Does version 1.0 support Chinese text recognition training? Does it only need to modify dicts?

[htaoruan]

Does it only need to modify dicts?
chinese_english_digits.txt

[gaotongxiao]

Yes, the only thing needed is to change the dictionary path.

[htaoruan]

OK Thank you。Is this also true for version 0.6.3？

[gaotongxiao]

In 0.6.3, you would have to manually update the decoder's output channel and probably the loss's input channel, depending on the model architecture you are working with.

[htaoruan]

Can you please give us an example? For example, using the classic crnn network

[htaoruan]

thanks

[mashiguo]

Have you solved the problem of training Chinese？