learned_primal_dual
learned_primal_dual copied to clipboard
Running issue
Hi, I had one issue when running the code in ellipses :
File "C:\Users\ZG\AppData\Local\Continuum\anaconda3\lib\site-packages\tensorflow\python\ops\variable_scope.py", line 848, in _get_single_variable name, "".join(traceback.format_list(tb))))
ValueError: Variable primal_iterate_0/conv2d/kernel already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope? Originally defined at:
Do you know what's the issue? Thanks.
You need to restart the graph between each run. If you don't then it will try creating new variables for each time you run the code (and fail).
See e.g. https://stackoverflow.com/questions/45951037/tensorflow-variable-w3-already-exists-disallowed/45951579 for more information.
Adler, Thanks a lot for your prompt reply. Yes, when restart the spyder, it seems working fine now. And i am running the learned_chambolle_pock demo, hopefully some interesting results will be shown soon.
Another issue is that it is running with CPU, which is kind of slow. It prints at the beginning that: Found no GPUs, doing nothing. Any hint for this? I have GPUs on the machine. Thanks again .
Do you have a GPU installation of tensorflow? There are separate installations for CPU and GPU
That's true, Adler, thanks for your kind help. After install gpu version of tensorflow, it show one error as below when running learned_chambolle_pock demo :
File "C:\Users\ZANGG\AppData\Local\Continuum\anaconda3\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 74, in
ImportError: Traceback (most recent call last):
File "C:\Users\ZANGG\AppData\Local\Continuum\anaconda3\lib\site-packages\tensorflow\python\pywrap_tensorflow.py", line 58, in
Failed to load the native TensorFlow runtime.
See https://www.tensorflow.org/install/errors
for some common reasons and solutions. Include the entire stack trace above this error message when asking for help.
Do you know what i can do for this? Thanks.
Have you read https://www.tensorflow.org/install/errors? This is typically due to missing cuDNN.