import os from pydoc import describe os.system('pip install torch==1.9 torchvision') os.system('pip install detectron2==0.5 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cu102/torch1.9/index.html') os.system('pip install timm opencv-python-headless') import gradio as gr from demo.predictor import VisualizationDemo from detectron2.config import get_cfg from opendet2 import add_opendet_config model_cfgs = { "FR-CNN": ["configs/faster_rcnn_R_50_FPN_3x_baseline.yaml", "frcnn_r50.pth"], "OpenDet-R50": ["configs/faster_rcnn_R_50_FPN_3x_opendet.yaml", "opendet2_r50.pth"], "OpenDet-SwinT": ["configs/faster_rcnn_Swin_T_FPN_3x_opendet.yaml", "opendet2_swint.pth"], } def setup_cfg(model): cfg = get_cfg() add_opendet_config(cfg) model_cfg = model_cfgs[model] cfg.merge_from_file(model_cfg[0]) cfg.MODEL.WEIGHTS = model_cfg[1] cfg.MODEL.DEVICE = "cpu" cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.5 cfg.MODEL.ROI_HEADS.VIS_IOU_THRESH = 0.8 cfg.freeze() return cfg def inference(input, model): cfg = setup_cfg(model) demo = VisualizationDemo(cfg) # use PIL, to be consistent with evaluation predictions, visualized_output = demo.run_on_image(input) output = visualized_output.get_image()[:, :, ::-1] return output examples = [ ["demo/000000002157.jpg", "OpenDet-R50"], ["demo/000000020059.jpg", "OpenDet-R50"]] iface = gr.Interface( inference, [ "image", gr.inputs.Radio( ["FR-CNN", "OpenDet-R50", "OpenDet-SwinT"], default='OpenDet-R50'), ], "image", examples=examples, title="OpenDet", article="
", description="Online demo for: Expanding Low-Density Latent Regions for Open-Set Object Detection. Please upload your image, or click one of the examples to load them", ) iface.launch(enable_queue=True, cache_examples=True)