合成时想采用训练集中的数据的embed数据的均值应该怎样修改ui.py文件

**Summary[问题简述（一句话）]**
A clear and concise description of what the issue is.

**Env & To Reproduce[复现与环境]**
描述你用的环境、代码版本、模型

**Screenshots[截图（如有）]**
If applicable, add screenshots to help

Start by reading ui.py and locating the synthesis entry point and how embeddings are currently selected. The issue provides no reproduction details, model version, or tests; clarify the intended training-set mean behavior first, then verify that synthesis uses it consistently.

Written by the indexing model from the issue text.