Some issues and suggestions I discovered during the replication process

[MrHandsomeljn]

Some issues and suggestions I discovered during the replication process:

1. in README.md step2, delete extra space at the end of -- imd_size=256 \ , which may cause the command to end prematurely.
2. in roma/models/model_zoo/roma_models.py, add from roma.models.encoders import CNNandDinov2 at the beginning of the file.
3. in check_points/mvinpainter_f_256/config.yaml and check_points/mvinpainter_o_256/config.yaml, add item beta_schedule: linear, change the value of prediction_type to epsilon.
4. in mask_adaption.py, the two files lists clean_fs and planet_mask_fs may be out of order, it's best to sort them (by clean_fs.sort() and planet_mask_fs.sort()), otherwise ROMA may have difficulty outputting normal results between two images with a large gap.