This forum uses cookies
This forum makes use of cookies to store your login information if you are registered, and your last visit if you are not. Cookies are small text documents stored on your computer; the cookies set by this forum can only be used on this website and pose no security risk. Cookies on this forum also track the specific topics you have read and when you last read them. Please confirm whether you accept or reject these cookies being set.

A cookie will be stored in your browser regardless of choice to prevent you being asked this question again. You will be able to change your cookie settings at any time using the link in the footer.

[BUG] VSGAN models produce green frames, and more
#11
Regarding vsgan: might be some driver incompatibility with the torch&co versions used by VSGAN.
Quote:Would it make sense to try an older Hybrid version?
No, since that wouldn't change the torch&vsmlrt files.

Quote:I think in earlier versions Hybrid always generated a new name for the output when using "Output name generation"...
When the file name generation is triggered, Hybrid will check if a file with the current name exists and add '_new' to the file name, until the file name does not exist.
This is the only functionality Hybrid has (and ever had).

Quote:I always use the "Add to queue and start queue" button to start a job
Ah, so it is as I expected. Smile (I never do,..)
I'll do some testing and see whether I can come up with a solution.

Cu Selur
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#12
Send you a link to a dev version (via pm) which ask whether to proceed or abort when create&start is used. (and the warnings are enabled)

Cu Selur
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#13
Selur Wrote:Send you a link to a dev version (via pm) which ask whether to proceed or abort when create&start is used. (and the warnings are enabled)

Hi there, sorry for only getting back now.

Sooo, thanks for the dev version. : ) I like it this way - being able to abort in that warning popup instead of just being warned and the overwrite will happen no matter what. Thanks! : ))

And I do have another question regarding the partially broken filters... For me the only strategy really is to use the R65 version. My system does not seem to be compatible with R68 - sadly. But some filters seem to not be present in the R65 7z? E.g. I'm using ESRGAN often for upscaling, and where in the R68 7z a folder Lib>site-packages>vsrealesrgan is present, it's missing in the R65 7z.

I'd appreciate some advice - merci. : )

Cheers!
Reply
#14
Seems like I forgot to add it to the last R65 version I had lying around.
You can try whether calling:
pyhton -m pip install vsrealesrgan
python -m vsrealesrgan
inside the Hybrid/64bit/Vapoursynth-folder installs it.

Cu Selur
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#15
Selur Wrote:Seems like I forgot to add it to the last R65 version I had lying around.
You can try whether calling:
pyhton -m pip install vsrealesrgan
python -m vsrealesrgan
inside the Hybrid/64bit/Vapoursynth-folder installs it.

Hi there, thank you. I must say I'm not familiar with Python. I just wanna clean up videos. ; )

Just installed Python from the MS Store, and then calling those two commands from the command prompt within Hybrid's Vapoursynth folder seems to install files into the Python folder (AppData>Local>Packages>Python...), so this doesn't work.

Since a complete R65 package might be useful for others with the same R68 issues, maybe you could create it, complete with all filters...? 😇

Thanks!
Reply
#16
Quote:Just installed Python from the MS Store,...
Why? Huh
Hybrid does come with its own portable Python (in the Vapoursynth folder), installing a system-wide Python will only create problems.
=> Deinstall the system-wide Python, and like I ask, call those calls inside the Vapoursynth-folder.

Quote: Since a complete R65 package might be useful for others with the same R68 issues, maybe you could create it, complete with all filters...? 😇
This is unlikely to happen, since current Vapoursynth is R69 now, so I in a few weeks, R68 will be abandoned too.

Cu Selur
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#17
Ah, R65 did not properly set the dependencies, so you might have to additionally call
set PATH="PATH TO Vapoursynth\torch_dependencies\bin\";%PATH%
where you have to adjust 'PATH TO Vapoursynth' before calling 'python -m vsrealesrgan'

Cu Selur
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#18
Selur Wrote:
Quote:Just installed Python from the MS Store,...
Why?  Huh

Because I'm a noob. ; )

Selur Wrote:This is unlikely to happen, since current Vapoursynth is R69 now, so I in a few weeks, R68 will be abandoned too.

Ah nice, maybe I can use Hybrid normally again then. : )

Ok, so I uninstalled Python, I set the path, everything seemed to install fine, but when clicking on Preview this pops up:

Failed to evaluate the script:
Python exception: could not find MARK

Traceback (most recent call last):
File "src\cython\vapoursynth.pyx", line 3115, in vapoursynth._vpy_evaluate
File "src\cython\vapoursynth.pyx", line 3116, in vapoursynth._vpy_evaluate
File "V:\tempPreviewVapoursynthFile14_06_08_617.vpy", line 45, in
clip = RealESRGAN(clip=clip, model=5, device_index=0, denoise_strength=0.00) # 4096x2944
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "V:\_Tools\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\utils\_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "V:\_Tools\Hybrid\64bit\Vapoursynth\Lib\site-packages\vsrealesrgan\__init__.py", line 159, in realesrgan
loadnet = torch.load(model_path, map_location="cpu")
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "V:\_Tools\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\serialization.py", line 1028, in load
return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "V:\_Tools\Hybrid\64bit\Vapoursynth\Lib\site-packages\torch\serialization.py", line 1246, in _legacy_load
magic_number = pickle_module.load(f, **pickle_load_args)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
_pickle.UnpicklingError: could not find MARK
Reply
#19
No clue,... I have never seen that error.
Maybe trying to get R68 working might be the better goal. :/
----
Dev versions are in the 'experimental'-folder of my GoogleDrive, which is linked on the download page.
Reply
#20
Selur Wrote:No clue,... I have never seen that error.
Maybe trying to get R68 working might be the better goal. :/

Bummer. : ( I certainly regret having deleted the old R65 download I got from your site earlier this year.

If you happen to find yourself so utterly bored that you create a new R65 package including ESRGAN, do let me know. ; )
Reply


Forum Jump:


Users browsing this thread: 4 Guest(s)