Continuous download of ch_PP-OCRv3_det_infer.tar

[pepperav]

ehi 👋,
I'm going to create a service by an AWS Lambda as Docker Image.

Even though I've putted rec_model_dir, det_model_dir and cls_model_dir to a local path with models already there,
every run it still download ch_PP-OCRv3_det_infer.tar

Code:
`

          import sys
          from paddleocr import PaddleOCR,draw_ocr
          
          def handler(event, context):
              ocr = PaddleOCR(ocr = PaddleOCR(det=False, det_model_dir='/tmp/whl/det/en/en_PP-OCRv3_det_infer/', rec_model_dir='/tmp/whl/rec/en/en_PP-OCRv3_det_infer/', cls_model_dir='/tmp/whl/cls/en/en_PP-OCRv3_det_infer/', use_angle_cls=False, use_gpu=False, lang='en'))
          
              return 'Hello from AWS Lambda using Python' + sys.version + '!'

`

Log:
download https://paddleocr.bj.bcebos.com/PP-OCRv3/chinese/ch_PP-OCRv3_det_infer.tar to /home/sbx_user1051/.paddleocr/whl/det/ch/ch_PP-OCRv3_det_infer/ch_PP-OCRv3_det_infer.tar

How can I avoid that?
Thank you

[Sayaka91]

@pepperav
I got the similar problem. It seems cause of paddleocr version. You can try down version to 2.6.1.2

[pepperav]

Thank you @Sayaka91 , I will try.

Actually with this folders tree:

and this code line:

it works without any other download.

[BisarAhmed]

any solution for this error it cause the failing for the docker compose up command start downloading before the download end it give this error
FileNotFoundError: [Errno 2] No such file or directory: '/root/.paddleocr/whl/det/ml/Multilingual_PP-OCRv3_det_infer/Multilingual_PP-OCRv3_det_infer.tar'

[pepperav]

I see multilingual on the log printed by you. In my case I forced to be only english lang='en'.

I guess this is the reason why it try to download other stuff.

[BisarAhmed]

i passed use_onnx as true it stopped downloading but failed at reading .tar files. i am still debugging

use_onnx=True, det_model_dir="whl/det/en/en_PP-OCRv3_rec_infer.tar", rec_model_dir="whl/rec/en/en_PP-OCRv3_rec_infer.tar", cls_model_dir="whl/cls/en/ch_ppocr_mobile_v2.0_cls_infer.tar"

    ` INVALID_PROTOBUF : Load model from whl/cls/en/ch_ppocr_mobile_v2.0_cls_infer failed:Protobuf parsing failed.`

[pepperav]

I don't know what use_onnx does but you should specify folders with tar files already extracted!

eg:
wrong: det_model_dir="whl/det/en/en_PP-OCRv3_rec_infer.tar"
valid: det_model_dir="whl/det/en/en_PP-OCRv3_rec_infer/"

[khansajeel]

@BisarAhmed you should not put in tar files there, use extracted directories. it should work

[BisarAhmed]

@khansajeel @pepperav thanks it solved the download issue

[madhavi1102]

@Sayaka91 without following the mentioned structure by you for downloading det, rec, pls models, it is working fine with specifying any local path used for referring these extracted models.

[oscarjose9423]

@pepperav Hey I am doing this right now, would you mind sharing your Dockerfile? the solution and folder structure did not work for me. Additionally I am getting an error:

able to unmarshal input: 'utf-8' codec can't decode byte 0x9c in position 147: invalid start byte

when testing the container locally

[pepperav]

@pepperav Hey I am doing this right now, would you mind sharing your Dockerfile? the solution and folder structure did not work for me. Additionally I am getting an error:

able to unmarshal input: 'utf-8' codec can't decode byte 0x9c in position 147: invalid start byte

when testing the container locally

Sure!

FROM public.ecr.aws/lambda/python:latest-x86_64

COPY requirements.txt ./

RUN pip3 install -r requirements.txt --target "${LAMBDA_TASK_ROOT}"
RUN yum check-update; if [ $? -eq 100 ]; then exit 0; else exit $?; fi
RUN yum install -y libgomp mesa-libGL

COPY app.py ${LAMBDA_TASK_ROOT}
COPY whl ${LAMBDA_TASK_ROOT}/whl

CMD [ "app.handler" ]

This is the requirements.txt file:

paddleocr
paddlepaddle==2.4.2 -i https://pypi.tuna.tsinghua.edu.cn/simple

Note: LAMBDA_TASK_ROOT – The path to your Lambda function code.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.

[UserWangZz]

This issue has not been updated for a long time. This issue is temporarily closed and can be reopened if necessary.