func.return %0#0 : tensor<*xf32>
}

// CHECK-LABEL: @RewriteSendRecvOps

  func.return %0 : tensor<2x?x?x1xf32>
  // CHECK-LABEL: einsum_matmul_dynamic_size
  // CHECK-DAG: %[[cst:.*]] = arith.constant dense<[2, -1, 1, 1]> : tensor<4xi64>
  // CHECK: %[[v0:.*]] = "tf.Reshape"(%arg1, %cst) : (tensor<2x?xf32>, tensor<4xi64>) -> tensor<2x?x1x1xf32>
  // CHECK: %[[v1:.*]] = "tf.BatchMatMulV2"(%arg0, %0) <{adj_x = false, adj_y = false}> : (tensor<2x?x?x?xf32>, tensor<2x?x1x1xf32>) -> tensor<2x?x?x1xf32>

  func.func @init_all_tables()
  attributes {tf_saved_model.exported_names = ["__tf_saved_model_session_initializer"]} {
    %cst = arith.constant dense<[1, 2, 3, 4]> : tensor<4xi64>
    %cst_0 = arith.constant dense<["a", "b", "c", "d"]> : tensor<4x!tf_type.string>

func.func @typeNotSupportedNotLegalized(%arg0: tensor<!tf_type.variant<tensor<*xf64>>>, %arg1: tensor<i32>, %arg2: tensor<*xf64>) -> tensor<!tf_type.variant<tensor<*xf64>>> {
  %0 = "tf.TensorListSetItem"(%arg0, %arg1, %arg2) : (tensor<!tf_type.variant<tensor<*xf64>>>, tensor<i32>, tensor<*xf64>) -> tensor<!tf_type.variant<tensor<*xf64>>>
  // CHECK-NOT: "tfl.custom"

  %1 = mhlo.constant dense<0.000000e+00> : tensor<f32>
  %2 = "mhlo.pad"(%arg0, %1) {edge_padding_high = dense<[0, 0, 1, 1]> : tensor<4xi64>, edge_padding_low = dense<[0, 0, 1, 1]> : tensor<4xi64>, interior_padding = dense<0> : tensor<4xi64>} : (tensor<1x3x6x6xf32>, tensor<f32>) -> tensor<1x3x8x8xf32>
  %3 = "mhlo.reduce_window"(%2, %1) ({
  ^bb0(%arg1: tensor<f32>, %arg2: tensor<f32>):

  // CHECK-DAG: %[[MATMUL_RHS_SHAPE:.*]] = "tf.Const"() <{value = dense<[5, 6]> : tensor<2xi64>}>
  // CHECK-DAG: %[[RESULT_SHAPE:.*]] = "tf.Const"() <{value = dense<[2, 3, 4, 6]> : tensor<4xi64>}>

  // CHECK: %[[LHS_RESHAPED:.*]] = "tf.Reshape"(%arg0, %[[LHS_RESHAPED_SHAPE]]) : (tensor<2x3x4x5xf32>, tensor<3xi64>) -> tensor<6x4x5xf32>

  // CHECK-LABEL: init_all_tables
  func.func @init_all_tables()
  attributes {tf_saved_model.exported_names = ["__tf_saved_model_session_initializer"]} {
    %cst = arith.constant dense<[1, 2, 3, 4]> : tensor<4xi64>
    %cst_0 = arith.constant dense<["a", "b", "c", "d"]> : tensor<4x!tf_type.string>

// CHECK-LABEL: broadcast_to_to_reshape_i64
func.func @broadcast_to_to_reshape_i64(%arg0: tensor<4x4x4xf32>, %arg1 : tensor<4xi64>) -> tensor<1x4x4x4xf32> {
  %0 = "tfl.broadcast_to"(%arg0, %arg1) : (tensor<4x4x4xf32>, tensor<4xi64>) -> tensor<1x4x4x4xf32>
  // CHECK: "tfl.cast"
  // CHECK-SAME: (tensor<4xi64>) -> tensor<4xi32>
  // CHECK-NEXT: "tfl.reshape"
  // CHECK-SAME: (tensor<4x4x4xf32>, tensor<4xi32>) -> tensor<1x4x4x4xf32>

  %1 = mhlo.constant dense<[[[[0, 1, 2, 3]], [[0, 1, 2, 3]]]]> : tensor<1x2x1x4xi32>
  %2 = "mhlo.broadcast_in_dim"(%0) <{broadcast_dimensions = dense<[0, 1, 2, 3]> : tensor<4xi64>}> : (tensor<1x1x1x4xi32>) -> tensor<1x2x1x4xi32>
  %3 = mhlo.multiply %1, %2 : tensor<1x2x1x4xi32>
  // CHECK:      return %[[RES]] : tensor<1x2x1x4xi32>
  return %3 : tensor<1x2x1x4xi32>
}

  func.return %2 : tensor<?x?x?x?xf32>

// CHECK-DAG: %[[CST:.*]] = arith.constant dense<0> : tensor<4xi64>
// CHECK-DAG: %[[CST0:.*]] = arith.constant dense<[1, 9, 23, 768]> : tensor<4xi64>
// CHECK-DAG: %[[CST1:.*]] = arith.constant dense<1> : tensor<4xi64>