func.return %0 : tensor<4xf64>
}

func.func @i8() -> tensor<4xi8> {
  // CHECK-LABEL: @i8
  // CHECK: value = dense<[1, 2, 3, 4]> : tensor<4xi8>
  %0 = "tfl.pseudo_const" () { value = dense<[1, 2, 3, 4]> : tensor<4xi8> } : () -> tensor<4xi8>
  func.return %0 : tensor<4xi8>
}

func.func @i16() -> tensor<4xi16> {
  // CHECK-LABEL: @i16

module attributes {tf.versions = {producer = 888 : i32}} {
func.func @testReshapeShapeInference(%arg0: tensor<3x4xi32>) -> tensor<*xi32> {
  %cst = arith.constant dense<[1, 6, 2]> : tensor<3xi32>
  // CHECK: "tfl.reshape"(%arg0, %cst) : (tensor<3x4xi32>, tensor<3xi32>) -> tensor<1x6x2xi32>
  %0 = "tfl.reshape"(%arg0, %cst) : (tensor<3x4xi32>, tensor<3xi32>) -> tensor<*xi32>
  func.return %0 : tensor<*xi32>
}
}

// -----

  // CHECK: %[[WHILE_RESULT:.*]]:2 = mhlo.while(%[[ARG0:.*]] = %[[INT0]], %[[ARG1:.*]] = %[[INT1]])
  // CHECK-SAME: tensor<2x2xi8>, tensor<2x2xi8>

  // CHECK: cond

  // CHECK: do
  // CHECK: mhlo.return %[[ARG0]], %[[ARG1]] : tensor<2x?xi8>, tensor<?x2xi8>

  %2:2 = "tf.WhileRegion"(%0, %1) ({
  ^bb0(%carg0: tensor<2x2x!tf_type.qint8>, %carg1: tensor<2x2x!tf_type.qint8>):

// CHECK: %[[UNIFORM_QUANTIZE:.+]] = stablehlo.uniform_quantize %[[ARG_0]] : (tensor<3x4x2xf32>) -> tensor<3x4x2x!quant.uniform<i8:f32, {{.*}}>>
// CHECK: %[[CALL:.+]] = call @quantized_gather_fn(%[[UNIFORM_QUANTIZE]], %[[CONST]]) {_quantization_method = "static_range_ptq { }"} : (tensor<3x4x2x!quant.uniform<i8:f32, {{.*}}>>, tensor<2x3x2xi32>) -> tensor<2x3x2x2x!quant.uniform<i8:f32, {{.*}}>>

  func.return %0 : tensor<4xf64>
}

func.func @i8() -> tensor<4xi8> {
  // CHECK-LABEL: @i8
  // CHECK: value = dense<[1, 2, 3, 4]> : tensor<4xi8>
  %0 = "tfl.pseudo_const" () { value = dense<[1, 2, 3, 4]> : tensor<4xi8> } : () -> tensor<4xi8>
  func.return %0 : tensor<4xi8>
}

func.func @i16() -> tensor<4xi16> {
  // CHECK-LABEL: @i16

  %cst = "tfl.pseudo_qconst"() {qtype = tensor<1x2x!quant.uniform<u8:f32, 1.0>>, value = dense<-76> : tensor<1x2xi8>} : () -> tensor<1x2x!quant.uniform<u8:f32, 1.0>>

// CHECK-LABEL: @quantize_constant
// CHECK-SAME: %[[ARG0:.*]]: tensor<1x3xf32>
func.func @quantize_constant(%arg0: tensor<1x3xf32>) -> tensor<1x2xf32> {
  // CHECK-DAG: %[[QCST:.*]] = stablehlo.constant() <{value = dense<-78> : tensor<3x2xi8>}> : () -> tensor<3x2x!quant.uniform<i8<-127:127>:f32, 5.000000e-03>>
  // CHECK-DAG: %[[Q1:.*]] = stablehlo.uniform_quantize %[[ARG0]]
  // CHECK-NOT: "quantfork.qcast"
  // CHECK: %[[DOT:.*]] = stablehlo.dot %[[Q1]], %[[QCST]]

    return %0 : tensor<1x3xf32>
  }
}

// CHECK-LABEL: quantize_dot_general_per_tensor
// CHECK-SAME: %[[ARG0:.+]]: tensor<1x2xf32>
// CHECK: %[[CST:.+]] = stablehlo.constant() <{value = dense<127> : tensor<2x3xi8>}> : () -> tensor<2x3x!quant.uniform<i8<-127:127>:f32, 0.0023622048182750312>>

TEST_F(QuantizationReportTest, InitializeWithModuleOp) {
  constexpr absl::string_view kQuantizedDotGeneral = R"mlir(
    func.func @main(%arg0: tensor<1x2xf32>) -> tensor<1x3xf32> {
      %0 = stablehlo.constant() {value = dense<127> : tensor<2x3xi8>} : () -> tensor<2x3x!quant.uniform<i8<-127:127>:f32:1, {1.000000e+0,2.000000e+0,3.000000e+0}>>
      %1 = stablehlo.uniform_quantize %arg0 : (tensor<1x2xf32>) -> tensor<1x2x!quant.uniform<i8:f32, 4.000000e+0>>

  module attributes {tf.versions = {bad_consumers = [], min_consumer = 0 : i32, producer = 268 : i32}} {
  func.func @main() -> () {
    %cst0 = "tf.Const"(){ value = dense<0> : tensor<3x5xi1>} : () -> tensor<3x5xi1>
    %0 = "tf.Where"(%cst0) : (tensor<3x5xi1>) -> tensor<?x2xi64>
    func.return
  }
})";

absl::StatusOr<XlaCompiler::CompilationResult> CompileMlirModule(
    const char* mlir_module_str,