96e4af6e86e93da021c29a4273dcdfaeab8fdfcf,tensorflow_transform/beam/impl.py,_RunMetaGraphDoFn,process,#_RunMetaGraphDoFn#Any#Any#,262

Before Change

      A representation of output features as a dict mapping keys (logical column
      names) to values.
    
    if self._graph_state is None:
      if (getattr(self._thread_local, "graph_state", None) is None or
          self._thread_local.graph_state.saved_model_dir != saved_model_dir):
        start = datetime.datetime.now()
        tf_config = _maybe_deserialize_tf_config(self._serialized_tf_config)
        self._thread_local.graph_state = self._GraphState(
            saved_model_dir, self._input_schema, self._output_schema, tf_config)
        self._graph_load_seconds_distribution.update(
            int((datetime.datetime.now() - start).total_seconds()))
      self._graph_state = self._thread_local.graph_state
    else:
      assert self._graph_state.saved_model_dir == saved_model_dir

    self._batch.append(element)
    if len(self._batch) >= self._desired_batch_size:
      yield self._flush_batch()

After Change

    if self._graph_state is None:
      // If available, acquire will return a cached _GraphState, since calling
      // _make_graph_state is expensive.
      self._graph_state = self._shared_graph_state_handle.acquire(
          lambda: self._make_graph_state(saved_model_dir))

      self._session = tf.Session(
          graph=self._graph_state.graph, config=self._graph_state.tf_config)
      with self._session.graph.as_default():

Instances