PaddleX插件方式安装PaddleOCR，推理运行时间很长

[Longdexin]

🔎 Search before asking

• I have searched the PaddleOCR Docs and found no similar bug report.
• I have searched the PaddleOCR Issues and found no similar bug report.
• I have searched the PaddleOCR Discussions and found no similar bug report.

🐛 Bug (问题描述)

1. 产线模式运行，设置device=‘cpu'，识别一张图片需要38秒，不设置device，则需要10秒；
2. 普通模式运行，需要20秒；
3. 这些运行时间都太长了，感觉不太对。

🏃‍♂️ Environment (运行环境)

OS: Ubuntu 24.04.1 LTS
Mem: 32G
Python: 3.10.16
CPU: AMD Ryzen™ 9 5900X × 24

Package                   Version             Editable project location
------------------------- ------------------- -----------------------------------------
aiofiles                  24.1.0
aiohappyeyeballs          2.4.4
aiohttp                   3.11.11
aiosignal                 1.3.2
aistudio-sdk              0.2.6
albucore                  0.0.13
albumentations            1.4.10
annotated-types           0.7.0
anyio                     4.6.2.post1
argon2-cffi               23.1.0
argon2-cffi-bindings      21.2.0
arrow                     1.3.0
astor                     0.8.1
asttokens                 3.0.0
async-lru                 2.0.4
async-timeout             4.0.3
asyncio-atexit            1.0.1
attrdict3                 2.0.2
attrs                     24.3.0
babel                     2.16.0
backoff                   2.2.1
bce-python-sdk            0.9.25
beautifulsoup4            4.12.3
bleach                    6.2.0
blinker                   1.9.0
cachetools                5.5.0
certifi                   2024.8.30
cffi                      1.17.1
chardet                   5.2.0
charset-normalizer        3.4.1
chinese_calendar          1.10.0
click                     8.1.8
colorama                  0.4.6
colorlog                  6.9.0
comm                      0.2.2
cryptography              44.0.0
cssselect                 1.2.0
cssutils                  2.11.1
cycler                    0.12.1
Cython                    3.0.11
dataclasses-json          0.6.7
datasets                  3.2.0
debugpy                   1.8.11
decorator                 5.1.1
defusedxml                0.7.1
dill                      0.3.4
easydict                  1.13
editdistance              0.8.1
emoji                     2.14.0
erniebot                  0.5.0
erniebot_agent            0.5.0
et_xmlfile                2.0.0
eval_type_backport        0.2.2
exceptiongroup            1.2.2
executing                 2.1.0
faiss-cpu                 1.8.0.post1
fastapi                   0.115.6
fastjsonschema            2.21.1
filelock                  3.16.1
filetype                  1.2.0
fire                      0.7.0
Flask                     3.1.0
flask-babel               4.0.0
fonttools                 4.55.3
fqdn                      1.5.1
frozenlist                1.5.0
fsspec                    2024.9.0
future                    1.0.0
gast                      0.3.3
GPUtil                    1.4.0
greenlet                  3.1.1
grpcio                    1.51.3
h11                       0.14.0
html5lib                  1.1
httpcore                  1.0.6
httpx                     0.27.2
huggingface-hub           0.27.0
idna                      3.10
imageio                   2.36.1
imagesize                 1.4.1
imgaug                    0.4.0
ipykernel                 6.29.5
ipython                   8.31.0
isoduration               20.11.0
itsdangerous              2.2.0
jedi                      0.19.2
jieba                     0.42.1
Jinja2                    3.1.5
joblib                    1.4.2
json5                     0.10.0
jsonpatch                 1.33
jsonpath-python           1.0.6
jsonpointer               3.0.0
jsonschema                4.23.0
jsonschema-path           0.3.3
jsonschema-specifications 2023.12.1
jupyter_client            8.6.3
jupyter_core              5.7.2
jupyter-events            0.11.0
jupyter-lsp               2.2.5
jupyter_server            2.15.0
jupyter_server_terminals  0.5.3
jupyterlab                4.3.4
jupyterlab_pygments       0.3.0
jupyterlab_server         2.27.3
kiwisolver                1.4.8
langchain                 0.1.5
langchain-community       0.0.17
langchain-core            0.1.23
langdetect                1.0.9
langsmith                 0.0.87
lapx                      0.5.11.post1
lazy_loader               0.4
lazy-object-proxy         1.10.0
lmdb                      1.5.1
lxml                      5.3.0
markdown-it-py            3.0.0
MarkupSafe                3.0.2
marshmallow               3.23.2
matplotlib                3.5.2
matplotlib-inline         0.1.7
mdurl                     0.1.2
mistune                   3.0.2
more-itertools            10.5.0
motmetrics                1.4.0
multidict                 6.1.0
multiprocess              0.70.12.2
mypy-extensions           1.0.0
nbclient                  0.10.2
nbconvert                 7.16.4
nbformat                  5.10.4
nest-asyncio              1.6.0
networkx                  3.4.2
nltk                      3.9.1
notebook_shim             0.2.4
numpy                     1.24.4
olefile                   0.47
onnx                      1.17.0
openapi-schema-validator  0.6.2
openapi-spec-validator    0.7.1
opencv-contrib-python     4.10.0.84
opencv-python             4.5.5.64
opencv-python-headless    4.10.0.84
openpyxl                  3.1.5
opt-einsum                3.3.0
overrides                 7.7.0
packaging                 23.2
paddle2onnx               1.3.1
paddleclas                2.6.0
paddledet                 0.0.0
paddlefsl                 1.1.0
paddlenlp                 2.8.0.post0
paddleocr                 0.1.0.dev1+g3f32858
paddlepaddle              3.0.0b2
paddlex                   3.0.0b2             /home/longdexin/python-venvs/demo/PaddleX
pandas                    2.2.3
pandocfilters             1.5.1
Parsley                   1.3
parso                     0.8.4
pathable                  0.4.3
pexpect                   4.9.0
pillow                    11.0.0
pip                       24.3.1
platformdirs              4.3.6
premailer                 3.10.0
prettytable               3.12.0
prometheus_client         0.21.1
prompt_toolkit            3.0.48
propcache                 0.2.1
protobuf                  5.28.3
psutil                    5.9.8
ptyprocess                0.7.0
pure_eval                 0.2.3
pyarrow                   18.1.0
pybind11                  2.13.6
pyclipper                 1.3.0.post6
pycocotools               2.0.8
pycparser                 2.22
pycryptodome              3.21.0
pydantic                  2.9.2
pydantic_core             2.23.4
Pygments                  2.18.0
PyMuPDF                   1.25.1
pypandoc                  1.14
pyparsing                 3.2.0
pypdf                     5.1.0
python-dateutil           2.9.0.post0
python-docx               1.1.2
python-iso639             2024.10.22
python-json-logger        3.2.1
python-magic              0.4.27
python-oxmsg              0.0.1
pytz                      2024.2
PyYAML                    6.0.2
pyzmq                     26.2.0
qianfan                   0.0.3
RapidFuzz                 3.11.0
rarfile                   4.2
referencing               0.35.1
regex                     2024.11.6
requests                  2.32.3
requests-toolbelt         1.0.0
rfc3339-validator         0.1.4
rfc3986-validator         0.1.1
rich                      13.9.4
rpds-py                   0.22.3
ruamel.yaml               0.18.6
ruamel.yaml.clib          0.2.12
safetensors               0.4.5
scikit-image              0.25.0
scikit-learn              1.6.0
scipy                     1.14.1
Send2Trash                1.8.3
sentencepiece             0.2.0
seqeval                   1.2.2
setuptools                65.5.0
shapely                   2.0.6
shellingham               1.5.4
six                       1.17.0
sklearn                   0.0
sniffio                   1.3.1
soupsieve                 2.6
SQLAlchemy                2.0.36
stack-data                0.6.3
starlette                 0.41.3
tenacity                  8.5.0
termcolor                 2.5.0
terminado                 0.18.1
terminaltables            3.1.10
threadpoolctl             3.5.0
tifffile                  2024.12.12
tinycss2                  1.4.0
tokenizers                0.19.1
tomark                    0.1.4
tomli                     2.2.1
tool_helpers              0.1.2
tornado                   6.4.2
tqdm                      4.67.1
traitlets                 5.14.3
typeguard                 4.4.1
typer                     0.15.1
types-python-dateutil     2.9.0.20241206
typing_extensions         4.12.2
typing-inspect            0.9.0
tzdata                    2024.2
ujson                     5.10.0
unstructured              0.16.11
unstructured-client       0.28.1
uri-template              1.3.0
urllib3                   2.3.0
uvicorn                   0.34.0
visualdl                  2.5.3
wcwidth                   0.2.13
webcolors                 24.11.1
webencodings              0.5.1
websocket-client          1.8.0
Werkzeug                  3.1.3
wget                      3.2
wrapt                     1.17.0
xmltodict                 0.14.2
xxhash                    3.5.0
yacs                      0.1.8
yarl                      1.18.3

🌰 Minimal Reproducible Example (最小可复现问题的Demo)

OCR.yaml

Global:
  pipeline_name: OCR
  input: https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_001.png
  
Pipeline:
  text_det_model: PP-OCRv4_server_det
  text_rec_model: PP-OCRv4_server_rec
  text_rec_batch_size: 1

代码

import cv2
import time
from paddlex import create_pipeline
from paddleocr import PaddleOCR

pp_ocr=PaddleOCR(use_angle_cls=False)
ocr_pipeline = create_pipeline(pipeline='OCR.yaml',device='cpu')
for i in range(5):
    time1 = time.perf_counter()
    img = cv2.imread('FM00064655_7.jpg')
    height, width, _ = img.shape
    ocr_output = ocr_pipeline.predict(img)
    ocr_data=[]
    for res in ocr_output:
        for idx in range(len(res.json['dt_polys'])):
            box = res.json['dt_polys'][idx]
            score = res.json['rec_score'][idx]
            text = res.json['rec_text'][idx]
            ocr_data.append({'text':text,'score':score,'box':box})
    time2 = time.perf_counter()
    print(f"ocr1: {i}|{time2-time1} 秒")
    ocr_result = pp_ocr.ocr(img, cls=False)
    ocr_data=[]
    for idx in range(len(ocr_result)):
        res = ocr_result[idx]
        for line in res:
            box,text,score=line[0],line[1][0],line[1][1]
            ocr_data.append({"text":text,"score":score,"box":box})
    time3 = time.perf_counter()
    print(f"ocr2: {i}|{time3-time2} 秒")

[cuicheng01]

是否方便提供一下你的图像呢？

[Longdexin]

@cuicheng01 ，您好，非常感谢！图片是同事做的实验结果，不太适合贴出来，用这张图也是一样的，运行时间很长，一个8秒，另一个1秒多，https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_001.png