README.md 1.46 KB
Newer Older
zhanggzh's avatar
zhanggzh committed
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
# TVM 
## 模型介绍
```
    使用深度学习编译器TVM对ResNet50网络模型进行推理、调优及部署
```
## 模型结构
```
    ResNet50-v2
```
## 数据集及模型文件
模型文件下载地址: 
```
    "https://github.com/onnx/models/raw/main/vision/classification/resnet/model/resnet50-v2-7.onnx"
```
    

## 推理、自动调优及部署
### 环境配置
拉取镜像:
```
zhanggzh's avatar
zhanggzh committed
21
    docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:tvm-0.11_fp32_cpp_dtk22.10_py38_centos-7.6-latest
zhanggzh's avatar
zhanggzh committed
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
```

### 执行推理及调优
下载模型文件后执行以下命令进行推理测试及调优测试:
```
    python tune_resnet50-v2.py
```
    
    
### 部署推理
下载配置好镜像之后,将/tvm-0.11-dev0/apps/howto_deploy、路径下原有代码移除,将prepare_test_libs.py、run_example.sh、tvm_runtime_pack.cc、cow.jpg、Makefile、resnet50-v2_deploy.cc等代码及图片放置到该路径下执行
以下命令:
```
    mkdir -p lib
    python prepare_test_libs.py
    sh run_example.sh
```
## 准确率数据
```
    max_num:15.6692
    max_iter:0x28cda14
    max_num_index:345
```
## TVM版本
```
    TVM-0.11
```   
    

## 源码仓库及问题反馈
   * https://developer.hpccube.com/codes/modelzoo/tvm_tune_resnet50-v2
zhanggzh's avatar
zhanggzh committed
53

zhanggzh's avatar
zhanggzh committed
54
55
56
## 参考
   * [https://tvm.apache.org/docs/how_to/tune_with_autoscheduler/tune_network_cuda.html#sphx-glr-how-to-tune-with-autoscheduler-tune-network-cuda-py]()