Refactoring Faceset DL and usage

[camjac251]

Download now goes to the faceset.pak. It uses the new pack feature in DFL instead of using individual files.

[camjac251]

I had issues in the current version with unzipping so I switched it to the new one file pack system that iperov setup. I'm hosting the faceset.pak on my GDrive currently. If you like the changes, would it be possible to change the GDrive URL before it's pushed publicly?

Also I was thinking a more advanced faceset selection might be possible. Like importing a faceset and inputting the folder path as an option. Or possibly premade ones that are available publicly for pretraining.

[camjac251]

Nevermind this is a bust. It is not working as I'd hoped

Initializing models: 100% 5/5 [01:26<00:00, 16.39s/it]
Error: name 'samples_dat_path' is not defined
Traceback (most recent call last):
  File "/content/DeepFaceLab/samplelib/SampleLoader.py", line 50, in load
    result = samplelib.PackedFaceset.load(samples_path)
  File "/content/DeepFaceLab/samplelib/PackedFaceset.py", line 132, in load
    raise NotImplementedError
NotImplementedError

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/content/DeepFaceLab/mainscripts/Trainer.py", line 57, in trainerThread
    debug=debug,
  File "/content/DeepFaceLab/models/ModelBase.py", line 173, in __init__
    self.on_initialize()
  File "/content/DeepFaceLab/models/Model_SAEHD/Model.py", line 755, in on_initialize
    generators_count=src_generators_count ),
  File "/content/DeepFaceLab/samplelib/SampleGeneratorFace.py", line 42, in __init__
    samples = SampleLoader.load (SampleType.FACE, self.samples_path)
  File "/content/DeepFaceLab/samplelib/SampleLoader.py", line 52, in load
    io.log_err(f"Error occured while loading samplelib.PackedFaceset.load {str(samples_dat_path)}, {traceback.format_exc()}")
NameError: name 'samples_dat_path' is not defined
Traceback (most recent call last):
  File "DeepFaceLab/main.py", line 350, in <module>
    arguments.func(arguments)
  File "DeepFaceLab/main.py", line 185, in process_train
    Trainer.main(**kwargs)
  File "/content/DeepFaceLab/mainscripts/Trainer.py", line 217, in main
    e.wait() #Wait for inital load to occur.
  File "/usr/lib/python3.6/threading.py", line 551, in wait
    signaled = self._cond.wait(timeout)
  File "/usr/lib/python3.6/threading.py", line 295, in wait
    waiter.acquire()

[chervonij]

no longer needed. now loading packaged CelebA will be directly from GitHub.