"array to apply quantization.")};

  Option<QuantMethod> quantization_method_{
      *this, "quantization-method",
      llvm::cl::init(tensorflow::quantization::QuantizationMethod::
                         METHOD_DYNAMIC_RANGE_INT8),
      llvm::cl::desc("Choose quantization method."),
      llvm::cl::values(
          clEnumValN(tensorflow::quantization::QuantizationMethod::

namespace {

using QuantMethod =
    ::tensorflow::quantization::QuantizationMethod::PresetMethod;
using QuantizationUnit = std::pair<Operation*, int>;
using QuantizationUnits = llvm::SetVector<QuantizationUnit>;
using ::tensorflow::quantization::OpSet;

// Preprocesses ops to allow multi-axis quantization, prior to quantization
// passes. Currently, per-channel quantization only supports 1D results.
class PreprocessOpPass

      llvm::cl::values(
          clEnumValN(tensorflow::quantization::QuantizationMethod::
                         METHOD_STATIC_RANGE_INT8,
                     "ptq", "Post-training static-range quantization"),
          clEnumValN(tensorflow::quantization::QuantizationMethod::
                         METHOD_DYNAMIC_RANGE_INT8,
                     "drq", "Post-training dynamic-range quantizaiton"),

    CustomQuantizationMethod custom_quantization_method = 2;
  }
}

// Preset model quantization method for optimization.
//
// Common quantization methods are defined as preset methods in this message.
// Note that the quantization specs (ex: bit width) for preset quantization
// methods are fixed. To use a different quantization spec for a given method,
// use CustomQuantizationMethod.
// NEXT ID: 2

  // Configures the quantization debugger.
  DebuggerConfig debugger_config = 5;

  // Defines calibration options for quantization. This option is only used for
  // activation of static range quantization (SRQ). Quantization calibration
  // method is set to MIN_MAX by default.
  CalibrationOptions calibration_options = 6;

  // Path to file to save the quantization report, which is essentially a

namespace {

using QuantMethod = tensorflow::quantization::QuantizationMethod::PresetMethod;

// Applies prepare quantization on the model in TF dialect. This pass runs
// before the quantization pass and propagate the quantization parameters
// across ops. This step is necessary for post-training quantization and also
// making the quantization rule for some operations in the quantization-aware
// training quantization simpler.

        ":types",
        "//tensorflow/compiler/mlir/quantization/stablehlo:quantization_config_proto_cc",
        "//tensorflow/compiler/mlir/quantization/tensorflow:exported_model_proto_cc",
        "//tensorflow/compiler/mlir/quantization/tensorflow:passes",
        "//tensorflow/compiler/mlir/quantization/tensorflow/cc:convert_asset_args",
        "//tensorflow/compiler/mlir/quantization/tensorflow/cc:run_passes",

        "//tensorflow/compiler/mlir/quantization/stablehlo/cc:config",
        "//tensorflow/compiler/mlir/quantization/stablehlo/cc:static_range_ptq",
        "//tensorflow/compiler/mlir/quantization/stablehlo/cc:weight_only_ptq",
        "//tensorflow/compiler/mlir/quantization/tensorflow/python:py_function_lib",
        "//tensorflow/compiler/mlir/tensorflow/transforms:tensorflow_passes",

#include "mlir/IR/Types.h"  // from @llvm-project
#include "mlir/Support/LLVM.h"  // from @llvm-project
#include "tensorflow/compiler/mlir/quantization/common/quantization_lib/quantization_utils.h"
#include "tensorflow/compiler/mlir/quantization/tensorflow/quantization_options.pb.h"
#include "tensorflow/compiler/mlir/quantization/tensorflow/utils/tf_quantize_op_utils.h"
#include "tensorflow/compiler/mlir/tensorflow/ir/tf_ops.h"

namespace mlir {

    const ::stablehlo::quantization::QuantizationSpecs& specs,
    const ::stablehlo::quantization::DebuggerConfig& debugger_config);

// Adds passes for static-range quantization post-calibration. Utilizes tensor
// statistics collected from the calibration step and performs quantization.
void AddPostCalibrationPasses(
    OpPassManager& pm,