update

Author: SeaOfOcean

docs/en/tutorial/run.md

@@ -15,9 +15,6 @@ Select the job type as `PyTorch` and paste the command into the `Execution Comma
 
 
 
-For RLHF, DPO, OnlineDPO, GRPO training task, you need set the advanced setting as `customPortList=30000-30050,createSvcForAllWorkers=true`.
-
-
 ## Non-PAI-DLC environment
 
 If you want to submit distributed training in a non-PAI-DLC environment,

docs/en/tutorial/tutorial_llama2.md

@@ -221,7 +221,6 @@ In our training script, the resource requirements (assuming the resources are A1
 
 For the environment variables and configurations required for distributed execution, please refer to [Distributed Execution](run.md).
 
-Note that for RLHF tasks, if you are running on PAI DLC, you need to fill in the advanced configuration `customPortList=30000-30050,createSvcForAllWorkers=true`.
 
 ### Evaluation
 

docs/zh/tutorial/run.md

@@ -13,8 +13,6 @@
 ![image.png](../../images/dlc_2.jpg)
 
 
-**对于 RLHF/DPO/OnlineDPO/GRPO 训练任务，您需要填写高级配置`customPortList=30000-30050,createSvcForAllWorkers=true`。**
-
 
 ## 其他环境分布式执行
 

docs/zh/tutorial/tutorial_llama2.md

@@ -212,7 +212,6 @@ bash scripts/train_grpo_math_llama.sh
 3. llama2-70B RLHF: 4*8 GPU
 
 分布式执行所需的环境变量和配置参考 [分布式执行](run.md)。
-**注意对于 RLHF 任务，如果在 PAI DLC 上运行，您需要填写高级配置`customPortList=30000-30050,createSvcForAllWorkers=true`。**
 
 
 ### 效果评估