RTX 5090 cannot be used for GPU acceleration

[zhenglipeng]

问题描述 / Problem Description

Failed to enable GPU acceleration

运行环境 / Runtime Environment

python
cuda 12.9
RTX 5090
Package Version

---

aistudio-sdk 0.3.8
annotated-doc 0.0.4
annotated-types 0.7.0
antlr4-python3-runtime 4.9.3
anyio 4.12.1
astor 0.8.1
bce-python-sdk 0.9.71
blinker 1.9.0
certifi 2026.1.4
chardet 7.4.3
charset-normalizer 3.4.7
click 8.3.3
colorama 0.4.6
colorlog 6.10.1
contourpy 1.3.3
crc32c 2.8
cycler 0.12.1
decorator 5.2.1
distro 1.9.0
easyocr 1.7.2
et_xmlfile 2.0.0
filelock 3.29.0
Flask 3.1.2
flatbuffers 25.12.19
fonttools 4.63.0
fsspec 2026.4.0
future 1.0.0
h11 0.16.0
hf-xet 1.5.0
httpcore 1.0.9
httpx 0.28.1
huggingface_hub 1.14.0
idna 3.11
ImageIO 2.37.3
imagesize 2.0.0
itsdangerous 2.2.0
Jinja2 3.1.6
jiter 0.14.0
kiwisolver 1.5.0
lazy-loader 0.5
Levenshtein 0.27.3
markdown-it-py 4.2.0
MarkupSafe 3.0.3
matplotlib 3.10.9
mdurl 0.1.2
modelscope 1.36.3
mpmath 1.3.0
networkx 3.6.1
ninja 1.13.0
numpy 2.3.5
nvidia-cublas-cu12 12.9.0.13
nvidia-cuda-runtime-cu12 12.9.37
nvidia-cudnn-cu12 9.9.0.52
nvidia-cufft-cu12 11.4.0.6
nvidia-curand-cu12 10.3.10.19
nvidia-cusolver-cu12 11.7.4.40
nvidia-cusparse-cu12 12.5.9.5
nvidia-nvjitlink-cu12 12.9.86
omegaconf 2.3.0
onnxruntime 1.26.0
openai 2.30.0
opencv-contrib-python 4.10.0.84
opencv-python 4.13.0.92
opencv-python-headless 4.13.0.92
openpyxl 3.1.5
opt-einsum 3.3.0
packaging 26.0
paddleocr 3.5.0
paddlepaddle-gpu 3.3.1
paddlex 3.5.2
pandas 3.0.3
pandas-stubs 3.0.0.260204
pillow 12.1.0
pip 26.1.1
prettytable 3.17.0
protobuf 6.33.4
psutil 7.2.2
py-cpuinfo 9.0.0
pyclipper 1.4.0
pycryptodome 3.23.0
pydantic 2.13.4
pydantic_core 2.46.4
Pygments 2.20.0
pyparsing 3.3.2
pypdfium2 5.8.0
python-bidi 0.6.9
python-dateutil 2.9.0.post0
PyYAML 6.0.2
RapidFuzz 3.14.5
rapidocr 3.8.1
rapidocr-onnxruntime 1.4.4
rapidocr-paddle 1.4.5
requests 2.34.0
rich 15.0.0
ruamel.yaml 0.19.1
safetensors 0.7.0
scikit-image 0.26.0
scipy 1.17.1
setuptools 81.0.0
shapely 2.1.2
shellingham 1.5.4
six 1.17.0
sniffio 1.3.1
sympy 1.14.0
tifffile 2026.3.3
torch 2.8.0+cu129
torchaudio 2.8.0+cu129
torchvision 0.23.0+cu129
tqdm 4.67.3
typer 0.25.1
typing_extensions 4.15.0
typing-inspection 0.4.2
tzdata 2026.2
ujson 5.12.1
urllib3 2.7.0
wcwidth 0.7.0
Werkzeug 3.1.8
wheel 0.46.3

复现代码 / Reproduction Code

from rapidocr import  RapidOCR
import time

engine = RapidOCR(
    params={
        "Det.engine_type": "paddle",
        "EngineConfig.paddle.use_cuda": True,  # 使用PaddlePaddle GPU版推理
        "EngineConfig.paddle.cuda_ep_cfg.device_id": 0,  # 指定GPU id
    }
)

if __name__ == '__main__':
    result3 = engine(r'D:\tools\pythonproject\photodiff\photodiff\photo\1.png')
    gpu_time = time.time()
    result4 = engine(r'D:\tools\pythonproject\photodiff\photodiff\photo\1.png')
    gpu_cost = time.time() - gpu_time
    print("gputime", gpu_cost)
    print("gpu", result4)

可用加速器: ['AzureExecutionProvider', 'CPUExecutionProvider']
gputime 2.1445775032043457

可能解决方案 / Possible solutions

[SWHL]

Follow the steps below. It is recommended to completely uninstall rapidocr-onnxruntime and rapidocr-paddle, as they may cause conflicts.

1. Uninstall rapidocr-onnxruntime and rapidocr-paddle.
2. Check if paddlepaddle-gpu is installed successfully.

    >>> import paddle
    
    >>> 
    >>> paddle.utils.run_check()
    Running verify PaddlePaddle program ...
    I0818 15:35:08.335391 30540 program_interpreter.cc:173] New Executor is Running.
    I0818 15:35:08.398319 30540 interpreter_util.cc:529] Standalone Executor is Used.
    PaddlePaddle works well on 1 CPU.
    PaddlePaddle is installed successfully! Let's start deep learning with PaddlePaddle now.

3. Run the code below for testing:

   from rapidocr import EngineType, RapidOCR
   
   engine = RapidOCR(
       params={
       "Det.engine_type": EngineType.PADDLE,
       "EngineConfig.paddle.use_cuda": True,  # 使用PaddlePaddle GPU版推理
       "EngineConfig.paddle.cuda_ep_cfg.device_id": 0,  # 指定GPU id
       }
   )
   
   img_url = "https://www.modelscope.cn/models/RapidAI/RapidOCR/resolve/master/resources/test_files/ch_en_num.jpg"
   result = engine(img_url)
   print(result)
   
   result.vis('vis_result.jpg')

[zhenglipeng]

error Cannot find reference 'EngineType' in 'rapidocr.py' Cannot find reference 'RapidOCR' in 'rapidocr.py'

[zhenglipeng]

@SWHL

[yudin-s]

这里有两个问题要分开看。

第一，Cannot find reference 'EngineType' in 'rapidocr.py' 很像是本地文件名遮蔽了真正的 rapidocr 包。请先确认你的脚本文件不要叫 rapidocr.py，项目目录里也不要有同名的 rapidocr.py / rapidocr 文件夹。可以在当前环境里运行：

import rapidocr
print(rapidocr.__file__)
print(hasattr(rapidocr, "RapidOCR"), hasattr(rapidocr, "EngineType"))

如果输出路径指向你自己的项目文件，比如 D:\tools\pythonproject\...\rapidocr.py，那就先把这个文件改名，然后删除 __pycache__ 再重试。

第二，你现在日志里的可用 provider 是：

['AzureExecutionProvider', 'CPUExecutionProvider']

这说明当前实际走的是 ONNX Runtime/CPU provider，不是 Paddle GPU。按维护者的建议，先避免同时安装 rapidocr-onnxruntime 和 rapidocr-paddle，否则很容易选错后端。建议新建一个干净环境测试：

pip uninstall -y rapidocr-onnxruntime rapidocr-paddle rapidocr
pip install rapidocr rapidocr-paddle paddlepaddle-gpu

然后先验证 Paddle 本身能看到 GPU：

import paddle
print(paddle.device.cuda.device_count())
paddle.utils.run_check()

只有这一步确认是 GPU 可用之后，再用 EngineType.PADDLE 测 RapidOCR。否则 RapidOCR 层面的参数改了也可能还是落回 CPU/ONNX。

如果这样解决了，请把有帮助的回复标记为 answer，方便其他 RTX 50 系用户找到。

[SWHL]

@yudin-s 太感谢你的回答了，我看着很清楚，很详细。你是会中文吗？我看你是俄罗斯人。

[zhenglipeng]

The system requires installing rapidocr_onnxruntime when I install rapidocr why? @SWHL

[SWHL]

Because you haven't uninstalled other RapidOCR packages besides rapidocr.
You need to first uninstall the rapidocr_onnxruntime package. It is an outdated legacy package and is no longer being maintained.

[zhenglipeng]

Package Version

---

aistudio-sdk 0.3.8
annotated-doc 0.0.4
annotated-types 0.7.0
antlr4-python3-runtime 4.9.3
anyio 4.12.1
astor 0.8.1
bce-python-sdk 0.9.71
blinker 1.9.0
certifi 2026.1.4
chardet 7.4.3
charset-normalizer 3.4.7
click 8.3.3
colorama 0.4.6
colorlog 6.10.1
contourpy 1.3.3
crc32c 2.8
cycler 0.12.1
decorator 5.2.1
distro 1.9.0
easyocr 1.7.2
et_xmlfile 2.0.0
filelock 3.29.0
Flask 3.1.2
flatbuffers 25.12.19
fonttools 4.63.0
fsspec 2026.4.0
future 1.0.0
h11 0.16.0
hf-xet 1.5.0
httpcore 1.0.9
httpx 0.28.1
huggingface_hub 1.14.0
idna 3.11
ImageIO 2.37.3
imagesize 2.0.0
itsdangerous 2.2.0
Jinja2 3.1.6
jiter 0.14.0
kiwisolver 1.5.0
lazy-loader 0.5
Levenshtein 0.27.3
markdown-it-py 4.2.0
MarkupSafe 3.0.3
matplotlib 3.10.9
mdurl 0.1.2
modelscope 1.36.3
mpmath 1.3.0
networkx 3.6.1
ninja 1.13.0
numpy 2.3.5
nvidia-cublas-cu12 12.9.0.13
nvidia-cuda-runtime-cu12 12.9.37
nvidia-cudnn-cu12 9.9.0.52
nvidia-cufft-cu12 11.4.0.6
nvidia-curand-cu12 10.3.10.19
nvidia-cusolver-cu12 11.7.4.40
nvidia-cusparse-cu12 12.5.9.5
nvidia-nvjitlink-cu12 12.9.86
omegaconf 2.3.0
onnxruntime 1.26.0
openai 2.30.0
opencv-contrib-python 4.10.0.84
opencv-python 4.13.0.92
opencv-python-headless 4.13.0.92
openpyxl 3.1.5
opt-einsum 3.3.0
packaging 26.0
paddleocr 3.5.0
paddlepaddle-gpu 3.3.1
paddlex 3.5.2
pandas 3.0.3
pandas-stubs 3.0.0.260204
pillow 12.1.0
pip 26.1.1
prettytable 3.17.0
protobuf 6.33.4
psutil 7.2.2
py-cpuinfo 9.0.0
pyclipper 1.4.0
pycryptodome 3.23.0
pydantic 2.13.4
pydantic_core 2.46.4
Pygments 2.20.0
pyparsing 3.3.2
pypdfium2 5.8.0
pytesseract 0.3.13
python-bidi 0.6.9
python-dateutil 2.9.0.post0
PyYAML 6.0.2
RapidFuzz 3.14.5
rapidocr 3.8.1
requests 2.34.0
rich 15.0.0
ruamel.yaml 0.19.1
safetensors 0.7.0
scikit-image 0.26.0
scipy 1.17.1
setuptools 81.0.0
shapely 2.1.2
shellingham 1.5.4
six 1.17.0
sniffio 1.3.1
sympy 1.14.0
tifffile 2026.3.3
torch 2.8.0+cu129
torchaudio 2.8.0+cu129
torchvision 0.23.0+cu129
tqdm 4.67.3
typer 0.25.1
typing_extensions 4.15.0
typing-inspection 0.4.2
tzdata 2026.2
ujson 5.12.1
urllib3 2.7.0
wcwidth 0.7.0
Werkzeug 3.1.8
wheel 0.46.3
Which one hasn't been uninstalled?

[SWHL]

[zhenglipeng]

PaddlePaddle works perfectly on my machine, as I have already used the GPU version of PaddleOCR