Add adaptation for the Doubao-Seed-1.6-vision model

[lupengfei-johnson]

I tried adding code to integrate the Doubao-Seed-1.6-vision model into the project. Some preliminary tests have been conducted, and the Doubao-Seed-1.6-vision model is performing well so far. Chinese users can opt for the Doubao-Seed-1.6-vision model, which offers better network connectivity and more affordable pricing compared to models like OpenAI.

Here is the official website and model introduction for Doubao-Seed-1.6-vision:
https://console.volcengine.com/ark/region:ark+cn-beijing/model/detail?Id=doubao-seed-1-6-vision

Doubao-Seed-1.6-vision | A New-Generation Visual Deep Thinking Model

Doubao-Seed-1.6-vision is designed for highly complex scenarios such as video understanding, visual grounding, and GUI agent applications.
Compared to Doubao-1.5-thinking-vision-pro, it demonstrates stronger general multimodal comprehension and reasoning capabilities in areas such as education, image moderation, inspection and security, and AI-powered search and Q&A.