Simplify quick start doc and add demo config file

docs_roll/docs/English/QuickStart/alicloud_pipeline_quick_start_en.md → docs_roll/docs/English/QuickStart/single_node_quick_start_en.md

@@ -1,36 +1,14 @@
-# Quickstart: Singel Node based on Alibaba Cloud
+# Quickstart: Singel Node Deployment Guide
 
 ## Environment Preparation
-1. Purchase an Alibaba Cloud Server
-- For a single-machine setup, consider a GPU instance with **NVIDIA V100**.
-- **Recommendation:** When purchasing a GPU instance via the ECS console, it's advised to select the option to automatically install GPU drivers.
-2. Remote Connect to the GPU Instance and access the machine terminal
-3. Install NVIDIA Container Toolkit
+1. Purchase a machine and install GPU drivers simultaneously
+2. Connect remotely to the GPU instance and access the machine terminal
+3. Install Docker environment and NVIDIA Container Toolkit
 ```shell
-curl -s -L https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo | \
-  sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo
-
-sudo yum install -y nvidia-container-toolkit
-```
-4. Install Docker Environment：refer to https://developer.aliyun.com/mirror/docker-ce/
-```shell
-# step 1: install necessary system tools
-sudo yum install -y yum-utils
-
-# Step 2: add software repository information
-sudo yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
-
-# Step 3: install Docker
-sudo yum install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
-
-# Step 4: start Docker service
-sudo service docker start
-
-# Verify that GPUs are visible
-docker version
+curl -fsSL https://your-domain.com/install_docker_aliyun.sh | sudo bash  
+https://github.com/alibaba/ROLL/blob/main/examples/quick_start/install_docker_nvidia_container_toolkit.sh  
 ```
 
-
 ## Environment Configuration
 ```shell
 # 1. Pull Docker image
@@ -42,13 +20,13 @@ sudo docker pull <image_address>
 # torch2.5.1 + SGlang0.4.3: roll-registry.cn-hangzhou.cr.aliyuncs.com/roll/pytorch:nvcr-24.05-py3-torch251-sglang043
 # torch2.5.1 + vLLM0.7.3: roll-registry.cn-hangzhou.cr.aliyuncs.com/roll/pytorch:nvcr-24.05-py3-torch251-vllm073
 
-# 2. Start a Docker container with GPU support and keep the container running
+# 2. Start a Docker container with GPU support, expose the port, and keep the container running
 sudo docker images
 sudo docker run -dit \
   --gpus all \
-  --network=host \
+  -p 9001:22 \
   --ipc=host \
-  --shm-size=2gb \
+  --shm-size=10gb \
   <image_id> \
   /bin/bash
 
@@ -65,7 +43,7 @@ nvidia-smi
 apt update && apt install git -y
 git clone https://github.com/alibaba/ROLL.git
 
-# If Github is not accessible, download the zip file and unzip it
+# If Github is not accessible, download the zip file directly and unzip
 wget https://github.com/alibaba/ROLL/archive/refs/heads/main.zip
 unzip main.zip
 
@@ -76,20 +54,18 @@ pip install -r requirements_torch260_sglang.txt -i https://mirrors.aliyun.com/py
 
 ## Pipeline Execution
 ```shell
-# If you encounter "ModuleNotFoundError: No module named 'roll'", you need to add environment variables
-export PYTHONPATH="/workspace/ROLL-main:$PYTHONPATH"
+bash examples/quick_start/run_agentic_pipeline_frozen_lake_single_node_demo.sh  
+```
+
+Example Log Screenshots during Pipeline Execution:
+![log1](../../../static/img/log_1.png)
 
-# Method 1: Specify the YAML file path, with the script directory (examples) as the root
-python examples/start_agentic_pipeline.py --config_path qwen2.5-0.5B-agentic_ds  --config_name agent_val_frozen_lake
+![log2](../../../static/img/log_2.png)
 
-# Method 2: Execute the shell script directly
-bash examples/qwen2.5-0.5B-agentic_ds/run_agentic_pipeline_frozen_lake.sh
+![log3](../../../static/img/log_3.png)
 
-# Modify the configuration as needed
-vim examples/qwen2.5-0.5B-agentic_ds/agent_val_frozen_lake.yaml
-```
 
-Key Configuration Modifications for Single V100 GPU Memory:
+## Reference: V100 Single-GPU Memory Configuration Optimization
 ```yaml
 # Reduce the system's expected number of GPUs from 8 to your actual 1 V100
 num_gpus_per_node: 1 
@@ -135,11 +111,3 @@ val_env_manager.tags: [SimpleSokoban, FrozenLake]
 # Reduce the total number of training steps for quicker full pipeline runs, useful for rapid debugging.
 max_steps: 100
 ```
-
-Example Log Screenshots during Pipeline Execution:
-![log1](../../../static/img/log_1.png)
-
-![log2](../../../static/img/log_2.png)
-
-![log3](../../../static/img/log_3.png)
-

docs_roll/docs/简体中文/快速开始/alicloud_pipeline_quick_start_cn.md → docs_roll/docs/简体中文/快速开始/single_node_quick_start_cn.md

@@ -1,38 +1,14 @@
-# 快速上手：阿里云单机版部署指南
+# 快速上手：单机版部署指南
 
 ## 准备环境
-1. 购买阿里云服务器
-- 单机版本 可选择 GPU：NVIDIA V100
-- 建议您通过ECS控制台购买GPU实例时，同步选中安装GPU驱动
+1. 购买机器，并同步安装GPU驱动
 2. 远程连接GPU实例，进入机器终端
-3. 安装 NVIDIA容器工具包
+3. 安装 Docker环境 和 NVIDIA容器工具包
 ```shell
-curl -s -L https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo | \
-  sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo
-
-# 安装 NVIDIA Container Toolkit 软件包
-sudo yum install -y nvidia-container-toolkit
-
-```
-4. 安装 Docker 环境：参考 https://developer.aliyun.com/mirror/docker-ce/
-```shell
-# step 1: 安装必要的一些系统工具
-sudo yum install -y yum-utils
-
-# Step 2: 添加软件源信息
-sudo yum-config-manager --add-repo https://mirrors.aliyun.com/docker-ce/linux/centos/docker-ce.repo
-
-# Step 3: 安装Docker
-sudo yum install docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin
-
-# Step 4: 开启Docker服务
-sudo service docker start
-
-# 安装校验
-docker version
+curl -fsSL https://your-domain.com/install_docker_aliyun.sh | sudo bash
+https://github.com/alibaba/ROLL/blob/main/examples/quick_start/install_docker_nvidia_container_toolkit.sh
 ```
 
-
 ## 环境配置
 ```shell
 # 1. 拉取docker镜像
@@ -44,13 +20,13 @@ sudo docker pull <image_address>
 # torch2.5.1 + SGlang0.4.3: roll-registry.cn-hangzhou.cr.aliyuncs.com/roll/pytorch:nvcr-24.05-py3-torch251-sglang043
 # torch2.5.1 + vLLM0.7.3: roll-registry.cn-hangzhou.cr.aliyuncs.com/roll/pytorch:nvcr-24.05-py3-torch251-vllm073
 
-# 2. 启动一个docker容器，指定GPU支持，并始终保持容器运行
+# 2. 启动一个docker容器，指定GPU支持，暴露容器端口，并始终保持容器运行
 sudo docker images
 sudo docker run -dit \
   --gpus all \
-  --network=host \
+  -p 9001:22 \
   --ipc=host \
-  --shm-size=2gb \
+  --shm-size=10gb \
   <image_id> \
   /bin/bash
 
@@ -78,20 +54,18 @@ pip install -r requirements_torch260_sglang.txt -i https://mirrors.aliyun.com/py
 
 ## pipeline运行
 ```shell
-# 若执行报错 ModuleNotFoundError: No module named 'roll'，需要添加环境变量
-export PYTHONPATH="/workspace/ROLL-main:$PYTHONPATH"
+bash examples/quick_start/run_agentic_pipeline_frozen_lake_single_node_demo.sh
+```
+
+pipeline运行中的log截图示例：
+![log1](../../../static/img/log_1.png)
 
-# 方法一：指定yaml文件路径，需要以脚本目录即examples为根目录
-python examples/start_agentic_pipeline.py --config_path qwen2.5-0.5B-agentic_ds  --config_name agent_val_frozen_lake
+![log2](../../../static/img/log_2.png)
 
-# 方法二：直接执行sh脚本
-bash examples/qwen2.5-0.5B-agentic_ds/run_agentic_pipeline_frozen_lake.sh
+![log3](../../../static/img/log_3.png)
 
-# 根据需要修改config
-vim examples/qwen2.5-0.5B-agentic_ds/agent_val_frozen_lake.yaml
-```
 
-单卡V100显存config修改要点：
+## 参考：单卡V100显存 config修改要点
 ```yaml
 # 将系统预期的GPU数量从8块减少到你实际拥有的1块V100
 num_gpus_per_node: 1 
@@ -138,10 +112,5 @@ val_env_manager.tags: [SimpleSokoban, FrozenLake]
 max_steps: 100
 ```
 
-pipeline运行中的log截图示例：
-![log1](../../../static/img/log_1.png)
 
-![log2](../../../static/img/log_2.png)
-
-![log3](../../../static/img/log_3.png)