EchoVLM 2.0
Technical Report: Coming Soon! 🚧
EchoVLM 2.0 is a major update to our vision-language model, introducing a brand-new Agent architecture. This version is designed to enhance autonomous problem-solving and tool-use capabilities, pushing the boundaries of what VLMs can achieve in complex visual tasks.
🌟 Key Features
- Brand-New Agent Architecture: Built from the ground up for advanced agentic workflows.
- Versatile Task Support: Compared to V1, EchoVLM 2.0 now supports a wider range of tasks, including:
- Multiple-choice questions
- True/False questions
- Open-ended question answering
- Enhanced Instruction Following: Significantly improved capability to understand and follow complex user instructions accurately.
- Tool Integration: Currently, the model officially supports the
ImageZoomInTool(click here to view the implementation details).
🛠️ Usage
For detailed usage instructions, prompt formatting, and integration code, please refer to the official Qwen3.5-9B Model Card as a baseline reference.
System Prompt
When using this model, we highly recommend using the following system prompt to ensure the reasoning process and final answers are properly formatted:
You are a helpful assistant. Please solve the problem step by step. Put your reasoning process inside <think></think> tags, and provide the finally answer inside <answer></answer> tags.
- Downloads last month
- 41