This Skill is specifically designed for the Alibaba Cloud Tongyi Qianwen (Qwen) series of models, providing auxiliary tools for multimodal input (Qwen-VL) and long-text processing.
Capability Description
- Multimodal Adaptation: Automatically generate Qwen-VL's mixed image/text Prompt format (such as
<img>url</img>）。 - Prompt Optimization: Tune Prompts for Qwen's instruction-following capabilities.
Typical Usage
- Generate Prompts for Visual Tasks：
python3 .trae/skills/qwen-enhance/scripts/qwen_vl.py "描述这张图片" "https://example.com/image.jpg"
Output Example
<img>https://example.com/image.jpg</img> 描述这张图片