About accuracy in a sentences ?

[NguyenThaiHoc1]

I used your code which you talked achieved is 0.84 (CER). But when I run yes CER is 0.81 ~ 0.84 but the sentence error rate is very low and appears overfitting on your model? You can expand my problem ???
Thank you for reading
Thai Hoc

[lamhoangtung]

Hi, BiLTSM with CTC with many epochs can overfit easily. In order to train a more general OCR:

• Use as much as data as you can, apply image augmentation
• Try to pretrained the model on a synthesis dataset and then fine tune it on your smaller real dataset.

Hope this could help ;)

[lamhoangtung]

@NguyenThaiHoc1 May I close this issue ?