PaddleOCR/main/en/quick_start

[giscus[bot]]

PaddleOCR/main/en/quick_start

Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)

https://paddlepaddle.github.io/PaddleOCR/main/en/quick_start.html

[Darya1377]

Hello, i'm using the lates pandas==2.2.3 and python 3.12.10 and it raises an error: ERROR: Failed to build installable wheels for some pyproject.toml based projects (pandas) when i try to pip install paddleocr==3.0.0

[liuhongen1234567]

Hello, I successfully created a Python 3.12.9 environment using conda. You can try running pip install pandas==1.5.3, and then pip install paddleocr. If that doesn’t work, you might need to downgrade Python to 3.12.9 or below.

[timminator]

Pandas 1.5.3 which is required by PaddleX (dependency from PaddleOCR) is not compatible with Python 3.12. This does only work if you have a working C Compiler installed so it can be build from source.
This is a big issue and mentioned here and here.
PaddleOCR 3.0.0 is in this state not compatible anymore with Python 3.12.
The current solution is to set up a venv with Python 3.11 and install it there.

[DingdangGa0]

macOS 这样之后就卡住了，正常吗
d. Please be aware that recompiling all source files may be required. You can download and install ccache from: https://github.com/ccache/ccache/blob/master/doc/INSTALL.md
warnings.warn(warning_message)
MKL-DNN is not available. Using paddle instead.
Creating model: ('PP-OCRv5_server_det', None)
Using official model (PP-OCRv5_server_det), the model files will be automatically downloaded and saved in /Users/dingdang/.paddlex/official_models.
Fetching 6 files: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:00<00:00, 1499.30it/s]
Creating model: ('PP-OCRv5_server_rec', None)
Using official model (PP-OCRv5_server_rec), the model files will be automatically downloaded and saved in /Users/dingdang/.paddlex/official_models.
Fetching 6 files: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:00<00:00, 1585.15it/s]

[Cath1y]

我也

[Siddd25]

I have RTX 5050 on my laptop and CUDA 12.8 installed,
but the code is giving errors when tried to run with gpu
how should I use my GPU while running PaddleOCR

[liuhongen1234567]

Hello, the latest version of PaddlePaddle-GPU should already support the 50-series graphics cards. You can try upgrading your CUDA version to 12.9 or above, and then install PaddlePaddle-GPU==3.2.1. The reference documentation is as follows: https://www.paddleocr.ai/main/en/version3.x/pipeline_usage/PaddleOCR-VL-NVIDIA-Blackwell.html#12-method-2-manually-install-paddlepaddle-and-paddleocr

[Siddd25]

highest cuda version supported is 12.8, and it does not run on cuda 12.8..

[tomclancys]

It doesn't work on the 5070 Ti either. Errors pop up.
The PaddleOCR-VL-NVIDIA-Blackwell documentation does not work.

[anmerino-pnd]

Hello!

I've been trying to use this code from your tutorials and it's not working

from paddleocr import PaddleOCR
from PIL import Image

Initialize the OCR engine, specify the language (e.g., 'en' for English)

use_angle_cls=True helps with text at various angles

ocr = PaddleOCR(use_textline_orientation=False, lang='es')

Run OCR on an image path (can be a local file or a web URL)

img_path = '../datos/casa_ley/febrero_2026/04022026_pagina_01.png'
result = ocr.predict(img_path)

Print the results

for res in result:
res.print()
res.save_to_img("output")
res.save_to_json("output")

The Error says "NotImplementedError: (Unimplemented) ConvertPirAttribute2RuntimeAttribute not support [pir::ArrayAttributepir::DoubleAttribute] (at ..\paddle\fluid\framework\new_executor\instruction\onednn\onednn_instruction.cc:118)"

What am I doing wrong?

[Adammz123]

Receive 2 errors:

[transformers] Resampling is not supported in SLANeXt

sklearn.utils._param_validation.InvalidParameterError: The 'n_clusters' parameter of KMeans must be an int in the range [1, inf). Got 0 instead.

I am running the PP-StructureV3 [transformers] exactly as written above in a new venv after installing "transformers>=5.8.0" and "paddleocr[all]"

Ideas?