README.md 3.06 KB
Newer Older
1
2
3
4
5
6
7
8
# Generative Pre-Training2(GPT2)

## 模型介绍
GPT2模型:第二代生成式预训练模型(Generative Pre-Training2)。

## 模型结构
GPT2主要使用Transformer的Decoder模块为特征提取器,并对Transformer Decoder进行了一些改动,原本的Decoder包含了两个Multi-Head Attention结构,而GPT2只保留了Mask Multi-Head Attention。

liucong's avatar
liucong committed
9
10
## Python版本推理

liucong's avatar
liucong committed
11
本次采用GPT-2模型进行诗词生成任务,模型文件下载链接:https://pan.baidu.com/s/1KWeoUuakCZ5dualK69qCcw , 提取码:4pmh ,并将GPT2_shici.onnx模型文件保存在Resource/文件夹下。下面介绍如何运行python代码示例,Python示例的详细说明见Doc目录下的Tutorial_Python.md。
liucong's avatar
liucong committed
12

liucong's avatar
liucong committed
13
### 下载镜像
14

liucong's avatar
liucong committed
15
在光源中下载MIGraphX镜像: 
16
17

```python
18
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort1.14.0_migraphx3.0.0-dtk22.10.1
19
20
```

liucong's avatar
liucong committed
21
### 设置Python环境变量
liucong's avatar
liucong committed
22

liucong's avatar
liucong committed
23
24
25
```
export PYTHONPATH=/opt/dtk/lib:$PYTHONPATH
```
liucong's avatar
liucong committed
26

liucong's avatar
liucong committed
27
### 安装依赖
liucong's avatar
liucong committed
28
29

```python
liucong's avatar
liucong committed
30
31
# 进入gpt2 migraphx工程根目录
cd <path_to_gpt2_migraphx> 
liucong's avatar
liucong committed
32
33
34
35
36
37
38
39

# 进入示例程序目录
cd ./Python/

# 安装依赖
pip install -r requirements.txt
```

liucong's avatar
liucong committed
40
### 设置动态shape模式
41
42

```python
liucong's avatar
liucong committed
43
44
45
export MIGRAPHX_DYNAMIC_SHAPE=1
```

liucong's avatar
liucong committed
46
47
48
### 运行示例

在Python目录下执行如下命令运行该示例程序:
liucong's avatar
liucong committed
49
50
51

```python
python gpt2.py
52
53
```

liucong's avatar
liucong committed
54
55
56
57
58
59
如下所示,采用交互式界面,通过输入开头诗词,GPT2模型可以生成后续的诗句。

<img src="./Doc/Images/GPT_03.png" style="zoom:80%;" align=middle>

## C++版本推理

liucong's avatar
liucong committed
60
本次采用GPT-2模型进行诗词生成任务,模型文件下载链接:https://pan.baidu.com/s/1KWeoUuakCZ5dualK69qCcw , 提取码:4pmh ,并将GPT2_shici.onnx模型文件保存在Resource/文件夹下。下面介绍如何运行C++代码示例,C++示例的详细说明见Doc目录下的Tutorial_Cpp.md。
liucong's avatar
liucong committed
61

liucong's avatar
liucong committed
62
### 下载镜像
liucong's avatar
liucong committed
63

liucong's avatar
liucong committed
64
65
66
```
docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:ort1.14.0_migraphx3.0.0-dtk22.10.1
```
67

liucong's avatar
liucong committed
68
### 修改CMakeLists.txt
69

liucong's avatar
liucong committed
70
71
如果使用ubuntu系统,需要修改CMakeLists.txt中依赖库路径:
将"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib64/"修改为"${CMAKE_CURRENT_SOURCE_DIR}/depend/lib/"
72
73


liucong's avatar
liucong committed
74
### 构建工程
75
76
77
78
79
80
81
82
83
84
85
86

```
rbuild build -d depend
```

### 设置环境变量

将依赖库依赖加入环境变量LD_LIBRARY_PATH,在~/.bashrc中添加如下语句:

**Centos**:

```
liucong's avatar
liucong committed
87
export LD_LIBRARY_PATH=<path_to_gpt2_migraphx>/depend/lib64/:$LD_LIBRARY_PATH
88
89
90
91
92
```

**Ubuntu**:

```
liucong's avatar
liucong committed
93
export LD_LIBRARY_PATH=<path_to_gpt2_migraphx>/depend/lib/:$LD_LIBRARY_PATH
94
95
96
97
98
99
```

然后执行:

```
source ~/.bashrc
100
101
```

liucong's avatar
liucong committed
102
### 设置动态shape模式
103

liucong's avatar
liucong committed
104
105
106
107
108
```
export MIGRAPHX_DYNAMIC_SHAPE=1
```

### 运行示例
109
110

```python
liucong's avatar
liucong committed
111
112
# 进入gpt2 migraphx工程根目录
cd <path_to_gpt2_migraphx> 
113

liucong's avatar
liucong committed
114
# 进入build目录
115
116
cd ./build/

liucong's avatar
liucong committed
117
118
# 执行示例程序
./GPT2
119
120
121
```

如下所示,采用交互式界面,通过输入开头诗词,GPT2模型可以推理出后续的诗句。
122

liucong's avatar
liucong committed
123
<img src="./Doc/Images/GPT_04.png" style="zoom:100%;" align=middle>
124

liucong's avatar
liucong committed
125
## 源码仓库及问题反馈
126
127
128

https://developer.hpccube.com/codes/modelzoo/gpt2_migraphx

liucong's avatar
liucong committed
129
## 参考
130
131

https://github.com/Morizeyao/GPT2-Chinese