speed up GFPGAN inference

Job ID: 35381306

Budget: $250 – $750 USD

Using GFPGAN for face restoration, we need to speed up the inference.
https://github.com/TencentARC/GFPGAN

One possible solution would be:
https://github.com/bes-dev/MobileStyleGAN.pytorch

OR

Using half() and adjust the model.

***Any solution*** is welcome, as long as the inference time is reduced by 50%.
Related categories: Machine Learning (ML) Deep Learning