cleanup prototype transforms functional tests

[pmeier]

This PR cleans up the recent changes to our prototype transforms functional test frameworks. It merges the files test_prototype_transforms_kernels.py and test_prototype_transforms_dispatchers.py into test_prototype_transforms_functional.py while removing any obsolete code in the latter.

From here on there are the following tasks that are left to do:

1. Add KernelInfo's for the remaining kernels

   vision/test/test_prototype_transforms_kernels.py

   Lines 33 to 41 in 56e707b

   	"adjust_brightness_image_tensor",
   	"adjust_contrast_image_tensor",
   	"adjust_gamma_image_tensor",
   	"adjust_hue_image_tensor",
   	"adjust_saturation_image_tensor",
   	"clamp_bounding_box",
   	"five_crop_image_tensor",
   	"normalize_image_tensor",
   	"ten_crop_image_tensor",

   Afterwards, this can also be used to add the DispatcherInfo's.

2. Add dispatch tests for the dispatchers. Basically we need to make sure that

   1. if PIL.Image.Image gets passed, we dispatch to the PIL kernel
   2. if a simple tensor is passed, we dispatch to the image tensor kernel
   3. if a _Feature is passed we dispatch to the feature method

3. Reduce the number of sample inputs since they are used for multiple tests:

   Overview

   kernel	# sample inputs
   pad_image_tensor	2448
   pad_mask	864
   affine_image_tensor	480
   rotate_image_tensor	432
   center_crop_image_tensor	432
   resize_image_tensor	384
   crop_image_tensor	384
   center_crop_mask	324
   resize_mask	288
   convert_color_space_image_tensor	288
   crop_mask	288
   resized_crop_image_tensor	192
   convert_format_bounding_box	144
   resized_crop_mask	144
   pad_bounding_box	108
   center_crop_bounding_box	108
   perspective_image_tensor	80
   elastic_image_tensor	80
   affine_mask	72
   rotate_mask	72
   crop_bounding_box	72
   erase_image_tensor	48
   horizontal_flip_mask	36
   affine_bounding_box	36
   vertical_flip_mask	36
   rotate_bounding_box	36
   resized_crop_bounding_box	36
   perspective_bounding_box	36
   elastic_bounding_box	36
   gaussian_blur_image_tensor	32
   horizontal_flip_image_tensor	24
   resize_bounding_box	24
   vertical_flip_image_tensor	24
   invert_image_tensor	24
   solarize_image_tensor	24
   autocontrast_image_tensor	24
   adjust_sharpness_image_tensor	16
   perspective_mask	12
   elastic_mask	12
   posterize_image_tensor	12
   horizontal_flip_bounding_box	6
   vertical_flip_bounding_box	6
   equalize_image_tensor	4

4. Port the remaining correctness tests into references of the new framework.

[vfdev-5]

Thanks @pmeier !