Torch-rs

- Added RSVQAHR, Sydney Captions, UC Merced Captions, S2MTCP, ADVANCE, SAT-4, SAT-6, HRSCD, Inria AIL, TiSeLaC, GID-15, ZeuriCrop, AID, Dubai Segmentation, HKH Glacier Mapping, UC Merced, PatternNet, WHU-RS19, RSSCN7, Brazilian Coffee Scenes datasets & datamodules
- Added methods for creating train/val/test splits in datamodules
- Added ExtractChips transform
- Added dataset and datamodules local tests
- Added h5py, imagecodecs, torchaudio reqs
- Some minor fixes+additions to current datasets
- Added FCEF, FCSiamConc, FCSiamDiff PL modules

Added FAIR1M, OSCD, S2Looking, LEVIR-CD+, RSVQAxBEN datasets
Added EF, Siam, FC-EF, FC-Siam-conc, FC-Siam-diff change detection models

- Added PROBA-V, ETCI2021, RSVQA LR, RSICD, RESISC45, EuroSAT datasets
- Added RAMS model
- Added Compose, ToTensor, ToDtype transforms