paddleocr-convert 0.0.4

Tool for converting the PaddleOCR model to onnx format.

paddleocr_convert

1. Install package by pypi.

pip install paddleocr_convert

2. Run by command line.

• Usage:

  $ paddleocr_convert -h
  usage: paddleocr_convert [-h] [-url MODEL_URL] [-o SAVE_DIR]
                          [-txt_url TXT_URL]
  
  optional arguments:
  -h, --help            show this help message and exit
  -url MODEL_URL, --model_url MODEL_URL
                          The inference model url of paddleocr. e.g.
                          https://paddleocr.bj.bcebos.com/PP-
                          OCRv3/chinese/ch_PP-OCRv3_det_infer.tar
  -o SAVE_DIR, --save_dir SAVE_DIR
                          The directory of saving the model.
  -txt_url TXT_URL, --txt_url TXT_URL
                          The raw txt url, if the model is recognition model.
  

• Example:

  $ paddleocr_convert -url https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar -o models
  
  $ paddleocr_convert -url https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar \
                      -o models \
                      -txt_url https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.6/ppocr/utils/ppocr_keys_v1.txt
  

3. Run by script.

from paddleocr_convert import PaddleOCRModelConvert

converter = PaddleOCRModelConvert()

url = 'https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_rec_infer.tar'

save_dir = 'models'
txt_url = 'https://raw.githubusercontent.com/PaddlePaddle/PaddleOCR/release/2.6/ppocr/utils/ppocr_keys_v1.txt'

converter(url, save_dir, txt_url=txt_url)

See more details for README.