function_aliases = GetFunctionAliases(src_saved_model_path, tags);
  if (!function_aliases.ok()) {
    return absl::InternalError(absl::StrCat(
        "Failed to get function alias: ", function_aliases.status().message()));
  }

  TF_ASSIGN_OR_RETURN(
      OwningOpRef<ModuleOp> module,
      ImportSavedModel(src_saved_model_path, signature_keys, tags,

        }
        py_function_library.SaveExportedModel(
            dst_saved_model_path, *exported_model, src_saved_model_path, tags,
            signature_def_map);

        return absl::OkStatus();
      },
      R"pbdoc(
      Quantizes a model that went through quantization-aware training (QAT)
      saved at `src_saved_model_path`. The resulting model will be saved to

      return True
  return False


# TODO: b/310594193 - Export API to pip package.
def quantize_saved_model(
    src_saved_model_path: str,
    dst_saved_model_path: str,
    config: qc.QuantizationConfig,
) -> None:
  """Quantizes a saved model.

  Args:
    src_saved_model_path: Path to the directory for the source SavedModel.
    dst_saved_model_path: Path to the directory for the destination SavedModel.

    src_saved_model_path = self.create_tempdir().full_path
    dst_saved_model_path = self.create_tempdir().full_path
    signature_def_keys = ['serving_default']
    invalid_quant_opts_object = ('a', 'b', 'c')

    with self.assertRaisesRegex(TypeError, 'incompatible function arguments'):
      pywrap_quantize_model.quantize_ptq_static_range(
          src_saved_model_path,
          dst_saved_model_path,

  // LINT.IfChange(static_range_ptq)
  m.def("static_range_ptq", &PywrapQuantizeStaticRangePtq,
        R"pbdoc(
        Runs static-range post-training quantization (PTQ) on a SavedModel at
        `src_saved_model_path` and saves the resulting model to
        `dst_saved_model_path`.

        The user should pass a serialized `QuantizationConfig` for the
        `quantization_config_serialized` argument, and a signature key ->

      function_aliases = GetFunctionAliases(src_saved_model_path, tags);
  if (!function_aliases.ok()) {
    return absl::InternalError(absl::StrCat(
        "Failed to get function alias: ", function_aliases.status().message()));
  }

  TF_ASSIGN_OR_RETURN(
      auto module,
      ImportSavedModel(src_saved_model_path, signature_keys, tags,

namespace stablehlo::quantization::pywrap {

// Function used by the pywrap_quantization module to mirror
// `::mlir::quant::stablehlo::QuantizeStaticRangePtq`.
absl::Status PywrapQuantizeStaticRangePtq(
    absl::string_view src_saved_model_path,
    absl::string_view dst_saved_model_path, const QuantizationConfig& config,
    const std::vector<std::string>& signature_keys,
    const absl::flat_hash_map<std::string, tensorflow::SignatureDef>&

from tensorflow.compiler.mlir.quantization.tensorflow.python import representative_dataset as rd

# LINT.IfChange(quantize_qat_model)
def quantize_qat_model(
    src_saved_model_path: str,
    dst_saved_model_path: str,
    quantization_options_serialized: bytes,
    *,
    signature_keys: list[str],
    signature_def_map_serialized: dict[str, bytes],

def _run_static_range_qat(
    src_saved_model_path: str,
    dst_saved_model_path: str,
    quant_opts: _QuantizationOptions,
    signature_def_map: _SignatureDefMap,
) -> None:
  """Runs static-range quantization for a Quantization-Aware Trained model.

  Runs the quantization for a model trained using QAT.

  Args:
    src_saved_model_path: Path to the source SavedModel directory.

from tensorflow.compiler.mlir.quantization.tensorflow.python import representative_dataset as rd

# LINT.IfChange(static_range_ptq)
def static_range_ptq(
    src_saved_model_path: str,
    dst_saved_model_path: str,
    quantization_config_serialized: bytes,
    *,
    signature_keys: list[str],
    signature_def_map_serialized: dict[str, bytes],