Release v0.2.1 (#208)

Author: pan-x-c

README.md

@@ -22,6 +22,8 @@
 
 ## 🚀 News
 
+* [2025-08] ✨ Trinity-RFT v0.2.1 is released with enhanced features for Agentic RL and Async RL.
+* [2025-08] 🎵 We introduce [CHORD](https://github.com/modelscope/Trinity-RFT/tree/main/examples/mix_chord), a dynamic integration of SFT and RL for enhanced LLM fine-tuning ([paper](https://arxiv.org/pdf/2508.11408)).
 * [2025-08] We now support training on general multi-step workflows! Please check out examples for [ALFWorld](./docs/sphinx_doc/source/tutorial/example_step_wise.md) and [ReAct](./docs/sphinx_doc/source/tutorial/example_react.md).
 * [2025-07] Trinity-RFT v0.2.0 is released.
 * [2025-07] We update the [technical report](https://arxiv.org/abs/2505.17826) (arXiv v2) with new features, examples, and experiments.

README_zh.md

@@ -22,6 +22,8 @@
 
 ## 🚀 最新动态
 
+* [2025-08] ✨ 发布 Trinity-RFT v0.2.1 版本，强化了 Agentic RL 和 异步 RL 相关功能。
+* [2025-08] 🎵 我们推出了 [CHORD](https://github.com/modelscope/Trinity-RFT/tree/main/examples/mix_chord)，一种动态整合 SFT 和 RL 来微调 LLM 的方法（[论文](https://arxiv.org/pdf/2508.11408)）。
 * [2025-08] Trinity-RFT 现在已经支持通用多轮工作流的训练了，请参考 [ALFWorld](./docs/sphinx_doc/source/tutorial/example_step_wise.md) 和 [ReAct](./docs/sphinx_doc/source/tutorial/example_react.md) 的例子！
 * [2025-07] 发布 Trinity-RFT v0.2.0 版本，新增了多项功能优化。
 * [2025-07] 更新了[技术报告](https://arxiv.org/abs/2505.17826) (arXiv v2)，增加了新功能、示例和实验。

docs/sphinx_doc/source/main.md

@@ -8,6 +8,8 @@
 
 ## 🚀 News
 
+* [2025-08] ✨ Trinity-RFT v0.2.1 is released with enhanced features for Agentic RL and Async RL.
+* [2025-08] 🎵 We introduce [CHORD](https://github.com/modelscope/Trinity-RFT/tree/main/examples/mix_chord), a dynamic integration of SFT and RL for enhanced LLM fine-tuning ([paper](https://arxiv.org/pdf/2508.11408)).
 * [2025-08] We now support training on general multi-step workflows! Please check out examples for [ALFWorld](./docs/sphinx_doc/source/tutorial/example_step_wise.md) and [ReAct](./docs/sphinx_doc/source/tutorial/example_react.md).
 * [2025-07] Trinity-RFT v0.2.0 is released.
 * [2025-07] We update the [technical report](https://arxiv.org/abs/2505.17826) (arXiv v2) with new features, examples, and experiments.

trinity/trainer/verl/dp_actor.py

@@ -35,7 +35,6 @@
 from trinity.algorithm.kl_fn.kl_fn import DummyKLFn
 from trinity.algorithm.utils import prefix_metrics
 from trinity.common.config import AlgorithmConfig
-from trinity.utils.plugin_loader import load_plugins
 
 __all__ = ["DataParallelPPOActor"]
 
@@ -49,7 +48,6 @@ def __init__(
     ):
         """When optimizer is None, it is Reference Policy"""
         super().__init__(config, actor_module, actor_optimizer)
-        load_plugins()
         self.policy_loss_fn = None
         self.kl_loss_fn = None
         self.entropy_loss_fn = None

trinity/utils/plugin_loader.py

@@ -46,7 +46,10 @@ def load_plugin_from_dirs(plugin_dirs: Union[str, List[str]]) -> None:
                 continue
             logger.info(f"Loading plugin modules from [{file}]...")
             # load modules from file
-            load_from_file(os.path.join(plugin_dir, file))
+            try:
+                load_from_file(os.path.join(plugin_dir, file))
+            except Exception as e:
+                logger.error(f"Failed to load plugin from {file}: {e}")
 
 
 def load_from_file(file_path: str):