Skip to content

海光黑客松 初赛 / 打卡 赛题#1239

Merged
luotao1 merged 6 commits intoPaddlePaddle:masterfrom
Yun1Liu:master
Mar 19, 2026
Merged

海光黑客松 初赛 / 打卡 赛题#1239
luotao1 merged 6 commits intoPaddlePaddle:masterfrom
Yun1Liu:master

Conversation

@Yun1Liu
Copy link
Copy Markdown
Contributor

@Yun1Liu Yun1Liu commented Mar 13, 2026

海光赛道的初赛赛题,请审核 :)

@paddle-bot
Copy link
Copy Markdown

paddle-bot Bot commented Mar 13, 2026

你的PR提交成功,感谢你对开源项目的贡献!
请检查PR提交格式和内容是否完备,具体请参考示例模版
Your PR has been submitted. Thanks for your contribution!
Please check its format and content. For this, you can refer to Template and Demo.

@CLAassistant
Copy link
Copy Markdown

CLAassistant commented Mar 16, 2026

CLA assistant check
All committers have signed the CLA.

Copy link
Copy Markdown
Collaborator

@luotao1 luotao1 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  1. 请签署CLA
  2. 图片请统一放到images目录下
  3. 打卡md里需要附上截图链接和说明

onecatcn
onecatcn previously approved these changes Mar 16, 2026
Copy link
Copy Markdown
Collaborator

@onecatcn onecatcn left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@Yun1Liu
Copy link
Copy Markdown
Contributor Author

Yun1Liu commented Mar 18, 2026

@luotao1 你好,我更新过了,请review。


## 打卡任务
### 编写PaddleOCR应用
- 任务:下载 https://github.com/PaddlePaddle/community/pfcc/paddle-hardware/images/dcu/ 下的10张图片,编写python代码,调用paddleocr-vl 1.5的模型,把他扫描成文本。
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这10张图片仅作为输入的话,可以也放在你们的数据集内,没有必要上传到仓库里。类似

下载数据集:https://opendatalab.com/OpenDataLab/MSRA-TD500

```

## 打卡任务
### 编写PaddleOCR应用
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

打卡任务两个都要做的话,请改小标题

  • 第一步:编写PaddleOCR应用
  • 第二步:使用vllm后端加速OCR


### 使用vllm后端加速OCR
- 下载数据集:https://opendatalab.com/OpenDataLab/MSRA-TD500 ,扫描test下的200张图片。要求通过启动vllm后端的方式,通过打高batch的方式尽可能得到一个较高的性能。
- 打卡提交物:代码(python程序,后缀名改为'py_'),运行结果(txt文件),性能数据(屏幕截图)
Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

性能数据请上传屏幕截图,做为示例

@luotao1 luotao1 merged commit c98b5f5 into PaddlePaddle:master Mar 19, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants