I am trying to use deep speech to train mixed-code speech recognition which the sentence would mainly has English but some Chinese characters.
I read lots thread that I might can use transfer learning branch to make it work, but I wonder has anyone done this yet? I want the output is mixed-code text too. Thank you
