  }

#endif

  return m;
}

} // namespace torch::dynamo