module {
  func.func @main(%arg0: tensor<1x2xi32>) -> tensor<1x2x2xi32> {
  %0 = "tfl.custom"(%arg0) {custom_code = "stablehlo.broadcast_in_dim", custom_option = #tfl<const_bytes : "0x62726F6164636173745F64696D656E73696F6E73000201020119010101072C022401">} : (tensor<1x2xi32>) -> tensor<1x2x2xi32>
  func.return %0 : tensor<1x2x2xi32>
  }
}

// CHECK:      module {
// CHECK-NEXT:  func @main(%arg0: tensor<1x2xi32>) -> tensor<1x2x2xi32> {

// CHECK: return %[[GATHER]] : tensor<1x300x10xf32>

// -----

module {
  func.func @propagate_xlagather(%arg0: tensor<10x2xi32>) -> tensor<1x300x10xf32> {
    %cst = "tf.Const"() {value = dense<[1, 1, 300]> : tensor<3xi64>} : () -> tensor<3xi64>

func.func @main(%arg0: tensor<10xi32>, %arg1: tensor<10xi32>) -> tensor<10xi32>
attributes {tf.entry_function = {inputs = "input0,input1", outputs = "Add"}} {
  %graph = tf_executor.graph {
    %2:2 = tf_executor.island wraps "tf.Add"(%arg0, %arg1) {T = "tfdtype$DT_INT32", device = ""} : (tensor<10xi32>, tensor<10xi32>) -> tensor<10xi32> loc("Add")
    tf_executor.fetch %2 : tensor<10xi32>
  }
  func.return %graph : tensor<10xi32>
}

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

func.func @xla_func_0(%arg0: tensor<7xf32>, %arg1: tensor<f32>) -> tensor<10xf32> {
  %cst = "tf.Const"() {value = dense<0> : tensor<1x2xi32>} : () -> tensor<1x2xi32>
  %cst_0 = "tf.Const"() {value = dense<1> : tensor<1xi32>} : () -> tensor<1xi32>
  %cst_1 = "tf.Const"() {value = dense<2> : tensor<1xi32>} : () -> tensor<1xi32>

  func.return
}

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

  // CHECK-NEXT:     %[[GROUP:.*]] = "tf.Const"()
  // CHECK-SAME:       [0, 1, 2, 3]
  // CHECK-NEXT:     %[[REDUCED:.*]] = "tf.XlaAllReduce"(%[[ID]], %[[GROUP]]) <{mode = "CrossReplica", reduce_op = "Add"}> : (tensor<f32>, tensor<1x4xi32>) -> tensor<f32>
  // CHECK-NEXT:     "tf.OpA"(%[[REDUCED]]) : (tensor<f32>) -> ()
  tf_device.replicate {n = 4 : i32} {
    "tf_device.cluster"() ({
      "tf.OpA"(%arg0) : (tensor<f32>) -> ()
      tf_device.return

  // CHECK-LABEL @lowers_string_ops
  // CHECK-DAG: ml_program.global public @vars.Variable_1([]) : tensor<!tf_type.string>
  func.func @lowers_string_ops(%arg0: tensor<128xi32>, %arg1: tensor<128xi32>, %arg2: tensor<128x1xi32>, %arg3: tensor<128x90xi32>, %arg4: tensor<128x90xi32>, %arg5: tensor<128x90xi32>, %arg6: tensor<128x90x64xf32>, %arg7: tensor<128x90x64xf32>) -> tensor<!tf_type.string> {

// CHECK-NO-UNPACK-SAME: (%[[ARG_0:.+]]: tensor<1x1024xf32>) -> tensor<1x3xf32>
// CHECK-NO-UNPACK-DAG: %[[CONST:.+]] = stablehlo.constant() <{value = dense<{{.*}}> : tensor<1024x3xi8>}> : () -> tensor<1024x3x!quant.uniform<i8<-127:127>:f32:1, {{.*}}>>
// CHECK-NO-UNPACK: %[[QUANTIZE_0:.+]] = stablehlo.uniform_quantize %[[ARG_0]] : (tensor<1x1024xf32>) -> tensor<1x1024x!quant.uniform<i8:f32, {{.*}}>>

        LinkedHashMap<ChecksumAlgorithm, ChecksumCalculator> algMap = new LinkedHashMap<>();
        algorithms.forEach(f -> algMap.put(f, f.getCalculator()));
        final byte[] buffer = new byte[1024 * 32];
        for (; ; ) {
            int read = stream.read(buffer);
            if (read < 0) {
                break;
            }
            for (ChecksumCalculator checksumCalculator : algMap.values()) {