Question

我正在使用单词嵌入训练CNN，出于某种原因，每当我尝试保存模型的冻结版本供以后使用时，我都会遇到FailedPreconditionError异常。

尽管我在训练之前打电话给sess.run(tf.global_variables_initializer())，但我没有问题训练和检查模型。

当我尝试从检查点加载模型并保存冻结模型时，会出现问题。我正在使用的功能如下：

def freeze_model(checkpoint_path, model_save_path, output_node_names):
    checkpoint = tf.train.get_checkpoint_state(checkpoint_path)
    input_checkpoint = checkpoint.model_checkpoint_path

    saver = tf.train.import_meta_graph(input_checkpoint + '.meta', clear_devices=True)
    graph = tf.get_default_graph()
    input_graph_def = graph.as_graph_def()
    with tf.Session() as sess:
        saver.restore(sess, input_checkpoint)

        output_graph_def = graph_util.convert_variables_to_constants(
            sess,
            input_graph_def,
            output_node_names
        )

        with tf.gfile.GFile(model_save_path, "wb") as f:
            f.write(output_graph_def.SerializeToString())

我得到的错误是：

Traceback (most recent call last):
  File "myproject/train.py", line 522, in <module>
    tf.app.run()
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 44, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "myproject/train.py", line 518, in main
    trainer.save_model(preprocessor)
  File "myproject/train.py", line 312, in save_model
    ut.freeze_model(self.checkpoint_dir, model_save_path, C.OUTPUT_NODE_NAMES)
  File "/home/foo/anaconda2/lib/python2.7/site-packages/myproject/utils.py", line 224, in freeze_model
    output_node_names
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/framework/graph_util_impl.py", line 218, in convert_variables_to_constants
    returned_variables = sess.run(variable_names)
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 767, in run
    run_metadata_ptr)
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 965, in _run
    feed_dict_string, options, run_metadata)
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1015, in _do_run
    target_list, options, run_metadata)
  File "/home/foo/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1035, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value embeddings/W
         [[Node: embeddings/W/_20 = _Send[T=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_30_embeddings/W", _device="/job:localhost/replica:0/task:0/gpu:0"](embeddings/W)]]
         [[Node: conv_maxpool_4/W/_17 = _Recv[_start_time=0, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_26_conv_maxpool_4/W", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/cpu:0"]()]]

Answer 1

事实证明我在制作Saver之前构建了一个Session对象，因此会话中没有任何内容被保存。

尝试保存冻结模型时Tensorflow FailedPreconditionError

1 个答案: