mozilla / DeepSpeech

DeepSpeech is an open source embedded (offline, on-device) speech-to-text engine which can run in real time on devices ranging from a Raspberry Pi 4 to high power GPU servers.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Mandarin Chinese garbled

wingdi opened this issue · comments

I use pre-trained model deepspeech-0.9.3-models-zh-CN.pbmm, it generated Garbled in GBK or UTF-8 ;
is there any method to repair this ?

this is the generate json :
QQ图片20210314104814

when i print the word value:
word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94'

print(word)

it got :
UnicodeEncodeError: 'utf-8' codec can't encode characters in position 0-15: surrogates not allowed

os : win10 ; python Server ; Pycharm(UTF-8)

Why this issue closed ? I didn't find solution in Discourse. and other similar questions in Discouse and here didn't get a solution. I think this is a bug , it need to be fixed ...

Why this issue closed ? I didn't find solution in Discourse. and other similar questions in Discouse and here didn't get a solution. I think this is a bug , it need to be fixed ...

Please follow the guidelines for reaching for support and use discourse.

word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94'

I'm no mandarin speaker but this looks as valid utf8 for that lang.

UnicodeEncodeError: 'utf-8' codec can't encode characters in position 0-15: surrogates not allowed

This sounds more of a python / utf8 / Windows issue.

Since you dont care about sharing requested infos i can only speculate on the code you wrote but this seems to be your issue and valid solution https://stackoverflow.com/a/62935174

Since you dont care about sharing requested infos i can only speculate on the code you wrote but this seems to be your issue and valid solution https://stackoverflow.com/a/62935174

Thanks for your reply
word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94',
this is invalid to transcode to mandarin .

https://stackoverflow.com/a/62935174, this link don't give a solution.\

Since you dont care about sharing requested infos i can only speculate on the code you wrote but this seems to be your issue and valid solution https://stackoverflow.com/a/62935174

Thanks for your reply
word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94',
this is invalid to transcode to mandarin .

As I said, I'm no speaker, but this looks to be a valid python string.

https://stackoverflow.com/a/62935174, this link don't give a solution.\

Again, please use Discourse.

Since you dont care about sharing requested infos i can only speculate on the code you wrote but this seems to be your issue and valid solution https://stackoverflow.com/a/62935174

Thanks for your reply
word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94',
this is invalid to transcode to mandarin .

As I said, I'm no speaker, but this looks to be a valid python string.

https://stackoverflow.com/a/62935174, this link don't give a solution.\

Again, please use Discourse.

I am almost pretty sure '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94' not valid.
this mandarin encode range using Unicode :
Unicode-中文编码范围

thanks.

Since you dont care about sharing requested infos i can only speculate on the code you wrote but this seems to be your issue and valid solution https://stackoverflow.com/a/62935174

Thanks for your reply
word = '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94',
this is invalid to transcode to mandarin .

As I said, I'm no speaker, but this looks to be a valid python string.

https://stackoverflow.com/a/62935174, this link don't give a solution.\

Again, please use Discourse.

I am almost pretty sure '\udce5\udcb8\udce5\udce5\udcb9\udcb3\udce6\udcba\udca6\udce5\udc8f\udce6\udcb1\udc89\udce5\udc94' not valid.
this mandarin encode range using Unicode :
Unicode-中文编码范围

thanks.

Maybe. But until you follow the discourse steps to reach for support https://discourse.mozilla.org/t/what-and-how-to-report-if-you-need-support/62071:

  • we don't know what you did,
  • we don't know how you did,
  • other people have reported successfully using the mandarin model,
  • this is provided as experimental,

There's basically nothing I can do to help. Please follow the steps on Discourse.