PaddleOCRVL 使用 python api 使用 GPU 执行推理，但是内存一直升高导致内存溢出 #16971
Description
🔎 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)
我尝试通过手动安装 PaddlePaddle 和 PaddleOCR，希望进行一些的测试。
在程序执行的过程中我发现 RAM 一直在升高直至占满，最后进程由于内存溢出被杀死。
我不太清楚是否是因为我的显存不够的原因？还是说我需要一台更好的设备进行测试？
🏃♂️ Environment (运行环境)
OS Ubuntu24.04
RAM 32GB
GPU NVIDIA GeForce RTX 3060 Ti / 8GB+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.95.05 Driver Version: 580.95.05 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 3060 Ti Off | 00000000:01:00.0 Off | N/A |
| 33% 39C P0 31W / 200W | 0MiB / 8192MiB | 3% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| No running processes found |
+-----------------------------------------------------------------------------------------+Package Version
------------------------ -----------------
aiohappyeyeballs 2.6.1
aiohttp 3.12.13
aiosignal 1.3.2
aistudio-sdk 0.3.8
annotated-types 0.7.0
anyio 4.9.0
attrs 25.3.0
bce-python-sdk 0.9.49
beautifulsoup4 4.13.4
cachetools 6.1.0
certifi 2025.6.15
chardet 5.2.0
charset-normalizer 3.4.2
click 8.2.1
colorlog 6.9.0
conda-pack 0.8.1
cssselect 1.3.0
cssutils 2.11.1
dataclasses-json 0.6.7
decorator 5.2.1
distro 1.9.0
einops 0.8.1
et_xmlfile 2.0.0
fastapi 0.115.14
filelock 3.18.0
frozenlist 1.7.0
fsspec 2025.5.1
ftfy 6.3.1
future 1.0.0
GPUtil 1.4.0
greenlet 3.2.3
h11 0.16.0
hf-xet 1.1.4
httpcore 1.0.9
httptools 0.6.4
httpx 0.28.1
httpx-sse 0.4.0
huggingface-hub 0.33.0
idna 3.10
imagesize 1.4.1
Jinja2 3.1.6
jiter 0.10.0
joblib 1.5.1
jsonpatch 1.33
jsonpointer 3.0.0
langchain 0.3.25
langchain-community 0.3.25
langchain-core 0.3.65
langchain-openai 0.3.24
langchain-text-splitters 0.3.8
langsmith 0.3.45
loguru 0.7.3
lxml 5.4.0
MarkupSafe 2.1.5
marshmallow 3.26.1
modelscope 1.31.0
more-itertools 10.7.0
mpmath 1.3.0
multidict 6.5.0
mypy_extensions 1.1.0
natsort 8.4.0
networkx 3.5
numpy 2.3.0
nvidia-cublas-cu12 12.9.0.13
nvidia-cuda-cccl-cu12 12.9.27
nvidia-cuda-cupti-cu12 12.9.19
nvidia-cuda-nvrtc-cu12 12.9.41
nvidia-cuda-runtime-cu12 12.9.37
nvidia-cudnn-cu12 9.9.0.52
nvidia-cufft-cu12 11.4.0.6
nvidia-cufile-cu12 1.14.0.30
nvidia-curand-cu12 10.3.10.19
nvidia-cusolver-cu12 11.7.4.40
nvidia-cusparse-cu12 12.5.9.5
nvidia-cusparselt-cu12 0.7.1
nvidia-nccl-cu12 2.27.3
nvidia-nvjitlink-cu12 12.9.41
nvidia-nvtx-cu12 12.9.19
openai 1.88.0
opencv-contrib-python 4.10.0.84
opencv-python-headless 4.11.0.86
openpyxl 3.1.5
opt-einsum 3.3.0
orjson 3.10.18
packaging 24.2
paddleocr 3.3.1
paddlepaddle-gpu 3.2.1
paddlex 3.3.6
pandas 2.3.0
pillow 11.2.1
pip 25.1.1
premailer 3.10.0
prettytable 3.16.0
propcache 0.3.2
protobuf 6.31.1
psutil 7.1.3
py-cpuinfo 9.0.0
pyclipper 1.3.0.post6
pycryptodome 3.23.0
pydantic 2.11.7
pydantic_core 2.33.2
pydantic-settings 2.9.1
pypdfium2 4.30.1
python-bidi 0.6.7
python-dateutil 2.9.0.post0
python-dotenv 1.1.0
pytorch-triton 3.3.1+gitc8757738
pytz 2025.2
PyYAML 6.0.2
regex 2024.11.6
requests 2.32.4
requests-toolbelt 1.0.0
ruamel.yaml 0.18.14
ruamel.yaml.clib 0.2.12
safetensors 0.6.2.dev0
scikit-learn 1.7.0
scipy 1.15.3
sentencepiece 0.2.1
setuptools 70.2.0
shapely 2.1.1
six 1.17.0
sniffio 1.3.1
some-package 0.1
soupsieve 2.7
SQLAlchemy 2.0.41
starlette 0.46.2
sympy 1.14.0
tenacity 9.1.2
threadpoolctl 3.6.0
tiktoken 0.9.0
tokenizers 0.21.1
torch 2.8.0+cu129
torchvision 0.23.0+cu129
tqdm 4.67.1
triton 3.4.0
typing_extensions 4.14.0
typing-inspect 0.9.0
typing-inspection 0.4.1
tzdata 2025.2
ujson 5.10.0
urllib3 2.5.0
uvicorn 0.34.3
uvloop 0.21.0
watchfiles 1.1.0
wcwidth 0.2.13
websockets 15.0.1
yarl 1.20.1
zstandard 0.23.0🌰 Minimal Reproducible Example (最小可复现问题的Demo)
import os
import time
from natsort import natsorted
from paddleocr import PaddleOCRVL
ocr = PaddleOCRVL(
layout_detection_model_dir="/mnt/remote/model/.paddlex/official_models/PP-DocLayoutV2",
vl_rec_model_dir="/mnt/remote/model/.paddlex/official_models/PaddleOCR-VL",
use_doc_orientation_classify=False,
use_doc_unwarping=False,
use_chart_recognition=False,
enable_mkldnn=True,
layout_threshold=0.65
)
start_time = time.time()
dir_path = "./_cache/images"
sorted_file = natsorted([
os.path.join(dir_path, f)
for f in os.listdir(dir_path)
if os.path.isfile(os.path.join(dir_path, f))
])
result = ocr.predict_iter(input=sorted_file, use_queues=True)
for idx, res in enumerate(result):
print(res.json)
end_time = time.time()
duration = end_time - start_time
print(f"Time:[{duration}]")Metadata
Metadata
Assignees
Labels
No labels