[GH-ISSUE #21] [Feature request] Chinese support for summarize and classify modules #17

New issue

Closed

opened 2026-10-15 18:03:26 +00:00 by bsakaguchi · 4 comments

bsakaguchi commented

2026-10-15 18:03:26 +00:00

Owner

Originally created by @g51503john02 on GitHub (May 10, 2023).
Original GitHub issue: https://github.com/SillyTavern/SillyTavern-Extras/issues/21

After cross-testing with Chinese and English, I found that the summarize and classify modules do not support Chinese. For the summarize module, the Chinese language makes the summed plot completely irrelevant. For the classify module, the Chinese language will result in completely wrong emotion recognition.

I know that by design, the models or principles of these two modules may only support English, so it would be nice if they could support Chinese or more languages, thank you very much.

Originally created by @g51503john02 on GitHub (May 10, 2023). Original GitHub issue: https://github.com/SillyTavern/SillyTavern-Extras/issues/21 After cross-testing with Chinese and English, I found that the summarize and classify modules do not support Chinese. For the summarize module, the Chinese language makes the summed plot completely irrelevant. For the classify module, the Chinese language will result in completely wrong emotion recognition. I know that by design, the models or principles of these two modules may only support English, so it would be nice if they could support Chinese or more languages, thank you very much.

bsakaguchi closed this issue

2026-10-15 18:03:26 +00:00

bsakaguchi commented

2026-10-15 18:03:27 +00:00

Author

Owner

@Cohee1207 commented on GitHub (May 10, 2023):

I did not train any of these. Please take a look at Huggingface for compatible models and try them out. My readme document has instructions on how to replace BART/BERT models used

@Cohee1207 commented on GitHub (May 10, 2023): I did not train any of these. Please take a look at Huggingface for compatible models and try them out. My readme document has instructions on how to replace BART/BERT models used

bsakaguchi commented

2026-10-15 18:03:27 +00:00

Author

Owner

@g51503john02 commented on GitHub (May 18, 2023):

I found the classify module specifically for Chinese, you can add it to the read me file for others to reference.
touch20032003/xuyuan-trial-sentiment-bert-chinese

@g51503john02 commented on GitHub (May 18, 2023): I found the classify module specifically for Chinese, you can add it to the read me file for others to reference. [touch20032003/xuyuan-trial-sentiment-bert-chinese](https://huggingface.co/touch20032003/xuyuan-trial-sentiment-bert-chinese)

bsakaguchi commented

2026-10-15 18:03:27 +00:00

Author

Owner

@Cohee1207 commented on GitHub (May 18, 2023):

Added

@Cohee1207 commented on GitHub (May 18, 2023): Added

bsakaguchi commented

2026-10-15 18:03:28 +00:00

Author

Owner

@bash99 commented on GitHub (Jun 7, 2023):

After cross-testing with Chinese and English, I found that the summarize and classify modules do not support Chinese. For the summarize module, the Chinese language makes the summed plot completely irrelevant. For the classify module, the Chinese language will result in completely wrong emotion recognition.

I know that by design, the models or principles of these two modules may only support English, so it would be nice if they could support Chinese or more languages, thank you very much.

May I ask which main llm model you used has best chinese support in sillyTavern?

@bash99 commented on GitHub (Jun 7, 2023): > After cross-testing with Chinese and English, I found that the summarize and classify modules do not support Chinese. For the summarize module, the Chinese language makes the summed plot completely irrelevant. For the classify module, the Chinese language will result in completely wrong emotion recognition. > > I know that by design, the models or principles of these two modules may only support English, so it would be nice if they could support Chinese or more languages, thank you very much. May I ask which main llm model you used has best chinese support in sillyTavern?