[Feature] downstream tasks

[alaa-shubbak]

What is the problem this feature will solve?

i would like to ask if there any tutorial that helps in design a script to train a selfsupervised approach for another downstream tasks ,rather than only object detcetion and segmenation.

for example design a model for training for pose estimation downstream task?

I try to read and understand the models and the downstream tasks , but i could not find something to help in design a self supervised for pose estimation.

any help please.

What is the feature?

a general description on how to design a self-supervised for more downstream tasks

What alternatives have you considered?

No response