36978 | Fast image processor for DPT model (#37481)

* chore: ran codegen script

* test: test_image_processor_properties

* test: test_image_processor_from_dict_with_kwargs

* test: wip - test_padding

* test: test_padding

* test: test_keep_aspect_ratio

* wip

* test

* test: wip

* test: wip

* test: test_call_segmentation_maps, wip

* chore: tidy up

* test: test_call_segmentation_maps

* fix: test_save_load_fast_slow

* test: reduce labels

* chore: make fixup

* chore: rm comment

* chore: tidy

* chore remove comment

* refactor: no need to infer channel dimesnion

* refactor: encapsulate logic for preparing segmentation maps

* refactor: improve readability of segmentation_map preparation

* improvement: batched version of pad_image

* chore: fixup

* docs

* chore: make quality

* chore: remove unecessary comment

* fix: add SemanticSegmentationMixin

* feat: add post_process_depth_estimation to fast dpt image processor

* chore: fix formatting

* remove max_height, max_width

* fix: better way of processin segmentation maps
- copied from Beit Fast processor

* chore: formatting + remove TODO

* chore: fixup styles

* chore: remove unecessary line break

* chore: core review suggestion to remove autodocstring

* fix: add do_reduce_labels logic + refactor
- refactor preprocess logic to make it consistent with other processors
- add missing reduce labels logic

* refactor: remove deprecated mixin

* chore: fixup

* use modular for dpt + final nit changes

* fix style

---------

Co-authored-by: Samuel Rae <samuelrae@Samuels-Air.fritz.box>
Co-authored-by: yonigozlan <yoni.gozlan@huggingface.co>

docs/source/en/model_doc/dpt.md

@@ -78,7 +78,13 @@ If you're interested in submitting a resource to be included here, please feel f
 
 [[autodoc]] DPTImageProcessor
     - preprocess
+
+## DPTImageProcessorFast
+
+[[autodoc]] DPTImageProcessorFast
+    - preprocess
     - post_process_semantic_segmentation
+    - post_process_depth_estimation
 
 ## DPTModel