diff --git a/README.md b/README.md
index d0f6de411..2ee6a0eef 100644
--- a/README.md
+++ b/README.md
@@ -48,30 +48,46 @@ PaddleMIX是基于飞桨的多模态大模型开发套件，聚合图像、文
 https://github.com/PaddlePaddle/PaddleMIX/assets/29787866/8d32722a-e307-46cb-a8c0-be8acd93d2c8
 
 
-## 安装
+## 安装步骤
+### 1. 克隆PaddleMIX仓库
+```
+git clone https://github.com/PaddlePaddle/PaddleMIX
+cd PaddleMIX
+```
 
-1. 环境依赖
+### 2. 创建虚拟环境
+```
+conda create -n paddlemix python=3.10 -y
+conda activate paddlemix
 ```
-pip install -r requirements.txt
+### 3. 安装PaddlePaddle
+#### 方法 1: 一键安装（推荐）
+- CUDA 11.x或12.3
+- PaddlePaddle 3.0.0b1
+```
+sh build_paddle_env.sh
 ```
 
+#### 方法 2: 手动安装
 关于PaddlePaddle安装的详细教程请查看[Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html)。
 
-> 注：ppdiffusers部分模型需要依赖 CUDA 11.2 及以上版本，如果本地机器不符合要求，建议前往 [AI Studio](https://aistudio.baidu.com/index) 进行模型训练、推理任务。
 
-> 如果希望使用**bf16**训练推理，请使用支持**bf16**的GPU，如A100。
+### 4. 安装依赖
 
-2. 手动安装
-```
-git clone https://github.com/PaddlePaddle/PaddleMIX
-cd PaddleMIX
-pip install -e .
+#### 方法 1: 一键安装（推荐）
 
-#ppdiffusers 安装
-cd ppdiffusers
-pip install -e .
+运行以下命令来自动安装所有必要的依赖:
+```
+sh build_env.sh
 ```
 
+#### 方法 2: 手动安装（请参考 build_env.sh）
+
+> 注：ppdiffusers部分模型需要依赖 CUDA 11.2 及以上版本，如果本地机器不符合要求，建议前往 [AI Studio](https://aistudio.baidu.com/index) 进行模型训练、推理任务。
+
+> 如果希望使用**bf16**训练推理，请使用支持**bf16**的GPU，如A100。
+
+
 ## 教程
 
 - [快速开始](applications/README.md/#快速开始)
diff --git a/README_EN.md b/README_EN.md
index ded883111..4e93ccded 100644
--- a/README_EN.md
+++ b/README_EN.md
@@ -47,12 +47,40 @@ https://github.com/PaddlePaddle/PaddleMIX/assets/29787866/8d32722a-e307-46cb-a8c
 
 
 
-## Installation
+## Installation 
+### 1. Clone the PaddleMIX repository
+```
+git clone https://github.com/PaddlePaddle/PaddleMIX
+cd PaddleMIX
+```
 
-1. Environment Dependencies
+### 2. Create a virtual environment
+```
+conda create -n paddlemix python=3.10 -y
+conda activate paddlemix
 ```
-pip install -r requirements.txt
+### 3. Install PaddlePaddle
+#### Method 1: One-click installation (recommended)
+- CUDA 11.x or 12.3
+- PaddlePaddle 3.0.0b1
 ```
+sh build_paddle_env.sh
+```
+
+#### Method 2: Manual installation
+For detailed instructions on installing PaddlePaddle, please refer to [Installation](https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html).
+
+
+### 4. Install dependencies
+
+#### Method 1: One-click installation (recommended)
+
+Run the following command to automatically install all necessary dependencies:
+```
+sh build_env.sh
+```
+
+#### Method 2: Manual installation (Please refer to build_env.sh)
 
 Detailed [installation]((https://www.paddlepaddle.org.cn/install/quick?docurl=/documentation/docs/zh/develop/install/pip/linux-pip.html)) tutorials for PaddlePaddle
 
@@ -60,16 +88,6 @@ Detailed [installation]((https://www.paddlepaddle.org.cn/install/quick?docurl=/d
 
 > If you wish to train and infer using **bf16**, please use a GPU that supports **bf16**, such as the A100.
 
-2. Manual Installation
-```
-git clone https://github.com/PaddlePaddle/PaddleMIX
-cd PaddleMIX
-pip install -e .
-
-#ppdiffusers 安装
-cd ppdiffusers
-pip install -e .
-```
 
 ## Tutorial
 
diff --git a/build_env.sh b/build_env.sh
new file mode 100644
index 000000000..7777e4571
--- /dev/null
+++ b/build_env.sh
@@ -0,0 +1,28 @@
+#!/bin/bash
+
+# 设置错误时退出
+set -e
+
+echo "开始安装 PaddleMIX 及其依赖..."
+
+# 安装 PaddleMIX
+echo "安装 PaddleMIX..."
+pip install -e .
+
+# 安装 ppdiffusers
+echo "安装 ppdiffusers..."
+cd ppdiffusers
+pip install -e .
+cd ..
+
+# 安装依赖包
+echo "安装依赖包..."
+pip install -r requirements.txt
+
+# 安装自定义算子
+echo "安装自定义算子..."
+cd paddlemix/external_ops
+python setup.py install
+cd ../../
+
+echo "安装完成!"
\ No newline at end of file
diff --git a/build_paddle_env.sh b/build_paddle_env.sh
new file mode 100644
index 000000000..7c3d26362
--- /dev/null
+++ b/build_paddle_env.sh
@@ -0,0 +1,34 @@
+#!/bin/bash
+
+# 设置错误时退出
+set -e
+
+echo "开始安装 paddlepaddle ..."
+# 检测 CUDA 版本并安装相应的 paddlepaddle
+if command -v nvcc &> /dev/null; then
+    cuda_version=$(nvcc --version | grep "release" | sed -n 's/.*release \([0-9]\+\.[0-9]\+\).*/\1/p')
+    echo "检测到 CUDA 版本: $cuda_version"
+    if [[ "$cuda_version" == "11.2" ]]; then
+        echo "安装 CUDA 11.2 版本的 paddlepaddle..."
+        python -m pip install paddlepaddle-gpu==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cu112/
+    elif [[ "$cuda_version" == "11.6" ]]; then
+        echo "安装 CUDA 11.6 版本的 paddlepaddle..."
+        python -m pip install paddlepaddle-gpu==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cu116/
+    elif [[ "$cuda_version" == "11.7" ]]; then
+        echo "安装 CUDA 11.7 版本的 paddlepaddle..."
+        python -m pip install paddlepaddle-gpu==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cu117/
+    elif [[ "$cuda_version" == "11.8" ]]; then
+        echo "安装 CUDA 11.8 版本的 paddlepaddle..."
+        python -m pip install paddlepaddle-gpu==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
+    elif [[ "$cuda_version" == "12.3" ]]; then
+        echo "安装 CUDA 12.3 版本的 paddlepaddle..."
+        python -m pip install paddlepaddle-gpu==3.0.0b1 -i https://www.paddlepaddle.org.cn/packages/stable/cu123/
+    else
+        echo "警告: 不支持的 CUDA 版本。请手动安装适合您系统的 paddlepaddle 版本。"
+    fi
+else
+    echo "未检测到 CUDA。安装 CPU 版本的 paddlepaddle..."
+    pip install paddlepaddle-gpu==3.0.0b1
+fi
+
+echo "安装完成!"
\ No newline at end of file
diff --git a/paddlemix/appflow/requirements.txt b/paddlemix/appflow/requirements.txt
index 94bd5c55b..40b175e0a 100644
--- a/paddlemix/appflow/requirements.txt
+++ b/paddlemix/appflow/requirements.txt
@@ -4,3 +4,4 @@ paddlenlp>=2.6.0rc0
 https://bj.bcebos.com/v1/paddlenlp/models/community/paddlemix/appflow/paddlespeech/paddlespeech-0.0.1-py3-none-any.whl
 opencc==1.1.6
 matplotlib==3.7.5
+scipy==1.12.0
diff --git a/paddlemix/datasets/__init__.py b/paddlemix/datasets/__init__.py
index 73591ab5b..f9fd3f700 100644
--- a/paddlemix/datasets/__init__.py
+++ b/paddlemix/datasets/__init__.py
@@ -31,7 +31,7 @@
     if version.startswith('3'):
         from .internvl_dataset import *
     else:
-        print(f"paddlenlp version {version} is not 3.x, skipping import.")
+        print(f"paddlenlp version {version} is not 3.x, skipping import internvl2 datasets.")
 
 except ImportError:
     print("paddlenlp is not installed.")
\ No newline at end of file
diff --git a/paddlemix/examples/qwen2_vl/README.md b/paddlemix/examples/qwen2_vl/README.md
index f67a96058..6eebbc1ed 100644
--- a/paddlemix/examples/qwen2_vl/README.md
+++ b/paddlemix/examples/qwen2_vl/README.md
@@ -10,7 +10,13 @@
 - **python >= 3.10**
 - tiktoken
 > 注：tiktoken 要求python >= 3.8
-- paddlepaddle-gpu >= 2.6.1
+- **paddlepaddle-gpu 要求版本develop**
+```
+# 安装示例
+python -m pip install paddlepaddle-gpu==0.0.0.post118 -f https://www.paddlepaddle.org.cn/whl/linux/gpu/develop.html
+```
+
+
 - paddlenlp >= 3.0.0(默认开启flash_attn，推荐源码编译安装)
 
 > 注：
diff --git a/paddlemix/models/__init__.py b/paddlemix/models/__init__.py
index 0e20d808e..43f62eb2c 100644
--- a/paddlemix/models/__init__.py
+++ b/paddlemix/models/__init__.py
@@ -26,6 +26,7 @@
 from .minigpt4.modeling import *
 from .visualglm.configuration import *
 from .visualglm.modeling import *
+from .qwen_vl import *
 
 
 import pkg_resources
@@ -34,9 +35,10 @@
 try:
     if version.startswith('3'):
         from .internvl2 import *
-        from .qwen_vl import *
+        from .qwen2_vl import *
+
     else:
-        print(f"paddlenlp version {version} is not 3.x, skipping import.")
+        print(f"paddlenlp version {version} is not 3.x, skipping import internvl2 and qwen2_vl.")
 
 except ImportError:
     print("paddlenlp is not installed.")
diff --git a/paddlemix/models/llava/__init__.py b/paddlemix/models/llava/__init__.py
index 5cab65e3a..370ab97f8 100644
--- a/paddlemix/models/llava/__init__.py
+++ b/paddlemix/models/llava/__init__.py
@@ -28,7 +28,7 @@
         from .modeling_qwen import *
         from .configuration_qwen import *
     else:
-        print(f"paddlenlp version {version} is not 3.x, skipping import.")
+        print(f"paddlenlp version {version} is not 3.x, skipping import Qwen2Model for llava-next.")
 
 except ImportError:
     print("paddlenlp is not installed.")
diff --git a/ppdiffusers/README.md b/ppdiffusers/README.md
index 1eed8af3c..11b31f2ce 100644
--- a/ppdiffusers/README.md
+++ b/ppdiffusers/README.md
@@ -88,6 +88,11 @@ git clone https://github.com/PaddlePaddle/PaddleMIX
 cd PaddleMIX/ppdiffusers
 python setup.py install
 ```
+### 设置代理
+```shell
+export HF_HUB_ENABLE_HF_TRANSFER=1   
+export HF_ENDPOINT=https://hf-mirror.com
+```
 
 ## 快速开始
 我们将以扩散模型的典型代表**Stable Diffusion**为例，带你快速了解PPDiffusers。
diff --git a/ppdiffusers/requirements.txt b/ppdiffusers/requirements.txt
index 5b1ae2257..74640baf0 100644
--- a/ppdiffusers/requirements.txt
+++ b/ppdiffusers/requirements.txt
@@ -1,4 +1,4 @@
-paddlenlp==2.7.2
+paddlenlp>=2.7.2
 safetensors>=0.3.1
 ftfy
 regex
@@ -14,4 +14,5 @@ urllib3<=2.0.0
 einops>=0.6.1
 paddlesde
 ligo-segments
-huggingface_hub==0.23
\ No newline at end of file
+huggingface_hub>=0.22.0
+hf_transfer
\ No newline at end of file