# 2.3 跑通第一个例子：算法迁移

算法迁移部分主要的工作是使用BM1684提供的软件接口，实现原来有caffe/pytorch等框架实现的前后处理、推理等代码。具体接口细节可以参考${BMNNSDK}/examples/SSD\_*object/cpp\_cv\_bmcv\_bmrt*中提供的具体的代码。

下面介绍一下编译方法，此例子支持PCIE模式和SOC模式。

### 2.3.1 程序编译

进入examples/SSD\_object/cpp\_cv\_bmcv\_bmrt目录，执行make命令编译程序：

```bash
# 当前所在目录
cd /workspace/examples/SSD_object/cpp_cv_bmcv_bmrt
make -f Makefile.pcie # for PCIE MODE
make -f Makefile.arm  # for SoC MODE
```

 编译完成后，会生成ssd300\_*cv\_*bmcv\_bmrt.pcie或者ssd300\_*cv\_*bmcv\_bmrt.arm，支持图片检测或者视频检测，具体参数说明如下：

{% hint style="info" %}

```
# 程序参数说明:
# ./ssd300_cv_bmcv_bmrt.xxx image <image file> <bmodel path> <test count> <device id>
# ./ssd300_cv_bmcv_bmrt.xxx video <video url>  <bmodel path> <test count> <device id>
```

{% endhint %}

### 2.3.2 PCIE模式下运行

直接在docker容器终端内执行即可：

```bash
# 使用图片测试
# ./ssd300_cv_bmcv_bmrt.xxx image <image file> <bmodel path> <test count> <device id>
./ssd300_cv_bmcv_bmrt.pcie image /workspace/res/image/vehicle_1.jpg \
  ../model/out/fp32_ssd300.bmodel 1 0
```

运行后，会在当前目录创建results文件夹保存检测结果。

```
# 终端输出
...
set device id:0
bmcpu init: skip cpu_user_defined
open usercpu.so, init user_cpu_init
[BMRT][load_bmodel:823] INFO:Loading bmodel from [../model/out/fp32_ssd300.bmodel]. Thanks for your patience...
[BMRT][load_bmodel:787] INFO:pre net num: 0, load net num: 1
Open /dev/bm-sophon0 successfully, device index = 0, jpu fd = 27, vpp fd = 27
class id:  6 upper-left: ( 652.67542,  253.06639)  object-size: ( 945.39197,  676.06012)
class id:  6 upper-left: (1376.94324,  438.03113)  object-size: ( 521.96301,  643.66248)
class id:  6 upper-left: ( 832.24683,   49.16213)  object-size: ( 345.22144,  162.27859)
class id:  7 upper-left: ( 236.51318,  802.92780)  object-size: ( 486.17816,  269.36493)
class id:  7 upper-left: (1282.00513,  209.12936)  object-size: ( 202.10974,  115.12851)
class id:  7 upper-left: ( 372.09344,  535.28412)  object-size: ( 309.26434,  229.60358)
class id:  7 upper-left: ( 502.21912,  130.76361)  object-size: ( 141.95667,  148.55127)
class id:  7 upper-left: (1552.48718,  262.08255)  object-size: ( 273.29871,  141.04590)
class id:  7 upper-left: (1101.15466,  173.13582)  object-size: ( 152.95129,   96.85536)
##############################################

############################
SUMMARY: detect
############################
[         ssd overall]  loops:    1 avg: 54431 us
[          read image]  loops:    1 avg: 9746 us
[        attach input]  loops:    1 avg: 3 us
[           detection]  loops:    1 avg: 35894 us
[     ssd pre-process]  loops:    1 avg: 1367 us
[ ssd pre-process-vpp]  loops:    1 avg: 1201 us
[ssd pre-process-linear_tranform]  loops:    1 avg: 163 us
[       ssd inference]  loops:    1 avg: 34220 us
[    ssd post-process]  loops:    1 avg: 302 us
...

# results目录内容
# log.txt  out-batch-fp32-t_0_vehicle_1.jpg
```

### 2.3.3 SOC模式下运行

将生成的ssd300\_cv\_bmcv\_bmrt.arm文件，以及bmodel文件和测试图片文件，均拷贝到SE5小盒子上/data目录下，然后再运行（需要配置好环境变量，参考[1.5.3 环境配置-SoC](/bmnnsdk2-bm1684/bmnnsdk2/setup/on-soc.md)）。

输出内容和PCIE的内容类似，此处不再详述。


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://sophgo-doc.gitbook.io/bmnnsdk2-bm1684/tutorial/alg-optimize.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.