https://github.com/songquanpeng/crack-captcha

Crack captcha of http://dyxt.jw.scut.edu.cn/
https://github.com/songquanpeng/crack-captcha

Last synced: 4 months ago
JSON representation

Crack captcha of http://dyxt.jw.scut.edu.cn/

• Host: GitHub
• URL: https://github.com/songquanpeng/crack-captcha
• Owner: songquanpeng
• Created: 2021-08-22T02:37:01.000Z (about 4 years ago)
• Default Branch: master
• Last Pushed: 2021-08-22T11:53:35.000Z (about 4 years ago)
• Last Synced: 2025-05-07T05:32:19.451Z (5 months ago)
• Language: Python
• Homepage:
• Size: 588 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Crack Captcha

The captcha we about to crack is very simple, for example: ![example](archive/example.gif)

## Model Accuracy History

|Commit Hash|Accuracy|Remark|

|---|---|---|

|`f15750`|50.5208%|Batch size 64, 20k.|

|`f15750`|55.4688%|Batch size 128, 20k.|

|`not save`|48.4375%|Disable random translate, 20k.|

|`bf6b08`|45.3125%|Use VGG16, 10k.|

The loss don't drop, always ~2.7. :(

## TODOs

+ [x] Construct the model.

+ [x] Crawl the captcha dataset.

+ [x] Label the dataset.

+ [x] Make it runnable.

+ [x] Finish the model training.

+ [ ] Construct API for the trained model.

+ [ ] Data augmentation: add salt & pepper noise.

## Others

I referred [xmcp/elective-dataset-2021spring](https://github.com/xmcp/elective-dataset-2021spring) when crafting this project.