app.py

import gradio as gr

from matplotlib import gridspec
import matplotlib.pyplot as plt
import numpy as np
from PIL import Image
import tensorflow as tf
from transformers import SegformerFeatureExtractor, TFSegformerForSemanticSegmentation

feature_extractor = SegformerFeatureExtractor.from_pretrained(
    "nvidia/segformer-b0-finetuned-ade-512-512"
)
model = TFSegformerForSemanticSegmentation.from_pretrained(
    "nvidia/segformer-b0-finetuned-ade-512-512"
)

def ade_palette():
    """ADE20K palette that maps each class to RGB values."""
    return [
        [22, 122, 213],
        [240, 3, 156],
        [87, 176, 33],
        [154, 88, 111],
        [63, 54, 244],
        [201, 235, 59],
        [102, 66, 183],
        [94, 147, 5],
        [39, 198, 247],
        [17, 149, 92],
        [130, 78, 184],
        [246, 119, 107],
        [225, 23, 68],
        [52, 189, 140],
        [142, 10, 22],
        [114, 161, 251],
        [168, 55, 34],
        [75, 203, 89],
        [32, 45, 235],
        [74, 1, 129],
        [31, 166, 96],
        [223, 51, 202],
        [57, 72, 27],
        [143, 191, 176],
        [111, 33, 244],
        [20, 155, 62],
        [128, 99, 209],
        [254, 120, 14],
        [229, 67, 175],
        [53, 206, 40],
        [198, 77, 10],
        [8, 166, 142],
        [133, 45, 111],
        [222, 199, 239],
        [56, 18, 90],
        [164, 98, 206],
        [239, 135, 60],
        [106, 28, 139],
        [49, 172, 224],
        [179, 109, 34],
        [12, 191, 157],
        [121, 64, 88],
        [243, 214, 127],
        [82, 11, 165],
        [158, 37, 192],
        [31, 144, 55],
        [176, 220, 252],
        [68, 5, 123],
        [220, 157, 73],
        [41, 183, 210],
        [173, 85, 14],
        [16, 131, 99],
        [135, 50, 177],
        [227, 202, 244],
        [47, 175, 217],
        [181, 112, 28],
        [15, 190, 160],
        [124, 66, 91],
        [241, 217, 130],
        [80, 13, 168],
        [157, 40, 195],
        [30, 147, 52],
        [175, 223, 249],
        [67, 7, 126],
        [218, 160, 76],
        [235, 141, 45],
        [101, 33, 149],
        [46, 178, 220],
        [182, 114, 31],
        [14, 193, 163],
        [122, 69, 94],
        [240, 219, 133],
        [79, 16, 171],
        [156, 43, 198],
        [29, 150, 58],
        [225, 207, 243],
        [51, 27, 121],
        [159, 107, 229],
        [234, 143, 48],
        [100, 35, 152],
        [239, 221, 136],
        [78, 19, 174],
        [155, 46, 201],
        [28, 152, 61],
        [173, 227, 243],
        [20, 127, 108],
        [138, 59, 179],
        [224, 209, 245],
        [50, 29, 124],
        [161, 109, 232],
        [233, 145, 51],
        [99, 37, 155],
        [44, 174, 226],
        [184, 118, 20],
        [12, 195, 169],
        [125, 73, 100],
        [238, 223, 139],
        [77, 22, 177],
        [154, 49, 204],
        [27, 154, 64],
        [51, 86, 205]
    ]

labels_list = []

with open(r'labels.txt', 'r') as fp:
    for line in fp:
        labels_list.append(line[:-1])

colormap = np.asarray(ade_palette())

def label_to_color_image(label):
    if label.ndim != 2:
        raise ValueError("Expect 2-D input label")

    if np.max(label) >= len(colormap):
        raise ValueError("label value too large.")
    return colormap[label]

def draw_plot(pred_img, seg):
    fig = plt.figure(figsize=(20, 15))

    grid_spec = gridspec.GridSpec(1, 2, width_ratios=[6, 1])

    plt.subplot(grid_spec[0])
    plt.imshow(pred_img)
    plt.axis('off')
    LABEL_NAMES = np.asarray(labels_list)
    FULL_LABEL_MAP = np.arange(len(LABEL_NAMES)).reshape(len(LABEL_NAMES), 1)
    FULL_COLOR_MAP = label_to_color_image(FULL_LABEL_MAP)

    unique_labels = np.unique(seg.numpy().astype("uint8"))
    ax = plt.subplot(grid_spec[1])
    plt.imshow(FULL_COLOR_MAP[unique_labels].astype(np.uint8), interpolation="nearest")
    ax.yaxis.tick_right()
    plt.yticks(range(len(unique_labels)), LABEL_NAMES[unique_labels])
    plt.xticks([], [])
    ax.tick_params(width=0.0, labelsize=25)
    return fig

def sepia(input_img):
    input_img = Image.fromarray(input_img)

    inputs = feature_extractor(images=input_img, return_tensors="tf")
    outputs = model(**inputs)
    logits = outputs.logits

    logits = tf.transpose(logits, [0, 2, 3, 1])
    logits = tf.image.resize(
        logits, input_img.size[::-1]
    )  # We reverse the shape of `image` because `image.size` returns width and height.
    seg = tf.math.argmax(logits, axis=-1)[0]

    color_seg = np.zeros(
        (seg.shape[0], seg.shape[1], 3), dtype=np.uint8
    )  # height, width, 3
    for label, color in enumerate(colormap):
        color_seg[seg.numpy() == label, :] = color

    # Show image + mask
    pred_img = np.array(input_img) * 0.5 + color_seg * 0.5
    pred_img = pred_img.astype(np.uint8)

    fig = draw_plot(pred_img, seg)
    return fig

demo = gr.Interface(fn=sepia,
                    inputs=gr.Image(shape=(800, 600)),
                    outputs=['plot'],
                    examples=["indoor.jpg", "indoor1.jpg", "indoor2.jpg", "indoor3.jpg"],
                    allow_flagging='never')


demo.launch()