search

jingsongliujing / OnnxOCR

基于PaddleOCR重构，并且脱离PaddlePaddle深度学习训练框架的轻量级OCR，推理速度超快
Apache License 2.0
715 stars 65 forks source link

Could you make it pip installable? #15

Open Jasonzhangzzf opened 3 months ago

Jasonzhangzzf commented 3 months ago

It would be much faster to deploy this model as a standalone module. Could you add the setup.py file?

jingsongliujing commented 3 months ago

I've also thought about setup.py, but due to the various styles of computers, there's no suitable storage location for the model, which may be supported in the future.

---Original--- From: "Zhifeng @.> Date: Sat, Jul 20, 2024 03:24 AM To: @.>; Cc: @.***>; Subject: [jingsongliujing/OnnxOCR] Could you make it pip installable? (Issue#15)

It would be much faster to deploy this model as a standalone module. Could you add the setup.py file?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>

Jasonzhangzzf commented 3 months ago

Can't we include the basic model in the module? I am happy to contribute to this feature since this OCR model is highly usable and I am tired of severe and unsolvable memory leakage problems in the original implementation.

jingsongliujing commented 3 months ago

Sure, this might be considered in the future. If you are willing to contribute to this feature, you are welcome to submit your code via a pull request (PR).