Add HF integration

[NielsRogge]

Hi @Verg-Avesta and team,

Thanks for this nice work!

I wrote a quick PoC to showcase that you can easily have integration with the 🤗 hub so that you can automatically load the various CounTR models using from_pretrained (and push them using push_to_hub), track download numbers for your models (similar to models in the Transformers library), have nice model cards on a per-model basis, and leverage safetensors for serialization. It leverages the PyTorchModelHubMixin class which allows to inherits these methods.

This way, people can also find your models easier, as currently they are hosted on Google Drive.

Usage is as follows:

from models_mae_cross import SupervisedMAE

# load model
model = SupervisedMAE()

# load weights
model.load_state_dict(...)

# push to the hub
model.push_to_hub("your-hf-org-or-username/countr-finetuned-fsc147")

# reload
model = SupervisedMAE.from_pretrained("your-hf-org-or-username/countr-finetuned-fsc147")

This means people don't need to manually download a checkpoint first in their local environment, it just loads automatically from the hub.

Would you be interested in this integration?

Kind regards,

Niels

Note

Please don't merge this PR before pushing the model to the hub :)

[NielsRogge]

I've also indexed the paper https://huggingface.co/papers/2208.13721, so that we can link the models to the paper.