funcation calling failed for xinference with qwen model

[cyflhn]

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

0.10.

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

bug for ollama provider:
1. add ollama provider in dify model settings.
2. launch a model with ollama
3. create an agent app in dify
4. check agent mode for this app.

bug for xinference provider:

1. add xinference provider in dify model settings
2. launch a qwen2.5 model with xinference.
3. create an agent app in dify
4. add a tool in this agent app.
5. run this agent app.

✔️ Expected Behavior

for ollama provider
the agent mode should be function calling.
for xinference provider
the agent app runs ok with correct function calling result.

❌ Actual Behavior

for ollama provider
the agent mode is always reAct.
for xinference provider
the agent app could not run with tool integrated. There is an error showing that the parameter was incorrect with function calling for qwen model.

[dosubot]

Hello @cyflhn! I'm here to help you with any bugs, questions, or contributions you have for Dify. Let's work together to solve this issue.

For the issues you're experiencing:

1. Ollama Provider: The Ollama provider in Dify's agent mode does not support function calling, which is why the agent mode defaults to ReAct. This is a known limitation when using the Ollama provider ^[1].

2. Xinference Provider with Qwen Model: There is a known issue with function calling for the qwen model in the xinference provider, where incorrect parameter errors can occur. Unfortunately, specific troubleshooting steps or solutions for this issue are not detailed in the available documentation. You might want to check if there are any updates or patches that address this problem in the repository or related discussions ^[2].

If you haven't already, you might want to look into any recent updates or community discussions that might provide a workaround or solution for these issues.

To continue talking to Dosu, mention @dosu.

[crazywoola]

Please report these in 2 issues, or you can just remove Ollama part. We have a pull request opened #10784 already. You can remove things related in you pull request.

[cyflhn]

Please report these in 2 issues, or you can just remove Ollama part. We have a pull request opened #10784 already. You can remove things related in you pull request.

I have also make a pr related to this bug

[dosubot]

Hi, @cyflhn. I'm Dosu, and I'm helping the Dify team manage their backlog. I'm marking this issue as stale.

Issue Summary

• You reported a bug with the xinference provider when using the qwen model in a self-hosted Dify setup.
• The issue involves errors due to incorrect parameters for function calling.
• @crazywoola suggested separating issues or removing the Ollama part, referencing pull request feat: support function call for ollama block chat api #10784.
• You confirmed making a pull request related to this bug.

Next Steps

• Please let me know if this issue is still relevant to the latest version of the Dify repository by commenting here.
• If there is no further activity, this issue will be automatically closed in 15 days.

Thank you for your understanding and contribution!