Skip to content

Commit

Permalink
Update readme
Browse files Browse the repository at this point in the history
  • Loading branch information
Nobody-Zhang committed Nov 1, 2023
1 parent f540339 commit ac34a7e
Show file tree
Hide file tree
Showing 3 changed files with 17 additions and 12 deletions.
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# 生产队的大萝卜 初赛代码(BEST SCORE: 0.9741)
# 生产队的大萝卜 初赛刷榜代码(BEST SCORE: 0.9741)

## 简介

Expand Down
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# 生产队的大萝卜(BEST SCORE: 0.9741)
# 生产队的大萝卜 复赛刷榜代码(BEST SCORE: 0.8807)

## 简介

Expand Down Expand Up @@ -35,13 +35,13 @@
4. 输出参数格式如下,其中,category为驾驶行为类别编号,duration为算法推理耗时(ms)

```json
 {
    "result":
        {
            "category":0,
            "duration":6000
        }
 }
{
"result":
{
"category":0,
"duration":6000
}
}

```

Expand Down
11 changes: 8 additions & 3 deletions README.md
Original file line number Diff line number Diff line change
@@ -1,13 +1,18 @@
# 第十八届“挑战杯”全国大学生课外学术科技作品竞赛“揭榜挂帅”专项赛 · 华为云赛道 国二

## [赛事首页](https://competition.huaweicloud.com/information/1000041855/introduction)

项目思路:

云端刷榜算法利用 yolov5 + Openvino + Huawei Cloud ModelArts 实现,
其中采用了二分以及分治的算法优化,使得不用遍历整个视频即可得到疲劳状态的起始、终止时间点。

端侧的算法是用的 Deepstream(C/C++) + yolov5 .
其中有 OTA [在线模型更新](https://docs.nvidia.com/metropolis/deepstream/6.0.1/dev-guide/text/DS_on_the_fly_model.html)
以及 SR [智能视频记录](https://docs.nvidia.com/metropolis/deepstream/6.0.1/dev-guide/text/DS_Smart_video.html)
端侧 ( Jetson TX2 NX ) 的算法是用的 Deepstream ( C / C++ 6.0.1 ) + yolov5 . 其中有

On-the-Air Model Update [在线模型更新](https://docs.nvidia.com/metropolis/deepstream/6.0.1/dev-guide/text/DS_on_the_fly_model.html)

Smart Video Record [智能视频记录](https://docs.nvidia.com/metropolis/deepstream/6.0.1/dev-guide/text/DS_Smart_video.html)

的具体实现代码(C/C++ ,目前Python的实现NVIDIA官方说的无法实现orz)


Expand Down

0 comments on commit ac34a7e

Please sign in to comment.