dataset pre-processing

[Jerric-wu]

It's clear that you did some pre-processing on the lip_mpv_dataset, stored an index in the pickle file and the results in the mentioned folders (dataloader.py)
self.path + "lip_images/"
self.path + "lip_dense/"
self.path + "lip_textures/"
self.path + "lip_apparels/"
these directories are missed in the code
Can you provide the tools for this pre-processing?

[tstkenny]

Same here.. quite some people are interested in your work. Would you provide the missing puzzle to help us out please?

[idinkov]

I managed to get it working with a bit of a magic, i will post info on how i got it later on my forked repo of this one.

[rubelchowdhury20]

@tstkenny please follow the link https://github.com/facebookresearch/DensePose/tree/main/notebooks to get the dense values and textures you need, which is not given in the repository.

[tstkenny]

@rubelchowdhury20 Hey man! It's really glad to see your reply. Would you be kind enough to make a simple workflow about how to make this repo work? The link you mentioned above contains just a few ipynb files, those do not fit in the codes here such as train.py:

with open(data_path + "lip_pairs.pkl", "rb") as fp: #Pickling

Appreciate your effort again!