GFPGAN tries to run even when I don't specify --gfpgan

[Cubox]

When running dream.py, on each image generated, I get the following message:
GFPGAN not initialized, it must be loaded via the --gfpgan argument█████████████████████| 50/50 [00:07<00:00, 6.74it/s]

And when running dream_web.py, I get

Request to generate with prompt: test
100%|███████████████████████████████████████████████████████████████████████████████████| 50/50 [00:09<00:00,  5.04it/s]
Error running GFPGAN - Your image was not enhanced.█████████████████████████████████████| 50/50 [00:09<00:00,  5.27it/s]
'>' not supported between instances of 'NoneType' and 'int'
Sampling: 100%|███████████████████████████████████████████████████████████████████████████| 1/1 [00:11<00:00, 11.62s/it]

I assume this isn't the intended behaviour, unless I explicitly call for GFPGAN, I don't want to try to run it/initialize it

Thanks for the work on this

[Oceanswave]

Thanks for the bug report! Yes, the default strength is unintentionally not None

[lstein]

This was fixed by Oceanswave's PR just above. Enter the stable-diffusion directory, do a "git pull", and possibly run "conda env update -f environment.yaml" if it doesn't work at first.

Feel free to enter a new Issue if it recurs.