Closed AndresPMD closed 6 years ago
Hello Pirate,
On 05/28/2018 05:13 PM, DreadPiratePsyopus wrote:
Hello @MichalBusta https://github.com/MichalBusta , At the main page it says that "For decoding final output, we provide just greedy and dictionary based prefix decoding" What do you mean by greedy prefix decoding? Is it basically to recognize out of vocabulary words?
it is not greedy prefix decoding but greedy and prefix decoding: - greedy = "incorrect" way how to decode softmax = take local maximum (correct way is to do full dyn. programing, but it brings you just 1% improvement) - prefix decoding = beam search with dictionary = take local maximum which leads to leaf node (dict. word)
Thanks for your anwer.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/MichalBusta/DeepTextSpotter/issues/40, or mute the thread https://github.com/notifications/unsubscribe-auth/AD6jsDhi6OW8N13CfJaHwksQriJ2q5a1ks5t3BQfgaJpZM4UQVPt.
Hello @MichalBusta , At the main page it says that "For decoding final output, we provide just greedy and dictionary based prefix decoding" What do you mean by greedy prefix decoding? Is it basically to recognize out of vocabulary words?
Thanks for your anwer.