I read the thesis by Alex graves on RNN LSTM CTC, and currently I am studying lstm.py and i ran code for training english alphabets for OCR purpose !!!
My query is with regards to implementation of CTC layer and decoding algorithm such as prefix decoding or token passing algorithm as specified by Thesis paper I am unable to locate decoding part in lstm.py !!!! Any information regards to this will be helpful On Monday, May 19, 2014 at 5:13:10 PM UTC+5:30, sudeep raja putta wrote: > > I am trying to understand and reverse engineer LSTM in lstm.py. But there > are many attributes with cryptic names > > What do the following names represent? > > WGI WGF WGO WIP WFP WOP cix ci gix gi gox go gfx gf > source state output gierr gferr goerr cierr stateerr > -- You received this message because you are subscribed to the Google Groups "ocropus" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/ocropus/97a08b65-e027-4a78-ba18-d5e9d5da79e3%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
