// CHECK-NEXT:  [[NEG_ONE_1:%.*]] = arith.constant dense<-1> : tensor<i32>
// CHECK-NEXT:  [[RESULT:%.*]] = "tf.Concat"([[ZERO]], [[INPUT]], [[EXTENDED_PART]]) : (tensor<i32>, tensor<3x10xf32>, tensor<?x10xf32>) -> tensor<?x10xf32>
// CHECK-NEXT:  return [[RESULT]] : tensor<?x10xf32>


// CHECK-LABEL:  func private @cond_false

    : ( tensor<1x28x28xf32>,
        tensor<20x28xf32>, tensor<20x28xf32>, tensor<20x28xf32>, tensor<20x28xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        none, none, none,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        none, none,

func.func @main(tensor<40x37xf32>, tensor<40x37xf32>) -> tensor<40x40xf32> {
^bb0(%arg0: tensor<40x37xf32>, %arg1: tensor<40x37xf32>):
  %cst = arith.constant dense<1.0> : tensor<40xf32>
  %0:2 = "tfl.fully_connected"(%arg0, %arg1, %cst) {fused_activation_function = "NONE", keep_num_dims = false, weights_format = "DEFAULT"} : (tensor<40x37xf32>, tensor<40x37xf32>, tensor<40xf32>) -> (tensor<40x40xf32>, tensor<40x40xf32>)
  func.return %0 : tensor<40x40xf32>

  // CHECK: %[[ADD:.*]] = tfl.add %[[VAR_VAL]], %arg0 {fused_activation_function = "NONE"} : tensor<1x10xf32>
  // CHECK: "tfl.assign_variable"(%[[RESOURCE]], %[[ADD]]) : (tensor<!tf_type.resource<tensor<1x10xf32>>>, tensor<1x10xf32>) -> ()
  // CHECK: %[[RESULT:.*]] = "tfl.read_variable"(%[[RESOURCE]]) : (tensor<!tf_type.resource<tensor<1x10xf32>>>) -> tensor<1x10xf32>

func.func @main(%arg0: tensor<40x37xf32>, %arg1: tensor<40x37xf32>) -> tensor<40x40xf32> {
// CHECK: [[NONE:%.*]] = "tfl.no_value"() <{value}> : () -> none
// CHECK: "tfl.fully_connected"(%arg0, %arg1, [[NONE]])
// CHECK-SAME: (tensor<40x37xf32>, tensor<40x37xf32>, none) -> (tensor<40x40xf32>, tensor<40x40xf32>)
  %cst = "tfl.no_value"() {value = unit} : () -> none

    %6 = "tfl.reshape"(%5, %cst_1) {tac.device = "GPU", tac.inference_type = "FLOAT"} : (tensor<200xf32>, tensor<2xi32>) -> tensor<2x100xf32>
    %7 = "tfl.quantize"(%6) {qtype = tensor<2x100x!quant.uniform<i8:f32, 2.000000e-01:-3>>, tac.device = "GPU", tac.inference_type = "FLOAT"} : (tensor<2x100xf32>) -> tensor<2x100x!quant.uniform<i8:f32, 2.000000e-01:-3>>

  %pi_0 = "tf.TPUPartitionedInput"(%arg0, %arg1) {_XlaSharding = "123", partition_dim = 0 : i64} : (tensor<10x3xf32>, tensor<10x3xf32>) -> tensor<20x3xf32>
  %ri = "tf.TPUReplicatedInput"(%pi_0) : (tensor<20x3xf32>) -> tensor<20x3xf32>
  // CHECK: return [[RI]]
  func.return %ri : tensor<20x3xf32>
}

// -----

// CHECK-LABEL:func @partitioned_2d_resource

    : ( tensor<1x28x28xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20x20xf32>, none,
        tensor<1x20xf32>, tensor<1x20xf32>,

  // CHECK-LABEL: @get_next
  func.func private @get_next(%arg0: tensor<!tf_type.resource>) -> tensor<200x10xf32> {
    // CHECK: %0 = "tf.Iterator"
    // CHECK-SAME: shared_name = "foo_iterator"
    // CHECK: "tf.IteratorGetNext"(%0)
    %0 = "tf.IteratorGetNext"(%arg0) : (tensor<!tf_type.resource>) -> tensor<200x10xf32>
    return %0 : tensor<200x10xf32>
  }

  // CHECK-LABEL: @main
  func.func @main()

// CHECK-LABEL: float_matmul
func.func @float_matmul(
  %arg0: tensor<1x10xf32>, %arg1: tensor<10x10xf32>) -> (tensor<*xf32>, tensor<*xf32>, tensor<*xf32>) {
  %cst = "tf.Const"() {value = dense<0.000000e+00> : tensor<10xf32>} : () -> tensor<10xf32>
  %0 = "tf.MatMul"(%arg0, %arg1) {
    transpose_a = false, transpose_b = false
  } : (tensor<1x10xf32>, tensor<10x10xf32>) -> tensor<*xf32>