Spent one hour on repair geometry in the geomagic,
rule of thumb
1. flip to the same normal before moving the imported item
2. then connect them using bridget
3. easy
Spent one hour on repair geometry in the geomagic,
rule of thumb
1. flip to the same normal before moving the imported item
2. then connect them using bridget
3. easy
https://github.com/tensorflow/tensorflow/issues/35403#
using a tf.Tensor
as a Python bool
is not allowed in Graph execution. Use Eager execution or decorate this function with
I solve it by just using Can you please try with !pip install tensorflow-gpu==2.1-rc2
and see if the error still persists. I am not seeing any issue with TF 2.1.Thanks!
Have some problem on compiling netflow, so I think the error resulted by
https://stackoverflow.com/questions/10727568/cannot-convert-parameter-1-from-const-char-to-lpcwstr
n your project, it sounds like all of these macros are referencing the -W versions. This is controlled by the UNICODE
preprocessor macro (which is defined if you choose the "Use Unicode Character Set" project option in Visual Studio). (Some of Microsoft's C and C++ run time library functions also have ANSI and wide versions. Which one you get is selected by the similarly-named _UNICODE
macro that is also defined by that Visual Studio project setting.)
Typically, both of the -A and -W functions exist in the libraries and are available, even if your application is compiled for Unicode. (There are exceptions; some newer functions are available only in "wide" versions.)
If you have a char *
that contains text in the proper ANSI code page, you can call the -A version explicitly (e.g., SetWindowTextA
). The -A versions are typically wrappers that make wide character copies of the string parameters and pass control to the -W versions.
An alternative is to make your own wide character copies of the strings. You can do this with MultiByteToWideChar. Calling it can be tricky, because you have to manage the buffers. If you can get away with calling the -A version directly, that's generally simpler and already tested. But if your char *
string is using UTF-8 or any encoding other than the user's current ANSI code page, you should do the conversion yourself
change Unicode fixed the problem
stuck in tensorflow for few days for PINN.
One of thing to get rid of placeholder is upgrade tensorflow
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()
#tf.compat.v1.disable_eager_execution()
https://stackoverflow.com/questions/55870127/module-tensorflow-has-no-attribute-contrib
https://stackoverflow.com/questions/30429491/how-to-use-cmakedefine-preprocessor-directive-properly
cannot find header, need to make first
find to fix the rendering error when install vtk add word like const
https://stackoverflow.com/questions/19816139/expression-having-type-const-comparevptrs-would-lose-some-const-volatile-quali
stop at here
File "C:\Users\Chiwei\Anaconda3\lib\site-packages\tensorflow\python\framework\ops.py", line 481, in _disallow_in_graph_mode
" this function with @tf.function.".format(task))
OperatorNotAllowedInGraphError: using a `tf.Tensor` as a Python `bool` is not allowed in Graph execution. Use Eager execution or decorate this function with @tf.function.
suspect is the save function and gpu plot will come back later.
hahahahaha I finally solve the problem in COMSOL myself. all these damn things come from the physics induced sequence.
next time cannot converge the model., right click mesh edit the physics induced sequence, then set the minimum order to be smaller then you can solve any model.
https://www.comsol.com/forum/thread/53241/ivar-kjelberg-problem-on-joule-heating-analysis-over-a-thin-geometry
Ivar kjelberg after so many years you are still the best mentor on the forum. thanks for tips.