question classifier always very slow #13640

Closed
opened 2026-02-21 19:13:07 -05:00 by yindo · 5 comments
Owner

Originally created by @gsq7474741 on GitHub (May 6, 2025).

Self Checks

  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
  • [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:)
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.3.0

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

Image

✔️ Expected Behavior

done in 2s

Actual Behavior

even the simple binary topic cost 4s

Originally created by @gsq7474741 on GitHub (May 6, 2025). ### Self Checks - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [x] I confirm that I am using English to submit this report (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:) - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.3.0 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce <img width="333" alt="Image" src="https://github.com/user-attachments/assets/188791ff-6844-4025-be90-c8f24a921e85" /> ### ✔️ Expected Behavior done in 2s ### ❌ Actual Behavior even the simple binary topic cost 4s
yindo closed this issue 2026-02-21 19:13:07 -05:00
Author
Owner

@junjiem commented on GitHub (May 6, 2025):

This is because your LLM generation speed is slow.

@junjiem commented on GitHub (May 6, 2025): This is because your LLM generation speed is slow.
Author
Owner

@gsq7474741 commented on GitHub (May 6, 2025):

I tried many models. The best performance is 2.142s@832 tokens with gpt-4.1. How can I reduce the time cost to less than 2s?

@gsq7474741 commented on GitHub (May 6, 2025): I tried many models. The best performance is 2.142s@832 tokens with gpt-4.1. How can I reduce the time cost to less than 2s?
Author
Owner

@kunlun122 commented on GitHub (May 6, 2025):

Try Qwen2.5:7B model

@kunlun122 commented on GitHub (May 6, 2025): Try Qwen2.5:7B model
Author
Owner

@hufangdong86 commented on GitHub (May 29, 2025):

me too

@hufangdong86 commented on GitHub (May 29, 2025): me too
Author
Owner

@Pandaaaa906 commented on GitHub (Feb 12, 2026):

me too
im using doubao-vision-pro-32k
always 6~7s
Image

then i write a langchain version, also using doubao-vision-pro-32k
first time it take 12s,but afterwards it only take ~1 second.
Image

@Pandaaaa906 commented on GitHub (Feb 12, 2026): me too im using doubao-vision-pro-32k always 6~7s <img width="751" height="408" alt="Image" src="https://github.com/user-attachments/assets/6a5b1c17-4e2f-4aad-8ace-abe5a47c9dcd" /> then i write a langchain version, also using doubao-vision-pro-32k first time it take 12s,but afterwards it only take ~1 second. <img width="993" height="407" alt="Image" src="https://github.com/user-attachments/assets/80834537-7e5f-4e42-8012-91be4b855f0f" />
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#13640