[Semantic Segmentation] - Add SegFormer Architecture, Weight Conversion Script and Presets

[DavidLandup0]

This PR adds:

• SegFormerBackbone and presets
• Preprocessor flow
• SegFormerImageSegmenter and presets for Cityscapes and ADE20k (B0...B5 each)
• Conversion script
• Tests

Basic Usage

preprocessor = keras_hub.models.ImageSegmenterPreprocessor.from_preset("segformer_b0_ade20k_512")
segmenter = keras_hub.models.SegFormerImageSegmenter.from_preset("segformer_b0_ade20k_512")
segmenter(np.random.rand(1, 512, 512, 3))

End-to-end example with preprocessor:

import urllib.request 
from PIL import Image 
import numpy as np
import keras_hub

preprocessor = keras_hub.models.ImageSegmenterPreprocessor.from_preset("segformer_b0_ade20k_512")
segmenter = keras_hub.models.SegFormerImageSegmenter.from_preset("segformer_b0_ade20k_512")

img_url = "https://www.vanorohotel.com/wp-content/uploads/2021/07/drz-vanoro_6737.jpg"  
urllib.request.urlretrieve(img_url, "image.png") 

img = np.array(Image.open("image.png").resize((512, 512)))
img = np.expand_dims(img, 0)
inputs = preprocessor(img)
outs = segmenter(inputs)

With Image Converter

converter = keras_hub.layers.ImageConverter(image_size=(512, 512))
preprocessor = keras_hub.models.ImageSegmenterPreprocessor.from_preset("segformer_b0_ade20k_512", image_converter=converter)
segmenter = keras_hub.models.SegFormerImageSegmenter.from_preset("segformer_b0_ade20k_512")

Training Pipeline Example

A few examples in the notebook below:

• Instantiation of Backbone and Segmenter with MiT Encoder
• Running on input images
• Training pipeline with TFDS on Oxford IIIT Pets as an example

https://colab.research.google.com/drive/1EBNg6nPKx_KzyRuQQtHZ_PG_Nsf2pAg2#scrollTo=V9Ub4NHKCx9e

After a few minutes of training from scratch (both encoder and segmenter):

[divyashreepathihalli]

Is this PR ready for review?

[DavidLandup0]

@divyashreepathihalli In essence - yes, but there's some noise when running predictions using the converted weights (example image below of huggingface outputs vs our outputs). Looking into the numerics again, but the rest of the PR is ready for review :)

[DavidLandup0]

Found the issue - a transpose call shuffling the order of a latent in the encoder incorrectly. I'll get the presets up on Kaggle now