// fixed upstream).
func.func @broadcast_add(%arg0: tensor<1xi32>, %arg1: tensor<1x2xi32>) -> tensor<1x2xi32> {
  // CHECK-NEXT: %[[LHS_BCAST:.+]] = "mhlo.broadcast_in_dim"(%arg0) <{broadcast_dimensions = dense<1> : tensor<1xi64>}>
  // CHECK-NEXT: mhlo.add %[[LHS_BCAST]], %arg1
  %0 = "tf.AddV2"(%arg0, %arg1) : (tensor<1xi32>, tensor<1x2xi32>) -> tensor<1x2xi32>
  func.return %0: tensor<1x2xi32>
}

// CHECK-LABEL: func @broadcast_multi_dim_add

    "tf.AssignVariableOp"(%arg0, %3) <{validate_shape = false}> : (tensor<*x!tf_type.resource<tensor<1xf32>>>, tensor<1xf32>) -> ()
    tf_device.return
  }
  func.return
}

func.func private @write_chain_func(%arg0: tensor<1xf32>) -> (tensor<1xf32>) {
  %cst = "tf.Const"() <{value = dense<[[0, 1]]> : tensor<1x2xi32>}> : () -> tensor<1x2xi32>

  func.return %0: tensor<3x2xf32>
}

func.func @sparse_f16() -> tensor<3x2xf16> {
  // CHECK-LABEL: @sparse_f16

  %1 = "quantfork.stats"(%0) {bitsNum = 8 : i64, layerStats = dense<[-2.0, 2.0]> : tensor<2xf32>, narrowRange = false} : (tensor<2x3xf32>) -> tensor<2x3xf32>
  %2 = stablehlo.convert %1 : (tensor<2x3xf32>) -> (tensor<2x3xf32>)
  func.return %2 : tensor<2x3xf32>
}

// -----

// CHECK-LABEL: func @dot_redundant_stats

  func.func @simple_chain(%arg0: tensor<1xf32>) -> tensor<*xf32> {
    // CHECK: %[[MUL:.*]] = "tf.Mul"{{.*}} (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    // CHECK: %[[ADD:.*]] = "tf.Add"(%[[MUL]], %[[MUL]]) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    // CHECK: return %[[ADD]] : tensor<1xf32>
    %0 = "tf.Mul"(%arg0, %arg0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<*xf32>

    %recurrent_stats = "quantfork.stats"(%recurrent_input) {layerStats = dense<[-2.0, 1.0]> : tensor<2xf32>} : (tensor<1x20xf32>) -> tensor<1x20xf32>
    %cell_input = arith.constant dense<1.0> : tensor<1x20xf32>
    %cell_stats = "quantfork.stats"(%cell_input) {layerStats = dense<[-2.73090601, 7.94872093]> : tensor<2xf32>} : (tensor<1x20xf32>) -> tensor<1x20xf32>

  // CHECK-LABEL: @xla_svd
  func.func @xla_svd(%arg0: tensor<1x1xf32>) -> (tensor<1xf32>, tensor<1x1xf32>, tensor<1x1xf32>) {
    // CHECK-NOT: XlaSvd
    %s, %u, %v = "tf.XlaSvd"(%arg0) {max_iter = 1, epsilon = 1.0E-09 : f32, precision_config = ""} : (tensor<1x1xf32>) -> (tensor<1xf32>, tensor<1x1xf32>, tensor<1x1xf32>)
    func.return %s, %u, %v : tensor<1xf32>, tensor<1x1xf32>, tensor<1x1xf32>
  }

  func.func @identity(%arg0: f32) -> f32 {

  %group_assignment = "tf.Const"() { value = dense<[[0, 1]]> : tensor<1x2xi32> } : () -> tensor<1x2xi32>
  %instance_key = "tf.Const"() { value = dense<3> : tensor<i32> } : () -> tensor<i32>
  %group_size, %group_key = "tf.CollectiveAssignGroupV2"(%group_assignment, %rank, %key_base) {} : (tensor<1x2xi32>, tensor<i32>, tensor<i32>) -> (tensor<i32>, tensor<i32>)
  // CHECK-NOT: "tf.CollectiveAssignGroupV2"

        }) : () -> tensor<f32>
        "tf_device.cluster"() ({
          %group = "tf.Const"() <{value = dense<[[0, 1]]> : tensor<1x2xi32>}> : () -> tensor<1x2xi32>
          %arg1_reduced = "tf.XlaAllReduce"(%arg1_id, %group) <{mode = "CrossReplica", reduce_op = "Add"}> : (tensor<f32>, tensor<1x2xi32>) -> tensor<f32>
          "tf.OpA"(%arg1_reduced) : (tensor<f32>) -> ()
          tf_device.return
        }) : () -> ()

func.func @main(%arg0: tensor<5x7xf32>) -> tensor<5x7xf32> {
  func.return %arg0: tensor<5x7xf32>
// CHECK-LABEL: main
// CHECK: return %arg0 : tensor<5x7xf32>
}

// - transpose
//
func.func @transpose_2d(%arg0: tensor<2x3xf32>) -> tensor<3x2xf32> {
  %0 = "mhlo.transpose"(%arg0) <{permutation = dense<[1, 0]> : tensor<2xi64>}> : (tensor<2x3xf32>) -> tensor<3x2xf32>
  func.return %0 : tensor<3x2xf32>