ollama - 💡(How to fix) Fix Tool calling fails or is inconsistent with qwen models [1 participants]

TL;DR

The issue may be related to compatibility problems between Qwen models and the current implementation, suggesting a need to review or adjust the integration with Qwen models.

Guidance

• Review the documentation and release notes of Qwen models (versions 2.5, 3, 3.5, 3.6) for any changes or specific requirements that might affect the function calling process.
• Compare the implementation details between Qwen and gemma models to identify any differences that could cause the malfunction.
• Test the function calling process with different Qwen models to see if the issue is consistent across all versions or specific to certain ones.
• Consider reaching out to the Qwen community or support for guidance on integrating their models, as the issue seems isolated to Qwen.

Notes

The lack of log output, OS, GPU, CPU, and Ollama version information limits the ability to provide a more specific diagnosis or solution. The guidance provided is based on the information given and may need adjustment based on additional details.

Recommendation

Apply workaround: Given the compatibility issues with Qwen models and the success with gemma models, it might be necessary to implement model-specific handling or adjustments in the function calling process until a more universal solution is found.