Add and manage evaluation results in Hugging Face model cards. Supports extracting eval tables from README content, importing scores from Artificial Analysis API, and running custom model evaluations with vLLM/lighteval. Works with the model-index metadata format.

在 Hugging Face 模型卡片中添加和管理评估结果。支持从 README 内容中提取评估表格、从 Artificial Analysis API 导入评分，以及使用 vLLM/lighteval 运行自定义模型评估。兼容 model-index 元数据格式。