Support for Batch Inference

Hi and Thanks for the great package.

I wanted to ask if there is any support for inferring Landmarks of a batch of images? Each `fa.get_landmarks(image)` takes a really long time compared to something like DLib's landmark detector even on a GPU and I was hoping I could use batch processing to process a batch of images at once and remedy this issue a bit.

Thanks again.

Look at the inference code in face_alignment/api.py and the underlying model forward passes. The issue is about modifying the pipeline to accept and process a batch of images instead of one at a time. Check how images are currently loaded and passed to the model, and see if the model itself supports batched input. 'Done' means the get_landmarks function can accept a list of images and return a list of landmarks efficiently.

Written by the indexing model from the issue text.