Created by: ganji15

I want to label the input sequence after I have trained a CTC network, but I have no idea to label the sequences. Acrodding to the related papers, the task is simliar with path decoding. There are 2 method to decode: "best path decoding" and "prefix search decoding" . So, how can I find most probable labelling with "best path decoding" and "prefix search decoding" in Paddle, or I have to do this labeling by meself?

Can anyone provide any suggestions or examples? Thank you?

ctc network

# Should the output of tmp be a sequnce of  probability vector such as [[f,...], [f,....]]?
tmp = fc_layer(input=x1,
              size=num_classes + 1,
              act=SoftmaxActivation())

ctc = ctc_layer(input=tmp,
                 label=y,
                 size=num_classes + 1)
ctc_eval = ctc_error_evaluator(input=ctc, label=y)