%3 = "quantfork.qcast"(%arg0) {volatile} : (tensor<1x2xf32>) -> tensor<1x2x!quant.uniform<i8:f32, 6.000000e-03:-128>>
   %4 = "quantfork.dcast"(%3) : (tensor<1x2x!quant.uniform<i8:f32, 6.000000e-03:-128>>) -> tensor<1x2xf32>
// expected-error @+2 {{Failed to find a valid entry function}}

    %7 = call @uniform_quantize_0(%arg0, %0, %1) : (tensor<1x2xf32>, tensor<1x1xf32>, tensor<1x1xi8>) -> tensor<1x2xi8>
    %8 = stablehlo.convert %7 : (tensor<1x2xi8>) -> tensor<1x2xf32>
    %9 = stablehlo.convert %2 : (tensor<2x3xi8>) -> tensor<2x3xf32>
    %10 = stablehlo.dot_general %8, %9, contracting_dims = [1] x [0] : (tensor<1x2xf32>, tensor<2x3xf32>) -> tensor<1x3xf32>

      %%arg1 = "tf.Const"() {value = dense<-3.0> : tensor<1x2x4xf32>} : () -> tensor<1x2x4xf32>

      %%1 = "tf.%s"(%%arg0, %%arg1) {T = f32, adj_x = false, adj_y = false, grad_x = false, grad_y = false, device = ""} : (tensor<1x4x2xf32>, tensor<1x2x4xf32>) -> tensor<1x4x4xf32>

      func.return %%1 : tensor<1x4x4xf32>
    }
  })";
  std::string mat_mul_method =

    %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-DAG: %[[input_19:.*]] = "quantfork.stats"({{.*}}) <{layerStats = dense<[-2.000000e+00, 4.000000e+00]> : tensor<2xf32>}> : (tensor<1x2xf32>) -> tensor<1x2xf32>

// CHECK: "tfl.unidirectional_sequence_lstm"({{.*}}, %[[input_18]], %[[input_19]], %{{[0-9]+}}, %{{[0-9]+}}, %{{[0-9]+}}, %{{[0-9]+}})

  // CHECK-DAG: %[[ITEMS0_0:.*]] = "tf.ExpandDims"(%[[ITEMS0]], %[[AXIS]])
  // CHECK-DAG: "tf.ConcatV2"(%[[ITEMS1_3]], %[[ITEMS1_2]], %[[ITEMS1_1]], %[[ITEMS1_0]], %[[ITEMS0_0]], %[[AXIS]]) : (tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<1x2xf32>, tensor<i64>) -> tensor<5x2xf32>

  %indices0 = "tf.Const"() {value = dense<4> : tensor<i32>} : () -> tensor<i32>

  %0 = "tf_device.cluster_func"(%arg0, %arg1) {func = @_func, use_spmd_for_xla_partitioning = true, use_tpu = true, num_cores_per_replica = 2 : i64} : (tensor<2x4xf32>, tensor<4x2xf32>) -> tensor<2x2xf32>
  %1:2 = "tf.TPUPartitionedOutputV2"(%0) {device = "", partition_dims = [2, 1]} : (tensor<2x2xf32>) -> (tensor<1x2xf32>, tensor<1x2xf32>)

    "tf_device.cluster"() ({
      %0 = "tf.RecvTPUEmbeddingActivations"() {config = "test_config_recv_embedding"} : () -> tensor<512x256xf32>
      "tf.SendTPUEmbeddingGradients"(%0) {N = 1 : i64, NN = 0 : i64, config = "test_config_send_embedding", operandSegmentSizes = array<i32: 1, 0>} : (tensor<512x256xf32>) -> ()
      "tf.A"() {_xla_outside_compilation = "cluster1"} : () -> ()
      tf_device.return

  // CHECK: [[Result:%.*]] = "tf.Case"(%arg0, %arg1)
  // CHECK-SAME: branches = [@tf.CaseRegion_branch0{{.*}}, @tf.CaseRegion_branch1{{.*}}]
  // CHECK-SAME: is_stateless = false
  %0 = "tf.CaseRegion"(%arg0) ({
    %1 = "tf.Cast"(%arg1) {Truncate = false} : (tensor<!tf_type.resource<tensor<1x2x3xf32>>>) -> tensor<!tf_type.resource>

// CHECK: return %[[v4]] : tensor<?x?x8x?xf32>
}

func.func @einsum_ellipsis(%arg0: tensor<1x512x128xf32>, %arg1: tensor<128x256xf32>) -> tensor<1x512x256xf32> {
  %0 = "tf.Einsum"(%arg0, %arg1) {device = "", equation = "...x,xy->...y"} : (tensor<1x512x128xf32>, tensor<128x256xf32>) -> tensor<1x512x256xf32>
  func.return %0 : tensor<1x512x256xf32>
// CHECK-LABEL: einsum_ellipsis