//
// Implementation steps:
//
//   1) For each function def in the SavedModel, register it with the runtime.
//   2) For each object in the object graph def, build it.
//   3) For each variable stored in the checkpoint in the SavedModel,
//      restore it, and attach it to the associated variable object.
//   4) For each polymorphic function, associate it with the appropriate
//      concrete function(s).

  init_op = variables.global_variables_initializer()
  saver = saver_lib.Saver(write_version=saver_pb2.SaverDef.V1)
  with session.Session() as sess:
    sess.run(init_op)
    sess.run(y.assign(y + 42))
    # Without the checkpoint, the variable won't be set to 42.
    ckpt = os.path.join(out_dir, 'test_graph_tfadd_with_ckpt.ckpt')
    saver.save(sess, ckpt)


def tfadd_with_ckpt_saver(out_dir):

// accordingly, if necessary.
func (c *context) retSpan() (n int, err error) {
	_, nSrc, err := c.ret()
	return nSrc, err
}

// checkpoint sets the return value buffer points for Transform to the current
// positions.
func (c *context) checkpoint() {
	if c.err == nil {
		c.nDst, c.nSrc = c.pDst, c.pSrc+c.sz
	}
}

// unreadRune causes the last rune read by next to be reread on the next

  absl::Status read_status =
      ReadFileToString(Env::Default(), meta_filename, &data);
  if (read_status.ok()) {
    return tensorflow::Fingerprint64(data);
  } else {
    LOG(WARNING) << "Failed to read checkpoint file: " << read_status;
    return 0;
  }
}

uint64_t HashSavedModel(const SavedModel& saved_model) {
  std::string saved_model_serialized;
  {
    // Local scope guarantees coded stream will be trimmed (ensures

  /// SavedObjectGraph from the MetaGraphDef.
  const SavedObjectGraph& saved_object_graph() {
    return meta_graph_def().object_graph_def();
  }

  /// TrackableObjectGraph loaded from the variable_reader() checkpoint.
  TrackableObjectGraph& trackable_object_graph() {
    return trackable_object_graph_;
  }

  /// BundleReader for accessing the variables bundle.
  BundleReader* variable_reader() { return variable_reader_.get(); }

// A pass that creates a new function that wraps the newly created SaveV2 op.
// The new function's name is "tf_quant__save". The function accepts a single
// string tensor as argument, which specifies the path to the checkpoint to
// which the variable's tensor values are saved. It finds
// `tf.AssignVariableOp(tf.VarHandleOp, tf.Const)` pattern in the initializer
// function of type "restore_op" to identify the VarHandleOps that should be

	case isSubpath(requestPath, logsPath):
		// "log" to match other log subresources (pods/log, etc)
		attrs.Subresource = "log"
	case isSubpath(requestPath, checkpointPath):
		attrs.Subresource = "checkpoint"
	}

	klog.V(5).InfoS("Node request attributes", "user", attrs.GetUser().GetName(), "verb", attrs.GetVerb(), "resource", attrs.GetResource(), "subresource", attrs.GetSubresource())

	return attrs

	"k8s.io/kubernetes/pkg/features"
	"k8s.io/kubernetes/pkg/kubelet/cm/devicemanager/checkpoint"
	kubecontainer "k8s.io/kubernetes/pkg/kubelet/container"
)

func TestGetContainerDevices(t *testing.T) {
	podDevices := newPodDevices()
	resourceName1 := "domain1.com/resource1"
	podID := "pod1"
	contID := "con1"
	devices := checkpoint.DevicesPerNUMA{0: []string{"dev1"}, 1: []string{"dev1"}}

    saver_def.set_filename_tensor_name(filename_tensor_name);
    saver_def.set_restore_op_name(restore_op_name);
    // :0 attached to indicate the first result tensor. This saves the model
    // checkpoint when fetched.
    saver_def.set_save_tensor_name(absl::StrCat(save_node_name, ":0"));
    return saver_def;
  } else {
    return absl::InternalError(

	return ft.orderS.OrderedOrEqual(ft.zero, v)
}

// checkpoint saves the current state of known relations.
// Called when descending on a branch.
func (ft *factsTable) checkpoint() {
	if ft.unsat {
		ft.unsatDepth++
	}
	ft.stack = append(ft.stack, checkpointFact)
	ft.limitStack = append(ft.limitStack, checkpointBound)
	ft.orderS.Checkpoint()
	ft.orderU.Checkpoint()
}