Spaces:
Runtime error
Runtime error
| try: | |
| import detectron2 | |
| except ImportError: | |
| import os | |
| os.system('pip install git+https://github.com/facebookresearch/detectron2.git') | |
| import streamlit as st | |
| import cv2 | |
| from PIL import Image | |
| import numpy as np | |
| import torch | |
| from detectron2.config import get_cfg | |
| from detectron2.engine import DefaultPredictor | |
| from detectron2.utils.visualizer import Visualizer | |
| from detectron2.data import MetadataCatalog | |
| cfg = get_cfg() | |
| cfg.merge_from_file('./faster_rcnn_R_50_FPN_3x.yaml') | |
| cfg.MODEL.ROI_HEADS.NUM_CLASSES = 13 | |
| cfg.MODEL.WEIGHTS = './fewshot_all_30.pth' | |
| my_metadata = MetadataCatalog.get("fewshot_all_30") | |
| my_metadata.thing_classes = ["clean water network", "communication network", "flood", "garbage", "gutter cover", | |
| "illegal parking", "layout and building", "park", "road", "sidewalk", "tree", "vandalism", "waterway"] | |
| if not torch.cuda.is_available(): | |
| cfg.MODEL.DEVICE = "cpu" | |
| def inference(image): | |
| im = np.array(image) | |
| #min_thresh = st.slider("Minimum score", 0.0, 1.0, 0.5) | |
| cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = min_thresh | |
| predictor = DefaultPredictor(cfg) | |
| outputs = predictor(im) | |
| v = Visualizer(im, | |
| metadata=my_metadata, | |
| scale=0.5 | |
| ) | |
| instances = outputs["instances"].to("cpu") | |
| # Customize the font_size | |
| v._default_font_size = 30 | |
| out = v.draw_instance_predictions(instances) | |
| return out.get_image() | |
| st.set_page_config(layout="wide", initial_sidebar_state="collapsed") # Set app layout to wide and hide sidebar by default on smaller screens | |
| st.title('Urban Problem Detection Model Demo') | |
| st.markdown('This demo introduces an interactive playground for our trained Detectron2 model for urban problem/report detection.') | |
| st.markdown('The model was trained on manually annotated image from image report in JAKI (Jakarta Kini).') | |
| st.markdown('The object categories it can detect are:') | |
| cl1, cl2, cl3 = st.columns(3) | |
| cl1.markdown( | |
| """ | |
| - Clean Water Network | |
| - Communication Network | |
| - Flood | |
| - Garbage | |
| - Gutter Cover | |
| """ | |
| ) | |
| cl2.markdown( | |
| """ | |
| - Illegal Parking | |
| - Layout and building | |
| - Park | |
| - Road | |
| - Sidewalk | |
| """ | |
| ) | |
| cl3.markdown( | |
| """ | |
| - Tree | |
| - Vandalism | |
| - Waterway | |
| """ | |
| ) | |
| uploaded_file = st.file_uploader('Upload Image', type=['jpg', 'jpeg', 'png']) | |
| min_thresh = st.slider("Minimum score", 0.0, 1.0, 0.5) | |
| col1, col2 = st.columns((1, 1)) | |
| with col1: | |
| #uploaded_file = st.file_uploader('Upload Image', type=['jpg', 'jpeg', 'png']) | |
| if uploaded_file is not None: | |
| image = Image.open(uploaded_file) | |
| st.image(image, caption='Uploaded Image', use_column_width=True) | |
| with col2: | |
| if uploaded_file is not None: | |
| result = inference(image) | |
| st.image(result, caption='Output', use_column_width=True) |