infer.py

from transformers import LayoutLMv3Processor
from datasets import load_dataset
from PIL import Image, ImageDraw, ImageFont
import numpy as np
from utils import get_data
import json


## INFERENCE

from transformers import AutoModelForTokenClassification


processor = LayoutLMv3Processor.from_pretrained("microsoft/layoutlmv3-base")
model = AutoModelForTokenClassification.from_pretrained("model")

with open('model/config.json') as f:
    id2label = json.load(f)["id2label"]
#id2label = {v: k for v, k in enumerate(labels)}
label2color = {'question':'blue', 'answer':'green', 'header':'orange', 'other':'violet'}

def unnormalize_box(bbox, width, height):
     return [
         width * (bbox[0] / 1000),
         height * (bbox[1] / 1000),
         width * (bbox[2] / 1000),
         height * (bbox[3] / 1000),
     ]

def iob_to_label(label):
    label = label[2:]
    if not label:
      return 'other'
    return label

def process_image(image):
    width, height = image.size

    # encode
    encoding = processor(image, truncation=True, return_offsets_mapping=True, return_tensors="pt")
    offset_mapping = encoding.pop('offset_mapping')

    # forward pass
    outputs = model(**encoding)

    # get predictions
    predictions = outputs.logits.argmax(-1).squeeze().tolist()
    token_boxes = encoding.bbox.squeeze().tolist()

    # only keep non-subword predictions
    is_subword = np.array(offset_mapping.squeeze().tolist())[:,0] != 0
    true_predictions = [id2label[str(pred)] for idx, pred in enumerate(predictions) if not is_subword[idx]]
    true_boxes = [unnormalize_box(box, width, height) for idx, box in enumerate(token_boxes) if not is_subword[idx]]

    # draw predictions over the image
    draw = ImageDraw.Draw(image)
    font = ImageFont.load_default()
    for prediction, box in zip(true_predictions, true_boxes):
        predicted_label = iob_to_label(prediction).lower()
        draw.rectangle(box, outline=label2color[predicted_label])
        draw.text((box[0]+10, box[1]-10), text=predicted_label, fill=label2color[predicted_label], font=font)
    
    return image

image = Image.open("sample.png").convert(mode="RGB")

image = process_image(image)

image.save("predicted.png")