Hi, I tried test my model but when I run it “deepspeech output_graph.pb quz_alphabet.txt achahakunatam.wav” the error is “Alphabet size does not match loaded model: alphabet has size 258, but model has 29 classes in its output. Make sure you’re passing an alphabet file with the same size as the one used for training”.
The problem is that the alphabet is the same that I used to train my model.
How did you specify the alphabet.txt used for training?
Is output_graph.pb the result of your training and not the release model?
Are you using the 0.1.1 deepspeech or did you compile one on you own?
Hi–
I’m very new here so this may not be the issue, but you may try the following:
deepspeech output_graph.pb achahakunatam.wav quz_alphabet.txt
The github page suggests the order that you are using, but I think at some point the order of the arguments may have changed.
Hi, I already solved this problem, thank.
Hi, How did you solve this? Even I have the same problem. Can you please elaborate?
Thanks 