STDC-th/README

36 lines
1.3 KiB
Plaintext
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

STDC语义分割模型
1.通过data里面的json文件修改任务, 配置文件如下:
"dspth":"../../data/RoadLane/", #数据文件夹
"cropsize":"1280,720", #模型的宽、高
"labelJson":"./data/RoadLane_info.json",#标签的信息
"n_classes":3,#语义分割的类别叔叔
"ignore_idx":255 #忽略的类别数
2. 数据组织
├── train
│   ├── images []
│   ├── labels []
│   └── t.txt
└── val
├── images []
└── labels []
图像放在images下面标签放在labels下面。
3. 标签格式 及数据说明文件
labels里的数据都是png格式里面放的是RGB彩色标签。如道路-256,0,0表示
数据说明文件:./data/RoadLane_info.json 一般放在./data 文件下面
{
"hasInstances": false,
"category": "void",
"catid": 0,
"name": "speedRoad",
"ignoreInEval": true,
"id":1,
"color": [
128,
0,
0
],
主要是"id"和"color"要对应上,"id"是从0开始编号
4. 模型训练
python train.py --parJson ./data/RoadLane.json --respath ./checkpooints/0430pm --gpuId 0
# ./checkpooints/0430pm --为之前保存的训练路径