-
aistudio 4卡gpu环境,paddleocr 2.6.0
中文古语识别,竖排文字,所以dataset里面的w、h我调换了顺序,设置成了[3, 640, 32] , reconaug设置了[640, 32, 3]
文本最长50字,已经根据gt切成了小图
启动命令
```
!python3 -m paddle.distributed.launch --ips="localhost" …
-
My goal is to finetune the Arabic model:
```https://paddleocr.bj.bcebos.com/PP-OCRv3/multilingual/arabic_PP-OCRv3_rec_train.tar``` on **sentence-level**
This is a sample from the multilingual dat…
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
- 系统环境/System Environment:Windows
- 版本号/Version 10:Paddle:paddlepaddle-gpu 2.4.0.post117
- 运行指令/Command …
-
- System Environment:Ubuntu 18.04 | CUDA 11.0
- Version:Paddle:2.3.0.post110 PaddleOCR:2.5.0.3 Related components:
- Command Code:python -m paddle.distributed.launch --gpus '0' tools/train.py -c…
-
我用的是centos7 ,cuda 10.2 + cudnn 8.2.0,paddle 版本是:paddlepaddle-gpu 2.4.1 (2.2.0也试过,都一样),paddleOCR版本是2.6.1.2。
paddle-gpu 2.4.1就是按照官方教程给的10.2版本cuda安装方案。
另一个issue3345(https://github.com/PaddlePaddle/…
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
- 系统环境/System Environment:docker
- 版本号/Version:Paddle:paddleocr-release-2.5 PaddleOCR: 问题相关组件/Related comp…
-
+ 模型:Rosetta;
+ 字典:数字字典,即 `0,1,2,3,4,5,6,7,8,9`;
+ 训练时文本的长度限制:15;
+ Paddle 版本:2.3;
+ PaddleOCR 版本:2.5。
-
After successful installation I run `example.py` as is and get the error:
```
% python3.8 example.py
[2022/10/26 17:46:34] ppocr DEBUG: Namespace(alpha=1.0, benchmark=False, …
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
- 系统环境/System Environment:
- 版本号/Version:Paddle: PaddleOCR: 问题相关组件/Related components:
- paddlepaddle-gpu…
-
请提供下述完整信息以便快速定位问题/Please provide the following information to quickly locate the problem
rec_svtrnet.yml 配置文件中数据集部分为LMDBDataSet,想修改为SimpleDataSet,如何修改参数
原始参数如下:
Train:
dataset:
name: LM…