func.func @tensorlistSetItem(%arg0: tensor<3x10xf32>, %arg1: tensor<1xi32>, %arg2: tensor<i32>, %arg3: tensor<10xf32>) -> tensor<3x10xf32> {
  %0 = "tf.TensorListFromTensor"(%arg0, %arg1) : (tensor<3x10xf32>, tensor<1xi32>) -> tensor<!tf_type.variant<tensor<10xf32>>>
  %1 = "tf.TensorListSetItem"(%0, %arg2, %arg3) : (tensor<!tf_type.variant<tensor<10xf32>>>, tensor<i32>, tensor<10xf32>) -> tensor<!tf_type.variant<tensor<10xf32>>>

// CHECK-NEXT:      %6 = "tfl.pseudo_const"() <{value = dense<1> : tensor<i32>}> : () -> tensor<i32>
// CHECK-NEXT:      %7 = "tfl.unsorted_segment_prod"(%3, %4, %6) : (tensor<4xi32>, tensor<4xi32>, tensor<i32>) -> tensor<1xi32>
// CHECK-NEXT:      %8 = "tfl.unsorted_segment_prod"(%3, %5, %6) : (tensor<4xi32>, tensor<4xi32>, tensor<i32>) -> tensor<1xi32>

  func.return %1 : tensor<1x1x257xi32>
  // CHECK: %[[CONST_DIM:.*]] = "tfl.pseudo_const"() <{value = dense<1> : tensor<1xi32>}> : () -> tensor<1xi32>
  // CHECK: %[[RED:.*]] = "tfl.sum"(%arg0, %[[CONST_DIM]]) <{keep_dims = true}> : (tensor<1x16384x257xi32>, tensor<1xi32>) -> tensor<1x1x257xi32>
}

// CHECK-LABEL: BatchmatmulToReduceSumF32
// CHECK-NOT: "tfl.batch_matmul"

func.func @foo() -> tensor<10xf32> {
  %0 = tf_executor.graph {
    %1:2 = tf_executor.island wraps "tf.Const"() {device = "", dtype = "tfdtype$DT_FLOAT", value = dense<1.000000e+00> : tensor<10xf32>} : () -> tensor<10xf32> loc("const_1")
    tf_executor.fetch %1#0 : tensor<10xf32>
  }
  func.return %0 : tensor<10xf32>
}

// CHECK-DAG: name: "bar"
func.func @bar() -> tensor<10xf32> {
  %0 = tf_executor.graph {

  %q_bias = "quantfork.qcast"(%bias) : (tensor<2xf32>) -> tensor<2x!quant.uniform<i32:f32, 0.044022349891595126>>
  %dq_bias = "quantfork.dcast"(%q_bias) : (tensor<2x!quant.uniform<i32:f32, 0.044022349891595126>>) -> tensor<2xf32>

  %multiples = "tf.Const"() { device = "/device:CPU:0", value = dense<[7,3]> : tensor<2xi32> } : () -> tensor<2xi32>
  // CHECK: T = !corert.string
  %output = "tf.Tile"(%arg, %multiples) { device = "/device:CPU:0" } : (tensor<1x2x!tf_type.string>, tensor<2xi32>) -> tensor<?x6x!tf_type.string>
  func.return %output : tensor<?x6x!tf_type.string>
}

// CHECK-LABEL: func @shape
func.func @shape() {

    %cst = "tf.Const"() {value = dense<-1> : tensor<i64>} : () -> tensor<i64>
    %cst_0 = "tf.Const"() {value = dense<[1, 0]> : tensor<2xi32>} : () -> tensor<2xi32>
    %cst_1 = "tf.Const"() {value = dense<1.000000e+00> : tensor<f32>} : () -> tensor<f32>
    %cst_2 = "tf.Const"() {value = dense<0.00235294132> : tensor<f32>} : () -> tensor<f32>

# CHECK-DAG:       %[[VAL_2:.*]] = arith.constant dense<[1, 0]> : tensor<2xi32>
# CHECK-DAG:       %[[VAL_3:.*]] = "tfl.no_value"() <{value}> : () -> none
# CHECK-DAG:       %[[VAL_6:.*]] = arith.constant dense<0> : tensor<i32>

library {
}
versions {
  producer: 76
}

# Verify that list element shape and dtype are expected.
# CHECK:  tf.TensorListReserve{{.*}}(tensor<2xi32>, tensor<i32>) -> tensor<!tf_type.variant<tensor<*xf32>>>

# Nested variant type.
# CHECK:  tf.TensorListReserve{{.*}}(tensor<2xi32>, tensor<i32>) -> tensor<!tf_type.variant<tensor<*x!tf_type.variant>>>

// PerChannel-DAG: %[[scale:.*]] = "tf.Const"() <{value = dense<{{[0-9\.Ee\+\-]+}}> : tensor<2xf32>}> : () -> tensor<2xf32>
// PerChannel-DAG: %[[zp:.*]] = "tf.Const"() <{value = dense<{{[0-9]+}}> : tensor<2xi32>}> : () -> tensor<2xi32>
// PerChannel: %[[out_1:.*]] = "tf.PartitionedCall"(%arg0, %[[q_w]], %[[scale]], %[[zp]]) <{config = "", config_proto = "", executor_type = "",