type: DT_INT32
      }
    }
  }
  attr {
    key: "output_shapes"
    value {
      list {
        shape {
        }
      }
    }
  }
}
node {
  name: "MapDataset"
  op: "MapDataset"
  input: "TensorSliceDataset"
  attr {
    key: "Targuments"
    value {
      list {
      }
    }
  }
  attr {
    key: "f"
    value {
      func {

    // CHECK: tfrt_fallback_async.executeop key({{[0-9]+}}) cost({{.*}}) device("/device:CPU:0") "tf.MapDataset"({{.*}}) {Targuments = [], metadata = "", output_shapes = [#corert.shape<>], output_types = [i64], preserve_cardinality = true, use_inter_op_parallelism = true} {f = "__inference_Dataset_map_lambda_16"} : 1

    // CHECK: tfrt_fallback_async.executeop key({{[0-9]+}}) cost({{.*}}) device("/device:CPU:0") "tf.MapDataset"({{.*}}) {Targuments = [], metadata = "", output_shapes = [#corert.shape<>], output_types = [i64], preserve_cardinality = true, use_inter_op_parallelism = true} {f = "funcC"} : 1

  // CHECK: "tf.MapAndBatchDataset"(%[[TSLICE]], %[[BSIZE:.*]], %[[NPC]]
  // CHECK-SAME: f = @"__inference_Dataset_map_<lambda>_80",
  %4 = "tf.MapDataset"(%3) {device = "",
           f = @"__inference_Dataset_map_<lambda>_80",
           output_shapes = [#tf_type.shape<>], output_types = [i32],
           preserve_cardinality = false, sloppy = false,

  } : (tensor<!tf_type.variant>) -> tensor<!tf_type.variant>

  func.return %0 : tensor<!tf_type.variant>
// CHECK: "tf.MapDataset"(
// CHECK-SAME: <{f = @{{.*}}, metadata = "", output_shapes = [#tf_type.shape<>], output_types = [!tf_type.string], preserve_cardinality = true, use_inter_op_parallelism = true}> {Targuments = []}
}

  // CHECK: tf_mlrt.executeop{{.*}}op: \22MapDataset\22
  // CHECK-SAME: \22__inference_Dataset_map_lambda_16\22

  // MapDatasetOp's relationship with its associated function is as
  // follows: first M function params are dictated by the set
  // output shapes and types, the next N are the last Ninputs from MapDataset
  // op. The MapDataset op always has N+1 inputs.
  // TODO(jpienaar): Avoid this lookup.
  auto module = op->getParentOfType<ModuleOp>();
  auto f = module.lookupSymbol<func::FuncOp>(op.getF());

    // For this test case the first 12 inputs to the function f below will be
    // from ParallelMapDataset, while the remaining one is a captured input.
    %75 = "tf.MapDataset"(%74, %arg3) {
      device = "/job:localhost/replica:0/task:0/device:CPU:0",
      f = @__inference_Dataset_map_lambda_96150,
      f._tf_data_function = true,

}

def TF_FlatMapDatasetOp : TF_Op<"FlatMapDataset", [Pure]> {
  let summary = [{
Creates a dataset that applies `f` to the outputs of `input_dataset`.
  }];

  let description = [{
Unlike MapDataset, the `f` in FlatMapDataset is expected to return a
Dataset variant, and FlatMapDataset will flatten successive results
into a single Dataset.
  }];

  let arguments = (ins