OS Ubuntu 22.04
Environment Anaconda3
Python 3.10
PaddleOCR 3.5.0
Install pip
RAM 16GB
CPU R5 5600x
CUDA 11.8
GPU GTX 1080 Ti
VRAM 11264MiB
from paddleocr import PaddleOCR
ocr = PaddleOCR(
#lang="chinese_cht",
#use_angle_cls=True,
#ocr_version="PP-OCRv3",
use_doc_orientation_classify=True,
use_textline_orientation=True,
return_word_box=True, # 返回词级框
#return_char=True, # 返回字符级四角坐标
enable_hpi=True,
#use_space_char=True,
device="gpu:0",
use_tensorrt=True,
use_doc_unwarping=False,
text_det_thresh=0.1,
text_det_box_thresh=0.1,
text_det_unclip_ratio=0.75
#device='cpu'
)
result = ocr.predict(input="0000570.png")
for line in result:
line.save_to_json('000057.json')
line.save_to_img("000057.png")
🔎 Search before asking
🐛 Bug (问题描述)
没有文字输出。
000057.json


🏃♂️ Environment (运行环境)
🌰 Minimal Reproducible Example (最小可复现问题的Demo)