ollama - 💡(How to fix) Fix 使用ollama api调用qwen3.5时如何关闭思考模式 [4 comments, 4 participants]

Fix Plan

The fix involves updating the thinking chain logic to accommodate the changes in qwen3.5.

Steps to Fix

• Review the thinking chain logic in the codebase and identify areas that need updates.
• Update the /no_think endpoint to handle the new thinking chain logic.
• Add conditional statements to handle different thinking chain versions.

Example Code

# Updated thinking chain logic
def thinking_chain(version):
    if version == "qwen3.5":
        # New thinking chain logic for qwen3.5
        return new_thinking_chain()
    else:
        # Old thinking chain logic
        return old_thinking_chain()

# Updated /no_think endpoint
@app.route("/no_think", methods=["GET"])
def no_think():
    version = request.args.get("version")
    if version == "qwen3.5":
        # Handle new thinking chain logic
        return thinking_chain(version)
    else:
        # Handle old thinking chain logic
        return old_no_think()

Verification

• Test the updated /no_think endpoint with different versions to ensure it handles both old and new thinking chain logic correctly.
• Verify that the new thinking chain logic is correctly implemented for qwen3.5.

Extra Tips

• Ensure to update documentation to reflect changes in the thinking chain logic.
• Consider adding versioning to the API to handle different thinking chain versions more elegantly.