search

ying09 / TextFuseNet

A PyTorch implementation of "TextFuseNet: Scene Text Detection with Richer Fused Features".
MIT License
474 stars 121 forks source link

'keyError: non-existent config key: TEXTFUSENET_MUTIL_PATH_FUSE_ON' #105

Open AI-Passionner opened 1 year ago

AI-Passionner commented 1 year ago

@ying09 I am using TextFuseNet to detect the scent text in an image. When I had detectron2 installed (MacPro + M1 chip) and tried to run the demo, it shows the error as 'keyError: non-existent config key: TEXTFUSENET_MUTIL_PATH_FUSE_ON'

I searched this issue around and it seems I am the first one in the TextFuseNet. Any suggestions?

A similar issue was found but I don't have an idea.

Thanks.

diwakar-vsingh commented 1 year ago

In all the scripts inside the demo folder, just insert a line (cfg.set_new_allowed(True)) into the function setup_cfg. It will look something like:

def setup_cfg(args):
    # load config from file and command-line arguments
    cfg = get_cfg()
    cfg.set_new_allowed(True)
    cfg.merge_from_file(args.config_file)
    cfg.merge_from_list(args.opts)
    # Set model
    cfg.MODEL.WEIGHTS = args.weights
    # Set score_threshold for builtin models
    cfg.MODEL.RETINANET.SCORE_THRESH_TEST = args.confidence_threshold
    cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = args.confidence_threshold
    cfg.MODEL.PANOPTIC_FPN.COMBINE.INSTANCES_CONFIDENCE_THRESH = args.confidence_threshold
    cfg.freeze()
    return cfg