CNN for solving TikTok captcha (NOT FINISHED)

Train Pipeline

1. Get a lot of data e.g Imagenet
2. Create Function for online(during training) generating Data
3. Use some CNN as Feature extractor
4. Train classifier

Classify as 0 - no angle rotate
            1 - image rotated

Classifier with 360 labels does not work

Result

First test model was trained with ResNet-18 and got:
* HIGHREST ACCURACY: 0.7892497518082542

0. Example image

1. Preprocessed Image

Angle = 90

2. Base Loss

3. Train/Test loss

4. Accuracy of model

Repo uses resources from:

https://github.com/DFGANDP/Pytorch-EarlyStopping

Currently TODO:

1. Train on more Data
2. Use more complex models (e.g. Vision Transformers)
3. Write Real Time solver With Selenium?