pycaptcha's Introduction

PyCAPTCHA 🔍

An End-to-end Pytorch-Lightning implemented CAPTCHA OCR model.
Training 2 epoch under 100k images to get over 96% acc on Val dataset 🤩

with 200k or even more training set you may get >98% acc

INSTALL ⚙️

Step0: Clone the Project

git clone https://github.com/ZiYang-xie/PyCAPTCHA
cd PyCAPTCHA

Step1: Create & Activate Conda Env

conda create -n "PyCaptcha" python=3.7
conda activate PyCaptcha

Step2: Install Pytorch

conda install torch==1.9.0 torchvision==0.10.0 -c pytorch

Step3: Install PIP Requirements

pip install -r requirement.txt

Docment 📃

Checkout the PyCAPTCHA Usage in WIKI Page

Check the Doc

pycaptcha's People

Contributors

Stargazers

Watchers

pycaptcha's Issues

RuntimeError

This error occurred during train:

requested resize to (60, 160) ((60, 160) elements in total), but the given tensor has a size of 3x60x160 (28800 elements). autograd's resize can only change the shape of a given tensor, while preserving the number of elements.

RuntimeError: The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimension 0

My dataset png images may have 4 channel instead of 3. Unbale to fix it. Any help?

Error in absence of GPU

I do not have a GPU on my PC and I want to run the training phase on my PC. There is no parameter to specify whether to use GPU or CPU.
Also, why not give a pre-trained network so that not everyone has to train the network independently?
Please help me with this issue.
Thanks

Predict directory is not found

! python predictor.py --input "/content/download (1) (1).png" # Start Predicting

FileNotFoundError: [Errno 2] No such file or directory: '/content/PyCAPTCHA/logs/main/version_2/checkpoints/epoch=4-step=1954.ckpt'

Recommend Projects

ziyang-xie / pycaptcha Goto Github PK