R&D CLIP model training and adding to the library

[arxyzan]

This issue specifies the roadmap for adding CLIP to Hezar (model architecture, task, datasets, pretrained weights, etc)

[arxyzan]

The general steps to take:

• R&D model architecture, code and possibility of training the model
• R&D the preferred dataset
• R&D the required preprocessors and check if we already have them in Hezar
• R&D the implemented and trained models on Persian

Then, for each of these steps, open an issue.

[pooya-mohammadi]

@arxyzan jan, I'm on it

[pooya-mohammadi]

Useful links: https://github.com/sajjjadayobi/CLIPfa https://huggingface.co/datasets/amir7d0/laion20M-fa

[arxyzan]

This task has been postponed for later.