Views
No views yet
1>>> from doctr.io import DocumentFile
2>>> from doctr.models import ocr_predictor, from_hub
3
4>>> img = DocumentFile.from_images(['<image_path>'])
5>>> # Load your model from the hub
6>>> model = from_hub('mindee/my-model')
7
8>>> # Pass it to the predictor
9>>> # If your model is a recognition model:
10>>> predictor = ocr_predictor(det_arch='db_mobilenet_v3_large',
11>>> reco_arch=model,
12>>> pretrained=True)
13
14>>> # If your model is a detection model:
15>>> predictor = ocr_predictor(det_arch=model,
16>>> reco_arch='crnn_mobilenet_v3_small',
17>>> pretrained=True)
18
19>>> # Get your predictions
20>>> res = predictor(img)