doc

Author: hjh0119

docs/source/Instruction/GRPO/DeveloperGuide/reward_function.md

@@ -81,6 +81,8 @@ swift 支持同时使用同步和异步奖励函数。训练器会自动检测
 - 同步奖励函数按顺序执行
 - 异步奖励函数使用 `asyncio.gather` 并行执行
 
+[plugin](https://github.com/modelscope/ms-swift/blob/main/examples/train/grpo/plugin/plugin.py)文件中提供了一个调用`swift deploy`服务的生成式奖励模型的例子(async_genrm)
+
 ## 内置奖励函数
 swift内置了五种基于规则的奖励函数(代码见swift/plugin/orm.py)
 

docs/source_en/Instruction/GRPO/DeveloperGuide/reward_function.md

@@ -80,6 +80,8 @@ Swift supports using both synchronous and asynchronous reward functions simultan
 - Synchronous reward functions are executed sequentially
 - Asynchronous reward functions are executed in parallel using `asyncio.gather`
 
+The [plugin](https://github.com/modelscope/ms-swift/blob/main/examples/train/grpo/plugin/plugin.py) file provides an example of a generative reward model (async_genrm) that calls the `swift deploy` service.
+
 ## Built-in Reward Functions
 Swift includes five rule-based reward functions (code can be found in swift/plugin/orm.py).