post_train

docs(posttrain): 添加法演化关系图及方法家族注释

docs(demo): 补充各训练方法的数据格式及示例说明

fix(env): 设置国内访问Huggingface的镜像源

Update README.md

Update README.md

Initial: 18-method post-training demo with uv (Qwen2.5-0.5B)