Hey everyone! I have a Python application using Tkinter that I have
bundled with PyInstaller. It has a number of large libraries such as
tensorflow, torch, and so on. It's been very challenging to get everything
in the right place.
When I finally got things working, one of my first issues I
Sadly, there are about 250 matches to that call within all the libraries
packaged in my application! I did some more logging. The last time
subprocess.run() gets called before a second instance of my application is
initialized is in the following custom logging I did...
*** check_output command
modify that function to behave
differently if it recognizes that it's an executable packaged by
PyInstaller. I saw something similar in a package called cpuinfo, where
they specifically mention Pyinstaller issues, but not exactly sure how to
adapt it to numpy.
On Sun, Sep 24, 2023 at 11:55 AM Chri
Thanks so much for the response! I have narrowed it down to the ultralytics
library but I haven't been able to identify where in that code it's calling
a subprocess. I'll try to investigate some more.
On Sun, Sep 24, 2023, 1:59 AM bwoodsend wrote:
> I’ve read some about how sys.executable() is
Ok, looks like I'll have to do some more digging.
The other issue I'm running into has to do with TensorFlow. My goal is to
create a single windowed executable with no console. Unfortunately, when I
package my application without a console, I run into this error:
WARNING - AutoGraph is not