Hi
I've been using the manual tool with pretty good success on video based sources. However, I have a bunch of additional images in png format that I'm also trying to create alignments for for later training.
The Extract and Sort process go just fine. When I load the output into the manual tool, I get some wonky behavior. As noted in another thread, the alignments are all shifted off the extracted images each time I advance through the images. And when I go to extract the final faces based on the redone alignments, I get the error below.
Reading the threads, it seems like the manual tool really wants the extraction process to use video, not source images. Is that true? DId I miss a way to fix alignments if the source is a folder of distinct photos? Or am I being a noob and missing something obvious?
Any help appreciated.
Code: Select all
Exception in Tkinter callback
Traceback (most recent call last):
File "C:\Users\rshah\MiniConda3\envs\faceswap\lib\tkinter\__init__.py", line 1883, in __call__
return self.func(*args)
File "C:\Users\rshah\MiniConda3\envs\faceswap\lib\tkinter\__init__.py", line 804, in callit
func(*args)
File "C:\Users\rshah\faceswap\tools\manual\detected_faces.py", line 370, in _monitor_extract
thread.check_and_raise_error()
File "C:\Users\rshah\faceswap\lib\multithreading.py", line 84, in check_and_raise_error
raise error[1].with_traceback(error[2])
File "C:\Users\rshah\faceswap\lib\multithreading.py", line 37, in run
self._target(*self._args, **self._kwargs)
File "C:\Users\rshah\faceswap\tools\manual\detected_faces.py", line 417, in _background_extract
self._alignments.data[basename]["faces"] = final_faces
KeyError: '16123556_0.png'