Skip to content

Latest commit

 

History

2 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

README

Mingxi Chen, Juan Li

Pre-trained part

The code used for pre-trained model is in the folder pretrain.

model.py is the model file and train.py is used to run and save models.

netD_490.pth is the saved weights file of discriminative model.

netG_490.pth is the saved weights file of generative model.

Downstream task part

The code used for the training of downstream task is in the folder downstream.

models/crnn_pre.py , train_pre.py are uesd to train our model with pre-trained.

models/crnn.py, train_crnn.py are used to train the pure CRNN model without pre-trained.

pre_crnn.pth is the saved weights file of model with pre-trained.

crnn.pth is the saved weights file of model without pre-trained.

About

Use the method of self-supervised learning to train a pre-trained model and applied it to the CRNN model to train a model that can recognize the Chinese characters sequence in the signboard with accuracy of 0.80.

Resources

Stars

1 star

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages