search

weinman / cnn_lstm_ctc_ocr

Tensorflow-based CNN+LSTM trained with CTC-loss for OCR
GNU General Public License v3.0
498 stars 170 forks source link

Can we use new training data? #19

Closed zsun227 closed 6 years ago

zsun227 commented 6 years ago

Hi, I was just wondering that can we use our own dataset(the images and labels may be different)? But it looks like your code already has a pre-trained model.....Do we need to delete that pre-trained model? Since that pre-trained model may not be helpful for the new training set...

Thanks!

weinman commented 6 years ago

Yes, you can rewire the code to use your own data; it should not be much trouble. You'd need to update train.py and perhaps the dimensions of model.py in a few places.

(Also, the repository does not come with a pretrained model, only a tiny fraction of prepopulated data.)