Question

我使用Lenet5架构定义了深度学习模型。我保存了结果和元数据。我试图重复使用它们来预测新图像。但即使该模型的验证准确度为97.7％，测试精度为95.8％，对新图像的预测精度为0％。

似乎问题是由于模型的不正确恢复而发生的似乎没有找到适当的张量。然后我命名张量来解决问题。最后，我遇到了一个命名问题。

以下是代码：＆lt; click me＆gt;
数据可在此处获得：＆lt; click me＆gt;

痕迹：

Training... model: model1, rate: 0.00085, epochs: 2, batch size: 100, keep_prob: 0.67

Epoch:   0 | cost : 0.631 | Val.accu : 0.816
Epoch:   1 | cost : 0.419 | Val.accu : 0.900
INFO:tensorflow:Restoring parameters from ./logs/nn_logs/171113x1008/model1.ckpt
Validation Accuracy = 0.900
INFO:tensorflow:Restoring parameters from ./logs/nn_logs/171113x1008/model1.ckpt
Test Accuracy = 0.880
< myLabel > = [0, 1, 2, 32, 35]
< repo name of the model > : 171113x1008

我收到以下错误消息：

Traceback (most recent call last):
  File "<ipython-input-1-e08ed0fc9454>", line 1, in <module>
    runfile('C:/Users/mo/_171109_1434_TSC-Debugging.py', wdir='C:/Users/mo/_Coding_')

  File "C:\Users\mo\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 866, in runfile
    execfile(filename, namespace)

  File "C:\Users\mo\Anaconda3\lib\site-packages\spyder\utils\site\sitecustomize.py", line 102, in execfile
    exec(compile(f.read(), filename, 'exec'), namespace)

  File "C:/Users/mo/_171109_1434_TSC-Debugging.py", line 910, in <module>
    myPredictionGRAY, ch = fctPrediction(myImg2, pathCkp)

  File "C:/Users/mo/_171109_1434_TSC-Debugging.py", line 889, in fctPrediction
    metaModel = tf.train.import_meta_graph(Ckp+ckptMeta)

  File "C:\Users\mo\Anaconda3\envs\tfgpu\Lib\site-packages\tensorflow\python\training\saver.py", line 1595, in import_meta_graph
    **kwargs)

  File "C:\Users\mo\Anaconda3\envs\tfgpu\Lib\site-packages\tensorflow\python\framework\meta_graph.py", line 499, in import_scoped_meta_graph
    producer_op_list=producer_op_list)

  File "C:\Users\mo\Anaconda3\envs\tfgpu\Lib\site-packages\tensorflow\python\framework\importer.py", line 308, in import_graph_def
    op_def=op_def)

  File "C:\Users\mo\Anaconda3\envs\tfgpu\Lib\site-packages\tensorflow\python\framework\ops.py", line 2339, in create_op
    self._add_op(ret)

  File "C:\Users\mo\Anaconda3\envs\tfgpu\Lib\site-packages\tensorflow\python\framework\ops.py", line 2033, in _add_op
    "is already used" % op.name)

错误讯息：
ValueError：无法添加名称为layer1 / layer1W / optAdam的op，因为该名称已被使用

感谢您的支持，
莫

为什么恢复元数据和保存的模型失败？

0 个答案: