Page 1 of 1

Error Logged While Extracting - 32bit limitation?

Posted: Fri Sep 23, 2022 1:58 pm
by ianstephens

We have stumbled upon a brand new error(s) I've not yet seen before in the extract process.

Code: Select all

09/22/2022 20:40:29 INFO     Initialized BiSeNet - Face Parsing (Mask) with batchsize of 8
Running pass 1 of 1: Extraction:  34%|████████████▊                         | 48925/144519 [6:12:05<8:17:28,  3.20it/s]
C:\Convert\plugins\extract\align\_base.py:584: RuntimeWarning: invalid value encountered in power
  size = ((roi[1][0] - roi[0][0]) ** 2 + (roi[1][1] - roi[0][1]) ** 2) ** 0.5
Running pass 1 of 1: Extraction:  42%|███████████████▉                      | 60689/144519 [7:30:28<6:29:21,  3.59it/s]
C:\Convert\plugins\extract\align\_base.py:584: RuntimeWarning: overflow encountered in long_scalars
  size = ((roi[1][0] - roi[0][0]) ** 2 + (roi[1][1] - roi[0][1]) ** 2) ** 0.5
09/23/2022 13:36:51 INFO     Aligner filtered: [min_scale (0.07): 3142, max_scale (2.00): 265, distance (0.15): 7335)
09/23/2022 13:37:03 INFO     Writing alignments to: 'C:\Convert AI\LSHORT\left_alignments.fsa'

I have had a brief look online and there is discussion about it being a 32-bit limitation. I am no expert.

My main question is - should we be worried? Will there be dropped/missed frames in the extraction process caused by this error?

Thank you for any help in advance :D


Re: Error Logged While Extracting - 32bit limitation?

Posted: Fri Sep 23, 2022 2:17 pm
by torzdf

Ok, that's annoying. This is part of the new filters, of which I have run literally hundreds of thousands of frames through now and not hit this issue :/

If there is anyway you can provide me a source file + config which reproduces this I would be able to look for a fix.

In the meantime you can disable the filters by going into Settings > Extract and turning all the filters to 0

As this is a warning, not an error, I would not expect any data to have been lost.


Re: Error Logged While Extracting - 32bit limitation?

Posted: Fri Sep 23, 2022 2:49 pm
by ianstephens

No problem, I am here to test things and report back :D

You would require the video source file? I can provide it however it's ~ 75GB. The actual video resolution is 3840x3840 - perhaps that plays a part.

With regards to (the now completed) extraction process - am I safe to assume everything has been extracted and masked correctly?

We are using this extraction as the master convert extraction (every 1 frame) so wouldn't want to miss anything. Once sorted and cleaned up, we then run a script to grab every X number of faces using a loop for the training set.


Re: Error Logged While Extracting - 32bit limitation?

Posted: Fri Sep 23, 2022 5:20 pm
by torzdf

Ah, ok. The res may well be why. I will artificially upscale some videos for testing.

I doubt that this would have caused any issues with output, but I can't guarantee it. I will look at what returns from numpy when this warning is hit. However, as it is just a test on whether the face is bigger than a certain threshold, the worst case scenario is that it has deleted a face that it was going to delete anyway.


Re: Error Logged While Extracting - 32bit limitation?

Posted: Tue Oct 04, 2022 11:14 am
by torzdf

I have updated the calculation which is causing you issues to 64bit. Hopefully that will remedy this error