[WIP] Add Grounding DINO

[EduardoPach]

What does this PR do?

This PR adds Grounding DINO

Fixes #25423

To-Do's:

• Port vision backbone
• Port Text backbone
• Port Feature Enhancer
• Port Cross-Modality Decoder

[sgugger]

cc @amyeroberts for information. Please let us know when your PR is ready for review!

[EduardoPach]

@amyeroberts Hey, I've actually started a new branch to port the model because in the branch from this PR, I've done add-new-model-like and then swin, but talking to @NielsRogge he mentioned that starting from deformable-detr would be easier (which was a SUPER helpful tip btw). Should I close this PR and open a new one with the correct branch? (still WIP, but almost there)

[amyeroberts]

@EduardoPach Do whatever is easiest for you in terms of code management. If you open a new PR, just make sure to link to it in a comment on this one so it can be easily tracked.

[EduardoPach]

@amyeroberts opened a new PR in #26087 as mentioned. Should I close this one?

[amyeroberts]

@EduardoPach I can close it for you :)