python - 如何在 v1.13.1 中恢复 tensorflow v1.1.0 保存的模型

Question

我正在尝试恢复提供的预训练模型 here并继续在不同的数据集上进行训练。那里可用的预训练模型是在 tensorflow_gpu-1.1.0 上训练的.但是我有 tensorflow_gpu-1.13.1。当我尝试恢复模型时，出现以下错误。

NotFoundError (see above for traceback): Restoring from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint.

是否可以将模型转换为当前的 tensorflow 版本？我尝试了提供的脚本 here , 但没有运气!

如果无法转换，我也可以使用旧的 tensorflow 版本。但是我也无法正确安装旧版本。github page 中提供的命令在下面

pip install --ignore-installed --upgrade https://storage.googleapis.com/tensorflow/linux/gpu/tensorflow_gpu-1.1.0-cp27-none-linux_x86_64.whl

但是当我使用上述命令安装tensorflow时，出现以下错误

Python 2.7.16 |Anaconda, Inc.| (default, Aug 22 2019, 16:00:36) 
[GCC 7.3.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/__init__.py", line 24, in <module>
    from tensorflow.python import *
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/python/__init__.py", line 51, in <module>
    from tensorflow.python import pywrap_tensorflow
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 52, in <module>
    raise ImportError(msg)
ImportError: Traceback (most recent call last):
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 41, in <module>
    from tensorflow.python.pywrap_tensorflow_internal import *
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
    _pywrap_tensorflow_internal = swig_import_helper()
  File "/media/nagabhushan/Data02/SoftwareFiles/Anaconda/anaconda3/envs/MCnet3/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
ImportError: libcublas.so.8.0: cannot open shared object file: No such file or directory


Failed to load the native TensorFlow runtime.

See https://www.tensorflow.org/install/install_sources#common_installation_problems

for some common reasons and solutions.  Include the entire stack trace
above this error message when asking for help.

如果我使用 conda 安装 tensorflow-1.1.0，导入工作正常，但恢复模型再次失败并出现相同的错误

NotFoundError (see above for traceback): Restoring from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint.

请帮忙!

Answer 1

我真的不确定是否可以移植模型，但我会尝试分享我是如何解决这个问题的。

首先，您应该能够独立于 TensorFlow 版本创建整个图。如果那里发生任何错误，那应该是最小的错误。然后，您可以简单地将所有变量从旧模型复制到新模型:

RESTORE_VARS_BLACKLIST = ['dont', 'load', 'this']
ckpt_vars = tf.train.list_variables(RESTORE_VARS_CKPT)
ass_ops = []
for dst_var in tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES):
    for (ckpt_var, ckpt_shape) in ckpt_vars:
        if dst_var.name.split(":")[0] == ckpt_var and dst_var.shape == ckpt_shape and ckpt_var not in RESTORE_VARS_BLACKLIST:
            value = tf.train.load_variable(RESTORE_VARS_CKPT, ckpt_var)
            ass_ops.append(tf.assign(dst_var, value))
# Run assign in a session
sess.run(ass_ops)

最后，只需保存您的新模型即可。