Files
Yi-sir 9e20dab0d6 [Example] Merge Download Paddle Model, Paddle->ONNX->MLIR->BModel (#1643)
* fix infer.py and README

* [Example] Merge Download Paddle Model, Paddle->Onnx->Mlir->Bmodel and
inference into infer.py. Modify README.md

* modify pp_liteseg sophgo infer.py and README.md

* fix PPOCR,PPYOLOE,PICODET,LITESEG sophgo infer.py and README.md

* fix memory overflow problem while inferring with sophgo backend

* fix memory overflow problem while inferring with sophgo backend

---------

Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com>
Co-authored-by: xuyizhou <yizhou.xu@sophgo.com>
2023-03-31 15:08:01 +08:00
..

English | 简体中文

PP-OCRv3 SOPHGO Python部署示例

本目录下提供infer.py快速完成 PP-OCRv3 在SOPHGO TPU上部署的示例。

1. 部署环境准备

在部署前需自行编译基于算能硬件的FastDeploy python wheel包并安装参考文档算能硬件部署环境

2.运行部署示例

2.1 模型准备

将Paddle模型转换为SOPHGO bmodel模型, 转换步骤参考文档

2.2 开始部署

# 下载部署示例代码
git clone https://github.com/PaddlePaddle/FastDeploy.git
cd  FastDeploy/examples/vision/ocr/PP-OCR/sophgo/python

# 如果您希望从PaddleOCR下载示例代码请运行
git clone https://github.com/PaddlePaddle/PaddleOCR.git
# 注意如果当前分支找不到下面的fastdeploy测试代码请切换到dygraph分支
git checkout dygraph
cd PaddleOCR/deploy/fastdeploy/sophgo/python

# 下载图片
wget https://gitee.com/paddlepaddle/PaddleOCR/raw/release/2.6/doc/imgs/12.jpg

# 下载字典文件
wget https://gitee.com/paddlepaddle/PaddleOCR/raw/release/2.6/ppocr/utils/ppocr_keys_v1.txt

# 推理
# --auto True自动完成下载数据、模型转换、推理
python3 infer.py --auto True --det_model '' --cls_model '' --rec_model '' --rec_label_file '' --image ''

# --auto False需要用户设置模型、字典、图片路径进行推理
python3 infer.py --auto False \
                 --det_model ocr_bmodel/ch_PP-OCRv3_det_1684x_f32.bmodel \
                 --cls_model ocr_bmodel/ch_ppocr_mobile_v2.0_cls_1684x_f32.bmodel \
                 --rec_model ocr_bmodel/ch_PP-OCRv3_rec_1684x_f32.bmodel \
                 --rec_label_file ../ppocr_keys_v1.txt \  
                 --image ../12.jpg

# 运行完成后返回结果如下所示
det boxes: [[42,413],[483,391],[484,428],[43,450]]rec text: 上海斯格威铂尔大酒店 rec score:0.952958 cls label: 0 cls score: 1.000000
det boxes: [[187,456],[399,448],[400,480],[188,488]]rec text: 打浦路15号 rec score:0.897335 cls label: 0 cls score: 1.000000
det boxes: [[23,507],[513,488],[515,529],[24,548]]rec text: 绿洲仕格维花园公寓 rec score:0.994589 cls label: 0 cls score: 1.000000
det boxes: [[74,553],[427,542],[428,571],[75,582]]rec text: 打浦路252935号 rec score:0.900663 cls label: 0 cls score: 1.000000

可视化结果保存在sophgo_result.jpg中

3. 其它文档