讲师中心 微信公众号
首页
文章
后端开发 web前端 数据库 开发工具 php框架 常见问题 人工智能 Java 系统教程 电脑教程 硬件教程 手机教程 软件教程 游戏教程 自媒体 新闻
专题
后端开发 web前端 数据库 开发工具 php框架 人工智能 Java 系统教程 电脑教程 硬件教程 手机教程 软件教程 游戏教程 新闻
AI工具
AI聊天问答 Agent智能体 AI文本写作 AI绘画作图 AI设计工具 AI视频创作 AI音频制作 AI办公学习 AI编程开发 AI提示词
学习
大前端 后端开发 数据库 移动端 运维开发 计算机基础
编程手册
大前端 后端开发 数据库 移动端 运维开发 计算机基础
下载
js特效 网站源码 工具下载 类库下载 网站素材 学习资源 插件扩展 手机游戏
最近更新
当前位置：首页 > 人工智能 > 人工智能 >

正文

0

0

基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例

P粉084495128

P粉084495128

发布时间：2025-07-23 11:08:23

|

836人浏览过

|

来源于php中文网

原创

本文介绍基于PaddleSeg 2.1实现DLRSD遥感影像数据集语义分割的流程。先准备数据集，按自定义格式整理，划分训练、验证集。接着用OCRNet模型训练，迭代1000次，最佳mIoU达0.6823。测试采用增强模式，mIoU为0.6839，最后完成单张影像推理，展示结果。

☞☞☞AI 智能聊天, 问答助手, AI 智能搜索, 免费无限量使用 DeepSeek R1 模型☜☜☜

基于paddleseg的自定义遥感数据语义分割——以dlrsd为例 - php中文网

基于paddleSeg进行自定义遥感影像数据集的语义分割——以DLRSD为例

  • 基于paddleseg 2.1
  • 使用自定义数据集DLRSD,其他遥感数据集
  • 实现训练、测试、推理
  • 脚本版任务

数据格式准备

DLRSD数据集

  • 基于UCMerced_LandUse数据集进行标注
  • 标注了17种类型的遥感地物类型
  • 图像尺寸256 * 256 * 3
  • 图像数21* 100 = 2100
  • gt格式，单通道8bit彩图，推荐使用PIL读取
  • DLRSD详细数据信息

数据展示： 基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例 - php中文网 基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例 - php中文网        

PaddleSeg数据格式

paddleSeg支持CityScapes、ADE20K、Pascal VOC、自定义等数据集格式，paddleseg数据集配置链接

       

本文使用的是自定义数据格式：

  • 数据存放在PaddleSeg/dataset/dlrsd文件夹下，
  • 影像、标注分别存储到image、gt下，
  • train、val、test等部分按txt分别存储

               

文件结构：

    custom_dataset
        |
        |--images        |  |--image1.jpg        |  |--image2.jpg        |  |--...        |
        |--labels        |  |--label1.jpg        |  |--label2.png        |  |--...        |
        |--train.txt        |
        |--val.txt        |
        |--test.txt
       

txt内容格式

images/image1.jpg labels/label1.png
images/image2.jpg labels/label2.png
...
       

数据集配置文件可以参考设置如下：

train_dataset:
  type: Dataset
  dataset_root: custom_dataset
  train_path: custom_dataset/train.txt
  num_classes: 2
  transforms:
    - type: ResizeStepScaling
      min_scale_factor: 0.5
      max_scale_factor: 2.0
      scale_step_size: 0.25
    - type: RandomPaddingCrop
      crop_size: [512, 512]    - type: RandomHorizontalFlip
    - type: Normalize
  mode: train
   

项目准备

In [ ]
# paddleseg下载解压# 从gitee下载PaddleSeg!git clone https://gitee.com/paddlepaddle/PaddleSeg.git# 进入PaddleSeg目录%cd PaddleSeg# 切换到release/2.1分支!git checkout -b release/2.1 origin/release/2.1# 安装依赖!pip install -r requirements.txt
       
fatal: destination path 'PaddleSeg' already exists and is not an empty directory.
/home/aistudio/PaddleSeg
fatal: Not a git repository (or any parent up to mount point /home/aistudio)
Stopping at filesystem boundary (GIT_DISCOVERY_ACROSS_FILESYSTEM not set).
Looking in indexes: https://pypi.tuna.tsinghua.edu.cn/simple
Requirement already satisfied: pre-commit in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 1)) (1.21.0)
Requirement already satisfied: yapf==0.26.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 2)) (0.26.0)
Requirement already satisfied: flake8 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 3)) (4.0.1)
Requirement already satisfied: pyyaml>=5.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 4)) (5.1.2)
Requirement already satisfied: visualdl>=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 5)) (2.2.0)
Requirement already satisfied: opencv-python in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 6)) (4.1.1.26)
Requirement already satisfied: tqdm in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 7)) (4.36.1)
Requirement already satisfied: filelock in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 8)) (3.0.12)
Requirement already satisfied: scipy in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 9)) (1.6.3)
Requirement already satisfied: prettytable in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from -r requirements.txt (line 10)) (0.7.2)
Requirement already satisfied: six in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (1.16.0)
Requirement already satisfied: toml in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (0.10.0)
Requirement already satisfied: nodeenv>=0.11.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (1.3.4)
Requirement already satisfied: virtualenv>=15.2 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (16.7.9)
Requirement already satisfied: identify>=1.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (1.4.10)
Requirement already satisfied: importlib-metadata in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (4.2.0)
Requirement already satisfied: cfgv>=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (2.0.1)
Requirement already satisfied: aspy.yaml in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from pre-commit->-r requirements.txt (line 1)) (1.3.0)
Requirement already satisfied: pycodestyle<2.9.0,>=2.8.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flake8->-r requirements.txt (line 3)) (2.8.0)
Requirement already satisfied: mccabe<0.7.0,>=0.6.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flake8->-r requirements.txt (line 3)) (0.6.1)
Requirement already satisfied: pyflakes<2.5.0,>=2.4.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flake8->-r requirements.txt (line 3)) (2.4.0)
Requirement already satisfied: flask>=1.1.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (1.1.1)
Requirement already satisfied: requests in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (2.27.1)
Requirement already satisfied: shellcheck-py in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (0.7.1.1)
Requirement already satisfied: bce-python-sdk in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (0.8.53)
Requirement already satisfied: protobuf>=3.11.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (3.14.0)
Requirement already satisfied: Flask-Babel>=1.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (1.0.0)
Requirement already satisfied: numpy in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (1.20.3)
Requirement already satisfied: matplotlib in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (2.2.3)
Requirement already satisfied: Pillow>=7.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (7.1.2)
Requirement already satisfied: pandas in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from visualdl>=2.0.0->-r requirements.txt (line 5)) (1.1.5)
Requirement already satisfied: Werkzeug>=0.15 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.0.0->-r requirements.txt (line 5)) (0.16.0)
Requirement already satisfied: itsdangerous>=0.24 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.0.0->-r requirements.txt (line 5)) (1.1.0)
Requirement already satisfied: Jinja2>=2.10.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.0.0->-r requirements.txt (line 5)) (3.0.0)
Requirement already satisfied: click>=5.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from flask>=1.1.1->visualdl>=2.0.0->-r requirements.txt (line 5)) (7.0)
Requirement already satisfied: pytz in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from Flask-Babel>=1.0.0->visualdl>=2.0.0->-r requirements.txt (line 5)) (2022.1)
Requirement already satisfied: Babel>=2.3 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from Flask-Babel>=1.0.0->visualdl>=2.0.0->-r requirements.txt (line 5)) (2.9.1)
Requirement already satisfied: typing-extensions>=3.6.4 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from importlib-metadata->pre-commit->-r requirements.txt (line 1)) (4.1.1)
Requirement already satisfied: zipp>=0.5 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from importlib-metadata->pre-commit->-r requirements.txt (line 1)) (3.7.0)
Requirement already satisfied: pycryptodome>=3.8.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from bce-python-sdk->visualdl>=2.0.0->-r requirements.txt (line 5)) (3.9.9)
Requirement already satisfied: future>=0.6.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from bce-python-sdk->visualdl>=2.0.0->-r requirements.txt (line 5)) (0.18.0)
Requirement already satisfied: python-dateutil>=2.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.0.0->-r requirements.txt (line 5)) (2.8.2)
Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.0.0->-r requirements.txt (line 5)) (3.0.7)
Requirement already satisfied: cycler>=0.10 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.0.0->-r requirements.txt (line 5)) (0.10.0)
Requirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from matplotlib->visualdl>=2.0.0->-r requirements.txt (line 5)) (1.1.0)
Requirement already satisfied: certifi>=2017.4.17 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.0.0->-r requirements.txt (line 5)) (2021.10.8)
Requirement already satisfied: idna<4,>=2.5 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.0.0->-r requirements.txt (line 5)) (3.3)
Requirement already satisfied: charset-normalizer~=2.0.0 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.0.0->-r requirements.txt (line 5)) (2.0.12)
Requirement already satisfied: urllib3<1.27,>=1.21.1 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from requests->visualdl>=2.0.0->-r requirements.txt (line 5)) (1.26.9)
Requirement already satisfied: MarkupSafe>=2.0.0rc2 in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from Jinja2>=2.10.1->flask>=1.1.1->visualdl>=2.0.0->-r requirements.txt (line 5)) (2.0.1)
Requirement already satisfied: setuptools in /opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages (from kiwisolver>=1.0.1->matplotlib->visualdl>=2.0.0->-r requirements.txt (line 5)) (56.2.0)
       

预修改的文件：

为正常运行网络模型，本项目已对以下文件进行修改，详情查看代码

Otter.ai
Otter.ai

一个自动的会议记录和笔记工具，会议内容生成和实时转录

下载
  • PaddleSeg/configs/base/dlrsd.yml——(数据集配置文件)
  • PaddleSeg/configs/unet/unet_dlrds.yml——(模型配置文件)
  • PaddleSeg/ocrnet_dlrsd.yml——(ocrnet模型配置文件)
  • work/img_copy.py——(文件拷贝代码)
  • work/data_split.py——(数据划分代码)
  • PaddleSeg/predict.py——(增加tif读取)

数据准备

In [ ]
# 文件解压!unzip -oq /home/aistudio/data/data55005/UCMerced_LandUse.zip -d /home/aistudio/data/image
!unzip -oq /home/aistudio/data/data55005/DLRSD.zip  -d /home/aistudio/data/gt_color
   
In [ ]
# 数据预处理，将文件组织自定义数据集格式# image copy!python /home/aistudio/work/img_copy.py /home/aistudio/data/image/UCMerced_LandUse/Images /home/aistudio/PaddleSeg/dataset/dlrsd/image# gt copy!python /home/aistudio/work/img_copy.py /home/aistudio/data/gt_color/DLRSD/Images /home/aistudio/PaddleSeg/dataset/dlrsd/gt# dataset split!python /home/aistudio/work/data_split.py 0.8 0.2 0 /home/aistudio/PaddleSeg/dataset/dlrsd/image /home/aistudio/PaddleSeg/dataset/dlrsd/gt
       
copy from /home/aistudio/data/image/UCMerced_LandUse/Images to /home/aistudio/PaddleSeg/dataset/dlrsd/image:
agricultural finish
airplane finish
tenniscourt finish
chaparral finish
freeway finish
parkinglot finish
mobilehomepark finish
buildings finish
sparseresidential finish
baseballdiamond finish
runway finish
denseresidential finish
beach finish
storagetanks finish
river finish
forest finish
overpass finish
golfcourse finish
harbor finish
intersection finish
mediumresidential finish
copy from /home/aistudio/data/gt_color/DLRSD/Images to /home/aistudio/PaddleSeg/dataset/dlrsd/gt:
agricultural finish
airplane finish
tenniscourt finish
chaparral finish
freeway finish
parkinglot finish
mobilehomepark finish
buildings finish
sparseresidential finish
baseballdiamond finish
runway finish
denseresidential finish
beach finish
storagetanks finish
river finish
forest finish
overpass finish
golfcourse finish
harbor finish
intersection finish
mediumresidential finish
total 2100, split: train 1680 - 0.80 rate, val 420 - 0.20 rate, test 0 - 0.00 rate
       

模型训练

参考paddleSeg

In [8]
# 模型训练, 其他超参数可参考paddleSeg，对unet_dlrds.yml进行修改# unet_dlrds.yml中的参数会覆盖dlrsd.yml中的参数，修改参数时，可以直接在unet_dlrds.yml中添加参数# !python train.py \#        --config configs/unet/unet_dlrsd.yml \#        --do_eval \#        --use_vdl \#        --save_interval 100 \#        --save_dir output# 使用精度更高的ocrnet进行训练, 可以尝试10000以上iter进行训练，得到更高精度!python train.py \
       --batch_size 32 \
       --iters 1000 \
       --config ocrnet_dlrsd.yml \
       --do_eval \
       --use_vdl \
       --save_interval 100 \
       --save_dir output
       
2022-04-14 12:18:35 [INFO]
------------Environment Information-------------
platform: Linux-4.13.0-36-generic-x86_64-with-debian-stretch-sid
Python: 3.7.4 (default, Aug 13 2019, 20:35:49) [GCC 7.3.0]
Paddle compiled with cuda: True
NVCC: Cuda compilation tools, release 10.1, V10.1.243
cudnn: 7.6
GPUs used: 1
CUDA_VISIBLE_DEVICES: None
GPU: ['GPU 0: Tesla V100-SXM2-16GB']
GCC: gcc (Ubuntu 7.5.0-3ubuntu1~16.04) 7.5.0
PaddlePaddle: 2.1.2
OpenCV: 4.1.1
------------------------------------------------
2022-04-14 12:18:35 [INFO]
---------------Config Information---------------
batch_size: 32
iters: 1000
loss:
  coef:
  - 1
  - 0.4
  types:
  - ignore_index: 255
    type: CrossEntropyLoss
  - ignore_index: 255
    type: CrossEntropyLoss
lr_scheduler:
  end_lr: 0
  learning_rate: 0.01
  power: 0.9
  type: PolynomialDecay
model:
  backbone:
    pretrained: https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
    type: HRNet_W18
  backbone_indices:
  - 0
  num_classes: 18
  type: OCRNet
optimizer:
  momentum: 0.9
  type: sgd
  weight_decay: 4.0e-05
train_dataset:
  dataset_root: dataset/dlrsd
  mode: train
  num_classes: 18
  train_path: dataset/dlrsd/train.txt
  transforms:
  - max_scale_factor: 2.0
    min_scale_factor: 0.5
    scale_step_size: 0.25
    type: ResizeStepScaling
  - crop_size:
    - 256
    - 256
    type: RandomPaddingCrop
  - type: RandomHorizontalFlip
  - type: Normalize
  type: Dataset
val_dataset:
  dataset_root: dataset/dlrsd
  mode: val
  num_classes: 18
  transforms:
  - type: Normalize
  - target_size:
    - 256
    - 256
    type: Resize
  type: Dataset
  val_path: dataset/dlrsd/val.txt
------------------------------------------------
W0414 12:18:35.587487   602 device_context.cc:404] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.0, Runtime API Version: 10.1
W0414 12:18:35.587553   602 device_context.cc:422] device: 0, cuDNN Version: 7.6.
2022-04-14 12:18:39 [INFO]	Loading pretrained model from https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
2022-04-14 12:18:40 [INFO]	There are 1525/1525 variables loaded into HRNet.
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/framework.py:706: DeprecationWarning: `np.bool` is a deprecated alias for the builtin `bool`. To silence this warning, use `bool` by itself. Doing this will not modify any behavior and is safe. If you specifically wanted the numpy scalar type, use `np.bool_` here.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  elif dtype == np.bool:
2022-04-14 12:18:47 [INFO]	[TRAIN] epoch: 1, iter: 10/1000, loss: 3.4345, lr: 0.009919, batch_cost: 0.6647, reader_cost: 0.03684, ips: 48.1407 samples/sec | ETA 00:10:58
2022-04-14 12:18:53 [INFO]	[TRAIN] epoch: 1, iter: 20/1000, loss: 2.7537, lr: 0.009829, batch_cost: 0.6197, reader_cost: 0.00049, ips: 51.6372 samples/sec | ETA 00:10:07
2022-04-14 12:18:59 [INFO]	[TRAIN] epoch: 1, iter: 30/1000, loss: 2.4857, lr: 0.009739, batch_cost: 0.5961, reader_cost: 0.00085, ips: 53.6849 samples/sec | ETA 00:09:38
2022-04-14 12:19:05 [INFO]	[TRAIN] epoch: 1, iter: 40/1000, loss: 2.2253, lr: 0.009648, batch_cost: 0.6049, reader_cost: 0.00106, ips: 52.8983 samples/sec | ETA 00:09:40
2022-04-14 12:19:11 [INFO]	[TRAIN] epoch: 1, iter: 50/1000, loss: 1.9626, lr: 0.009558, batch_cost: 0.5984, reader_cost: 0.00085, ips: 53.4791 samples/sec | ETA 00:09:28
2022-04-14 12:19:17 [INFO]	[TRAIN] epoch: 2, iter: 60/1000, loss: 1.8616, lr: 0.009467, batch_cost: 0.6030, reader_cost: 0.01686, ips: 53.0698 samples/sec | ETA 00:09:26
2022-04-14 12:19:23 [INFO]	[TRAIN] epoch: 2, iter: 70/1000, loss: 1.7680, lr: 0.009377, batch_cost: 0.5849, reader_cost: 0.00077, ips: 54.7057 samples/sec | ETA 00:09:04
2022-04-14 12:19:28 [INFO]	[TRAIN] epoch: 2, iter: 80/1000, loss: 1.6116, lr: 0.009286, batch_cost: 0.5890, reader_cost: 0.00095, ips: 54.3256 samples/sec | ETA 00:09:01
2022-04-14 12:19:35 [INFO]	[TRAIN] epoch: 2, iter: 90/1000, loss: 1.5829, lr: 0.009195, batch_cost: 0.6182, reader_cost: 0.00047, ips: 51.7606 samples/sec | ETA 00:09:22
2022-04-14 12:19:41 [INFO]	[TRAIN] epoch: 2, iter: 100/1000, loss: 1.5083, lr: 0.009104, batch_cost: 0.5941, reader_cost: 0.00166, ips: 53.8610 samples/sec | ETA 00:08:54
2022-04-14 12:19:41 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py:239: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.int32, but right dtype is paddle.bool, the right dtype will convert to paddle.int32
  format(lhs_dtype, rhs_dtype, lhs_dtype))
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py:239: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.int64, but right dtype is paddle.bool, the right dtype will convert to paddle.int64
  format(lhs_dtype, rhs_dtype, lhs_dtype))
420/420 [==============================] - 60s 142ms/step - batch_cost: 0.1419 - reader cost: 3.9444e-
2022-04-14 12:20:40 [INFO]	[EVAL] #Images: 420 mIoU: 0.3062 Acc: 0.6152 Kappa: 0.5491 
2022-04-14 12:20:40 [INFO]	[EVAL] Class IoU: 
[0.     0.     0.2566 0.1382 0.4141 0.4006 0.3423 0.     0.9608 0.4201
 0.0166 0.6178 0.2826 0.5871 0.15   0.     0.6299 0.294 ]
2022-04-14 12:20:40 [INFO]	[EVAL] Class Acc: 
[0.     0.     0.4912 0.9435 0.4735 0.4704 0.3853 0.     1.     0.8808
 0.086  0.6741 0.5215 0.85   0.1504 0.     0.6799 0.3236]
2022-04-14 12:20:41 [INFO]	[EVAL] The model with the best validation mIoU (0.3062) was saved at iter 100.
2022-04-14 12:20:47 [INFO]	[TRAIN] epoch: 3, iter: 110/1000, loss: 1.3279, lr: 0.009013, batch_cost: 0.5924, reader_cost: 0.01635, ips: 54.0138 samples/sec | ETA 00:08:47
2022-04-14 12:20:53 [INFO]	[TRAIN] epoch: 3, iter: 120/1000, loss: 1.2831, lr: 0.008922, batch_cost: 0.5966, reader_cost: 0.00139, ips: 53.6371 samples/sec | ETA 00:08:45
2022-04-14 12:20:59 [INFO]	[TRAIN] epoch: 3, iter: 130/1000, loss: 1.1546, lr: 0.008831, batch_cost: 0.6096, reader_cost: 0.00082, ips: 52.4938 samples/sec | ETA 00:08:50
2022-04-14 12:21:05 [INFO]	[TRAIN] epoch: 3, iter: 140/1000, loss: 1.2061, lr: 0.008740, batch_cost: 0.5993, reader_cost: 0.00117, ips: 53.3943 samples/sec | ETA 00:08:35
2022-04-14 12:21:11 [INFO]	[TRAIN] epoch: 3, iter: 150/1000, loss: 1.0794, lr: 0.008648, batch_cost: 0.5980, reader_cost: 0.00177, ips: 53.5158 samples/sec | ETA 00:08:28
2022-04-14 12:21:17 [INFO]	[TRAIN] epoch: 4, iter: 160/1000, loss: 1.0784, lr: 0.008557, batch_cost: 0.5858, reader_cost: 0.01779, ips: 54.6240 samples/sec | ETA 00:08:12
2022-04-14 12:21:23 [INFO]	[TRAIN] epoch: 4, iter: 170/1000, loss: 1.0746, lr: 0.008465, batch_cost: 0.5943, reader_cost: 0.00015, ips: 53.8473 samples/sec | ETA 00:08:13
2022-04-14 12:21:29 [INFO]	[TRAIN] epoch: 4, iter: 180/1000, loss: 1.0160, lr: 0.008374, batch_cost: 0.5921, reader_cost: 0.00016, ips: 54.0444 samples/sec | ETA 00:08:05
2022-04-14 12:21:35 [INFO]	[TRAIN] epoch: 4, iter: 190/1000, loss: 1.1226, lr: 0.008282, batch_cost: 0.5955, reader_cost: 0.00195, ips: 53.7374 samples/sec | ETA 00:08:02
2022-04-14 12:21:41 [INFO]	[TRAIN] epoch: 4, iter: 200/1000, loss: 1.1066, lr: 0.008190, batch_cost: 0.6310, reader_cost: 0.00169, ips: 50.7141 samples/sec | ETA 00:08:24
2022-04-14 12:21:41 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 61s 144ms/step - batch_cost: 0.1440 - reader cost: 4.2161e-0
2022-04-14 12:22:41 [INFO]	[EVAL] #Images: 420 mIoU: 0.5294 Acc: 0.7634 Kappa: 0.7237 
2022-04-14 12:22:41 [INFO]	[EVAL] Class IoU: 
[0.     0.4053 0.4327 0.627  0.599  0.0673 0.7334 0.     0.9533 0.5389
 0.5181 0.7464 0.4466 0.8169 0.552  0.6634 0.69   0.7388]
2022-04-14 12:22:41 [INFO]	[EVAL] Class Acc: 
[0.     0.6584 0.5536 0.7176 0.7343 0.8204 0.7856 0.     0.9868 0.6825
 0.7245 0.8834 0.8107 0.9022 0.5718 0.8478 0.7959 0.9131]
2022-04-14 12:22:42 [INFO]	[EVAL] The model with the best validation mIoU (0.5294) was saved at iter 200.
2022-04-14 12:22:48 [INFO]	[TRAIN] epoch: 5, iter: 210/1000, loss: 1.0923, lr: 0.008098, batch_cost: 0.5964, reader_cost: 0.01708, ips: 53.6531 samples/sec | ETA 00:07:51
2022-04-14 12:22:54 [INFO]	[TRAIN] epoch: 5, iter: 220/1000, loss: 1.0190, lr: 0.008005, batch_cost: 0.6013, reader_cost: 0.00018, ips: 53.2211 samples/sec | ETA 00:07:48
2022-04-14 12:23:00 [INFO]	[TRAIN] epoch: 5, iter: 230/1000, loss: 1.0221, lr: 0.007913, batch_cost: 0.5972, reader_cost: 0.00066, ips: 53.5837 samples/sec | ETA 00:07:39
2022-04-14 12:23:06 [INFO]	[TRAIN] epoch: 5, iter: 240/1000, loss: 0.9766, lr: 0.007821, batch_cost: 0.6007, reader_cost: 0.00062, ips: 53.2710 samples/sec | ETA 00:07:36
2022-04-14 12:23:12 [INFO]	[TRAIN] epoch: 5, iter: 250/1000, loss: 0.8989, lr: 0.007728, batch_cost: 0.6137, reader_cost: 0.00145, ips: 52.1419 samples/sec | ETA 00:07:40
2022-04-14 12:23:18 [INFO]	[TRAIN] epoch: 5, iter: 260/1000, loss: 0.9029, lr: 0.007635, batch_cost: 0.5650, reader_cost: 0.00090, ips: 56.6328 samples/sec | ETA 00:06:58
2022-04-14 12:23:24 [INFO]	[TRAIN] epoch: 6, iter: 270/1000, loss: 0.8537, lr: 0.007543, batch_cost: 0.6184, reader_cost: 0.01685, ips: 51.7439 samples/sec | ETA 00:07:31
2022-04-14 12:23:30 [INFO]	[TRAIN] epoch: 6, iter: 280/1000, loss: 0.8544, lr: 0.007450, batch_cost: 0.5992, reader_cost: 0.00116, ips: 53.4018 samples/sec | ETA 00:07:11
2022-04-14 12:23:36 [INFO]	[TRAIN] epoch: 6, iter: 290/1000, loss: 0.8475, lr: 0.007357, batch_cost: 0.5923, reader_cost: 0.00041, ips: 54.0301 samples/sec | ETA 00:07:00
2022-04-14 12:23:42 [INFO]	[TRAIN] epoch: 6, iter: 300/1000, loss: 0.7936, lr: 0.007264, batch_cost: 0.6050, reader_cost: 0.00058, ips: 52.8951 samples/sec | ETA 00:07:03
2022-04-14 12:23:42 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 144ms/step - batch_cost: 0.1435 - reader cost: 4.1186e-0
2022-04-14 12:24:42 [INFO]	[EVAL] #Images: 420 mIoU: 0.5801 Acc: 0.7912 Kappa: 0.7568 
2022-04-14 12:24:42 [INFO]	[EVAL] Class IoU: 
[0.     0.5324 0.4741 0.626  0.6086 0.5252 0.7589 0.007  0.9623 0.5803
 0.491  0.7786 0.5816 0.9076 0.5717 0.616  0.6964 0.7239]
2022-04-14 12:24:42 [INFO]	[EVAL] Class Acc: 
[0.     0.6716 0.7091 0.8305 0.8657 0.7009 0.8022 0.9908 0.9958 0.7105
 0.5133 0.8871 0.7654 0.9173 0.6074 0.6626 0.7855 0.9062]
2022-04-14 12:24:43 [INFO]	[EVAL] The model with the best validation mIoU (0.5801) was saved at iter 300.
2022-04-14 12:24:49 [INFO]	[TRAIN] epoch: 6, iter: 310/1000, loss: 0.8334, lr: 0.007170, batch_cost: 0.6216, reader_cost: 0.00088, ips: 51.4776 samples/sec | ETA 00:07:08
2022-04-14 12:24:55 [INFO]	[TRAIN] epoch: 7, iter: 320/1000, loss: 0.8206, lr: 0.007077, batch_cost: 0.6054, reader_cost: 0.01669, ips: 52.8564 samples/sec | ETA 00:06:51
2022-04-14 12:25:02 [INFO]	[TRAIN] epoch: 7, iter: 330/1000, loss: 0.7791, lr: 0.006983, batch_cost: 0.6115, reader_cost: 0.00038, ips: 52.3275 samples/sec | ETA 00:06:49
2022-04-14 12:25:08 [INFO]	[TRAIN] epoch: 7, iter: 340/1000, loss: 0.7893, lr: 0.006889, batch_cost: 0.5994, reader_cost: 0.00075, ips: 53.3874 samples/sec | ETA 00:06:35
2022-04-14 12:25:14 [INFO]	[TRAIN] epoch: 7, iter: 350/1000, loss: 0.7701, lr: 0.006796, batch_cost: 0.5998, reader_cost: 0.00048, ips: 53.3533 samples/sec | ETA 00:06:29
2022-04-14 12:25:19 [INFO]	[TRAIN] epoch: 7, iter: 360/1000, loss: 0.7948, lr: 0.006702, batch_cost: 0.5912, reader_cost: 0.00020, ips: 54.1302 samples/sec | ETA 00:06:18
2022-04-14 12:25:25 [INFO]	[TRAIN] epoch: 8, iter: 370/1000, loss: 0.7359, lr: 0.006607, batch_cost: 0.5961, reader_cost: 0.01591, ips: 53.6839 samples/sec | ETA 00:06:15
2022-04-14 12:25:31 [INFO]	[TRAIN] epoch: 8, iter: 380/1000, loss: 0.7732, lr: 0.006513, batch_cost: 0.5958, reader_cost: 0.00133, ips: 53.7118 samples/sec | ETA 00:06:09
2022-04-14 12:25:37 [INFO]	[TRAIN] epoch: 8, iter: 390/1000, loss: 0.7243, lr: 0.006419, batch_cost: 0.6008, reader_cost: 0.00178, ips: 53.2630 samples/sec | ETA 00:06:06
2022-04-14 12:25:43 [INFO]	[TRAIN] epoch: 8, iter: 400/1000, loss: 0.7617, lr: 0.006324, batch_cost: 0.5974, reader_cost: 0.00041, ips: 53.5635 samples/sec | ETA 00:05:58
2022-04-14 12:25:43 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 142ms/step - batch_cost: 0.1421 - reader cost: 3.9464e-0
2022-04-14 12:26:43 [INFO]	[EVAL] #Images: 420 mIoU: 0.6354 Acc: 0.8046 Kappa: 0.7733 
2022-04-14 12:26:43 [INFO]	[EVAL] Class IoU: 
[0.     0.5345 0.4958 0.6375 0.6488 0.5401 0.7947 0.4412 0.9566 0.6051
 0.5929 0.7681 0.6005 0.932  0.6718 0.6875 0.7183 0.812 ]
2022-04-14 12:26:43 [INFO]	[EVAL] Class Acc: 
[0.     0.6217 0.6435 0.6767 0.799  0.7619 0.8421 0.7266 0.9952 0.7753
 0.7934 0.917  0.7802 0.9715 0.7451 0.7686 0.8394 0.8891]
2022-04-14 12:26:44 [INFO]	[EVAL] The model with the best validation mIoU (0.6354) was saved at iter 400.
2022-04-14 12:26:50 [INFO]	[TRAIN] epoch: 8, iter: 410/1000, loss: 0.7373, lr: 0.006229, batch_cost: 0.5961, reader_cost: 0.00151, ips: 53.6830 samples/sec | ETA 00:05:51
2022-04-14 12:26:56 [INFO]	[TRAIN] epoch: 9, iter: 420/1000, loss: 0.6996, lr: 0.006134, batch_cost: 0.6003, reader_cost: 0.01966, ips: 53.3080 samples/sec | ETA 00:05:48
2022-04-14 12:27:02 [INFO]	[TRAIN] epoch: 9, iter: 430/1000, loss: 0.7439, lr: 0.006039, batch_cost: 0.6062, reader_cost: 0.00026, ips: 52.7906 samples/sec | ETA 00:05:45
2022-04-14 12:27:08 [INFO]	[TRAIN] epoch: 9, iter: 440/1000, loss: 0.8381, lr: 0.005944, batch_cost: 0.5920, reader_cost: 0.00020, ips: 54.0508 samples/sec | ETA 00:05:31
2022-04-14 12:27:14 [INFO]	[TRAIN] epoch: 9, iter: 450/1000, loss: 0.7520, lr: 0.005848, batch_cost: 0.5970, reader_cost: 0.00065, ips: 53.6005 samples/sec | ETA 00:05:28
2022-04-14 12:27:20 [INFO]	[TRAIN] epoch: 9, iter: 460/1000, loss: 0.8414, lr: 0.005753, batch_cost: 0.6021, reader_cost: 0.00129, ips: 53.1469 samples/sec | ETA 00:05:25
2022-04-14 12:27:26 [INFO]	[TRAIN] epoch: 10, iter: 470/1000, loss: 0.8407, lr: 0.005657, batch_cost: 0.5878, reader_cost: 0.01533, ips: 54.4415 samples/sec | ETA 00:05:11
2022-04-14 12:27:32 [INFO]	[TRAIN] epoch: 10, iter: 480/1000, loss: 0.7766, lr: 0.005561, batch_cost: 0.5939, reader_cost: 0.00088, ips: 53.8809 samples/sec | ETA 00:05:08
2022-04-14 12:27:38 [INFO]	[TRAIN] epoch: 10, iter: 490/1000, loss: 0.6825, lr: 0.005465, batch_cost: 0.6033, reader_cost: 0.00082, ips: 53.0438 samples/sec | ETA 00:05:07
2022-04-14 12:27:44 [INFO]	[TRAIN] epoch: 10, iter: 500/1000, loss: 0.6993, lr: 0.005369, batch_cost: 0.5977, reader_cost: 0.00220, ips: 53.5385 samples/sec | ETA 00:04:58
2022-04-14 12:27:44 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 143ms/step - batch_cost: 0.1428 - reader cost: 4.3921e-
2022-04-14 12:28:44 [INFO]	[EVAL] #Images: 420 mIoU: 0.6433 Acc: 0.8158 Kappa: 0.7854 
2022-04-14 12:28:44 [INFO]	[EVAL] Class IoU: 
[0.     0.5429 0.4983 0.6943 0.6581 0.5558 0.8024 0.4723 0.9608 0.6045
 0.5682 0.7844 0.6218 0.9213 0.6765 0.666  0.7303 0.8215]
2022-04-14 12:28:44 [INFO]	[EVAL] Class Acc: 
[0.     0.5851 0.7199 0.7802 0.8013 0.7535 0.8533 0.7028 0.9982 0.7434
 0.768  0.898  0.7861 0.9472 0.7861 0.7684 0.8051 0.9003]
2022-04-14 12:28:45 [INFO]	[EVAL] The model with the best validation mIoU (0.6433) was saved at iter 500.
2022-04-14 12:28:51 [INFO]	[TRAIN] epoch: 10, iter: 510/1000, loss: 0.7007, lr: 0.005272, batch_cost: 0.6111, reader_cost: 0.00197, ips: 52.3652 samples/sec | ETA 00:04:59
2022-04-14 12:28:56 [INFO]	[TRAIN] epoch: 10, iter: 520/1000, loss: 0.6916, lr: 0.005175, batch_cost: 0.5682, reader_cost: 0.00144, ips: 56.3156 samples/sec | ETA 00:04:32
2022-04-14 12:29:03 [INFO]	[TRAIN] epoch: 11, iter: 530/1000, loss: 0.6765, lr: 0.005078, batch_cost: 0.6610, reader_cost: 0.01684, ips: 48.4102 samples/sec | ETA 00:05:10
2022-04-14 12:29:09 [INFO]	[TRAIN] epoch: 11, iter: 540/1000, loss: 0.7006, lr: 0.004981, batch_cost: 0.6011, reader_cost: 0.00060, ips: 53.2397 samples/sec | ETA 00:04:36
2022-04-14 12:29:15 [INFO]	[TRAIN] epoch: 11, iter: 550/1000, loss: 0.6693, lr: 0.004884, batch_cost: 0.5991, reader_cost: 0.00319, ips: 53.4166 samples/sec | ETA 00:04:29
2022-04-14 12:29:21 [INFO]	[TRAIN] epoch: 11, iter: 560/1000, loss: 0.6371, lr: 0.004786, batch_cost: 0.6081, reader_cost: 0.00075, ips: 52.6256 samples/sec | ETA 00:04:27
2022-04-14 12:29:27 [INFO]	[TRAIN] epoch: 11, iter: 570/1000, loss: 0.6714, lr: 0.004688, batch_cost: 0.5879, reader_cost: 0.00069, ips: 54.4316 samples/sec | ETA 00:04:12
2022-04-14 12:29:33 [INFO]	[TRAIN] epoch: 12, iter: 580/1000, loss: 0.6552, lr: 0.004590, batch_cost: 0.5979, reader_cost: 0.01576, ips: 53.5168 samples/sec | ETA 00:04:11
2022-04-14 12:29:39 [INFO]	[TRAIN] epoch: 12, iter: 590/1000, loss: 0.7032, lr: 0.004492, batch_cost: 0.5926, reader_cost: 0.00015, ips: 53.9954 samples/sec | ETA 00:04:02
2022-04-14 12:29:45 [INFO]	[TRAIN] epoch: 12, iter: 600/1000, loss: 0.6684, lr: 0.004394, batch_cost: 0.5981, reader_cost: 0.00088, ips: 53.5068 samples/sec | ETA 00:03:59
2022-04-14 12:29:45 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 142ms/step - batch_cost: 0.1415 - reader cost: 4.0659e-0
2022-04-14 12:30:44 [INFO]	[EVAL] #Images: 420 mIoU: 0.6476 Acc: 0.8181 Kappa: 0.7884 
2022-04-14 12:30:44 [INFO]	[EVAL] Class IoU: 
[0.     0.5753 0.5149 0.6874 0.6624 0.5632 0.8085 0.4147 0.9513 0.6158
 0.6408 0.7854 0.6181 0.9192 0.6675 0.6941 0.7279 0.8105]
2022-04-14 12:30:44 [INFO]	[EVAL] Class Acc: 
[0.     0.7804 0.685  0.7804 0.806  0.7368 0.8634 0.8568 0.989  0.7441
 0.7843 0.9157 0.8449 0.9835 0.7029 0.7389 0.8188 0.874 ]
2022-04-14 12:30:45 [INFO]	[EVAL] The model with the best validation mIoU (0.6476) was saved at iter 600.
2022-04-14 12:30:51 [INFO]	[TRAIN] epoch: 12, iter: 610/1000, loss: 0.6927, lr: 0.004295, batch_cost: 0.6133, reader_cost: 0.00106, ips: 52.1739 samples/sec | ETA 00:03:59
2022-04-14 12:30:57 [INFO]	[TRAIN] epoch: 12, iter: 620/1000, loss: 0.6416, lr: 0.004196, batch_cost: 0.5878, reader_cost: 0.00156, ips: 54.4372 samples/sec | ETA 00:03:43
2022-04-14 12:31:03 [INFO]	[TRAIN] epoch: 13, iter: 630/1000, loss: 0.6696, lr: 0.004097, batch_cost: 0.5937, reader_cost: 0.01762, ips: 53.9003 samples/sec | ETA 00:03:39
2022-04-14 12:31:10 [INFO]	[TRAIN] epoch: 13, iter: 640/1000, loss: 0.6738, lr: 0.003997, batch_cost: 0.6349, reader_cost: 0.00130, ips: 50.4020 samples/sec | ETA 00:03:48
2022-04-14 12:31:16 [INFO]	[TRAIN] epoch: 13, iter: 650/1000, loss: 0.6204, lr: 0.003897, batch_cost: 0.5997, reader_cost: 0.00070, ips: 53.3608 samples/sec | ETA 00:03:29
2022-04-14 12:31:22 [INFO]	[TRAIN] epoch: 13, iter: 660/1000, loss: 0.5943, lr: 0.003797, batch_cost: 0.5970, reader_cost: 0.00115, ips: 53.6006 samples/sec | ETA 00:03:22
2022-04-14 12:31:28 [INFO]	[TRAIN] epoch: 13, iter: 670/1000, loss: 0.6692, lr: 0.003697, batch_cost: 0.5999, reader_cost: 0.00095, ips: 53.3394 samples/sec | ETA 00:03:17
2022-04-14 12:31:33 [INFO]	[TRAIN] epoch: 14, iter: 680/1000, loss: 0.6299, lr: 0.003596, batch_cost: 0.5912, reader_cost: 0.01579, ips: 54.1238 samples/sec | ETA 00:03:09
2022-04-14 12:31:40 [INFO]	[TRAIN] epoch: 14, iter: 690/1000, loss: 0.6693, lr: 0.003495, batch_cost: 0.6100, reader_cost: 0.00082, ips: 52.4570 samples/sec | ETA 00:03:09
2022-04-14 12:31:46 [INFO]	[TRAIN] epoch: 14, iter: 700/1000, loss: 0.6149, lr: 0.003394, batch_cost: 0.5998, reader_cost: 0.00157, ips: 53.3538 samples/sec | ETA 00:02:59
2022-04-14 12:31:46 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 143ms/step - batch_cost: 0.1428 - reader cost: 4.3040e-0
2022-04-14 12:32:46 [INFO]	[EVAL] #Images: 420 mIoU: 0.6746 Acc: 0.8354 Kappa: 0.8078 
2022-04-14 12:32:46 [INFO]	[EVAL] Class IoU: 
[0.     0.5989 0.5554 0.7251 0.6605 0.575  0.8147 0.4845 0.96   0.631
 0.6657 0.799  0.7254 0.9374 0.6921 0.7311 0.7419 0.8457]
2022-04-14 12:32:46 [INFO]	[EVAL] Class Acc: 
[0.     0.807  0.7245 0.8534 0.7856 0.7607 0.8763 0.7907 0.9987 0.7856
 0.7755 0.8704 0.8246 0.9818 0.7645 0.7729 0.8567 0.9267]
2022-04-14 12:32:46 [INFO]	[EVAL] The model with the best validation mIoU (0.6746) was saved at iter 700.
2022-04-14 12:32:52 [INFO]	[TRAIN] epoch: 14, iter: 710/1000, loss: 0.7421, lr: 0.003292, batch_cost: 0.5968, reader_cost: 0.00117, ips: 53.6205 samples/sec | ETA 00:02:53
2022-04-14 12:32:58 [INFO]	[TRAIN] epoch: 14, iter: 720/1000, loss: 0.6546, lr: 0.003190, batch_cost: 0.5940, reader_cost: 0.00032, ips: 53.8741 samples/sec | ETA 00:02:46
2022-04-14 12:33:04 [INFO]	[TRAIN] epoch: 15, iter: 730/1000, loss: 0.5932, lr: 0.003088, batch_cost: 0.5865, reader_cost: 0.01657, ips: 54.5621 samples/sec | ETA 00:02:38
2022-04-14 12:33:10 [INFO]	[TRAIN] epoch: 15, iter: 740/1000, loss: 0.5941, lr: 0.002985, batch_cost: 0.5966, reader_cost: 0.00024, ips: 53.6345 samples/sec | ETA 00:02:35
2022-04-14 12:33:16 [INFO]	[TRAIN] epoch: 15, iter: 750/1000, loss: 0.6004, lr: 0.002882, batch_cost: 0.6279, reader_cost: 0.00230, ips: 50.9615 samples/sec | ETA 00:02:36
2022-04-14 12:33:23 [INFO]	[TRAIN] epoch: 15, iter: 760/1000, loss: 0.6031, lr: 0.002779, batch_cost: 0.6034, reader_cost: 0.00289, ips: 53.0347 samples/sec | ETA 00:02:24
2022-04-14 12:33:28 [INFO]	[TRAIN] epoch: 15, iter: 770/1000, loss: 0.6118, lr: 0.002675, batch_cost: 0.5960, reader_cost: 0.00167, ips: 53.6929 samples/sec | ETA 00:02:17
2022-04-14 12:33:34 [INFO]	[TRAIN] epoch: 15, iter: 780/1000, loss: 0.6142, lr: 0.002570, batch_cost: 0.5754, reader_cost: 0.00081, ips: 55.6109 samples/sec | ETA 00:02:06
2022-04-14 12:33:40 [INFO]	[TRAIN] epoch: 16, iter: 790/1000, loss: 0.5906, lr: 0.002465, batch_cost: 0.6159, reader_cost: 0.01718, ips: 51.9571 samples/sec | ETA 00:02:09
2022-04-14 12:33:46 [INFO]	[TRAIN] epoch: 16, iter: 800/1000, loss: 0.5972, lr: 0.002360, batch_cost: 0.6070, reader_cost: 0.00100, ips: 52.7145 samples/sec | ETA 00:02:01
2022-04-14 12:33:46 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 143ms/step - batch_cost: 0.1431 - reader cost: 4.3473e-0
2022-04-14 12:34:47 [INFO]	[EVAL] #Images: 420 mIoU: 0.6764 Acc: 0.8370 Kappa: 0.8102 
2022-04-14 12:34:47 [INFO]	[EVAL] Class IoU: 
[0.     0.6189 0.56   0.7151 0.6693 0.5738 0.7879 0.5127 0.9634 0.6378
 0.669  0.8069 0.7086 0.9449 0.6956 0.7226 0.7423 0.8469]
2022-04-14 12:34:47 [INFO]	[EVAL] Class Acc: 
[0.     0.7452 0.7343 0.7984 0.8166 0.767  0.8257 0.7523 1.     0.7964
 0.7517 0.8972 0.8457 0.9782 0.7789 0.7646 0.8452 0.9181]
2022-04-14 12:34:48 [INFO]	[EVAL] The model with the best validation mIoU (0.6764) was saved at iter 800.
2022-04-14 12:34:54 [INFO]	[TRAIN] epoch: 16, iter: 810/1000, loss: 0.5991, lr: 0.002254, batch_cost: 0.5941, reader_cost: 0.00210, ips: 53.8640 samples/sec | ETA 00:01:52
2022-04-14 12:34:59 [INFO]	[TRAIN] epoch: 16, iter: 820/1000, loss: 0.5848, lr: 0.002147, batch_cost: 0.5925, reader_cost: 0.00075, ips: 54.0113 samples/sec | ETA 00:01:46
2022-04-14 12:35:05 [INFO]	[TRAIN] epoch: 16, iter: 830/1000, loss: 0.5802, lr: 0.002040, batch_cost: 0.5780, reader_cost: 0.00066, ips: 55.3610 samples/sec | ETA 00:01:38
2022-04-14 12:35:11 [INFO]	[TRAIN] epoch: 17, iter: 840/1000, loss: 0.6143, lr: 0.001933, batch_cost: 0.6128, reader_cost: 0.01615, ips: 52.2204 samples/sec | ETA 00:01:38
2022-04-14 12:35:17 [INFO]	[TRAIN] epoch: 17, iter: 850/1000, loss: 0.6197, lr: 0.001824, batch_cost: 0.6147, reader_cost: 0.00096, ips: 52.0584 samples/sec | ETA 00:01:32
2022-04-14 12:35:24 [INFO]	[TRAIN] epoch: 17, iter: 860/1000, loss: 0.5561, lr: 0.001715, batch_cost: 0.6057, reader_cost: 0.00072, ips: 52.8288 samples/sec | ETA 00:01:24
2022-04-14 12:35:30 [INFO]	[TRAIN] epoch: 17, iter: 870/1000, loss: 0.5752, lr: 0.001605, batch_cost: 0.5980, reader_cost: 0.00125, ips: 53.5096 samples/sec | ETA 00:01:17
2022-04-14 12:35:35 [INFO]	[TRAIN] epoch: 17, iter: 880/1000, loss: 0.5779, lr: 0.001495, batch_cost: 0.5963, reader_cost: 0.00278, ips: 53.6679 samples/sec | ETA 00:01:11
2022-04-14 12:35:41 [INFO]	[TRAIN] epoch: 18, iter: 890/1000, loss: 0.5226, lr: 0.001383, batch_cost: 0.5988, reader_cost: 0.01577, ips: 53.4440 samples/sec | ETA 00:01:05
2022-04-14 12:35:47 [INFO]	[TRAIN] epoch: 18, iter: 900/1000, loss: 0.5943, lr: 0.001270, batch_cost: 0.5943, reader_cost: 0.00091, ips: 53.8452 samples/sec | ETA 00:00:59
2022-04-14 12:35:47 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 143ms/step - batch_cost: 0.1424 - reader cost: 4.8283e-
2022-04-14 12:36:47 [INFO]	[EVAL] #Images: 420 mIoU: 0.6823 Acc: 0.8382 Kappa: 0.8114 
2022-04-14 12:36:47 [INFO]	[EVAL] Class IoU: 
[0.     0.6235 0.5526 0.7267 0.6715 0.5952 0.8155 0.516  0.964  0.6368
 0.6684 0.8126 0.7114 0.9239 0.699  0.7702 0.7441 0.8504]
2022-04-14 12:36:47 [INFO]	[EVAL] Class Acc: 
[0.     0.7182 0.683  0.8378 0.8101 0.7393 0.8766 0.763  0.9997 0.8103
 0.7642 0.8939 0.8346 0.9864 0.7791 0.8363 0.8568 0.9297]
2022-04-14 12:36:48 [INFO]	[EVAL] The model with the best validation mIoU (0.6823) was saved at iter 900.
2022-04-14 12:36:54 [INFO]	[TRAIN] epoch: 18, iter: 910/1000, loss: 0.5703, lr: 0.001156, batch_cost: 0.6115, reader_cost: 0.00312, ips: 52.3318 samples/sec | ETA 00:00:55
2022-04-14 12:37:00 [INFO]	[TRAIN] epoch: 18, iter: 920/1000, loss: 0.6093, lr: 0.001041, batch_cost: 0.6014, reader_cost: 0.00108, ips: 53.2099 samples/sec | ETA 00:00:48
2022-04-14 12:37:06 [INFO]	[TRAIN] epoch: 18, iter: 930/1000, loss: 0.5604, lr: 0.000925, batch_cost: 0.5959, reader_cost: 0.00160, ips: 53.7019 samples/sec | ETA 00:00:41
2022-04-14 12:37:12 [INFO]	[TRAIN] epoch: 19, iter: 940/1000, loss: 0.5809, lr: 0.000807, batch_cost: 0.5947, reader_cost: 0.01608, ips: 53.8086 samples/sec | ETA 00:00:35
2022-04-14 12:37:18 [INFO]	[TRAIN] epoch: 19, iter: 950/1000, loss: 0.5609, lr: 0.000687, batch_cost: 0.5905, reader_cost: 0.00034, ips: 54.1959 samples/sec | ETA 00:00:29
2022-04-14 12:37:24 [INFO]	[TRAIN] epoch: 19, iter: 960/1000, loss: 0.6061, lr: 0.000564, batch_cost: 0.6175, reader_cost: 0.00174, ips: 51.8181 samples/sec | ETA 00:00:24
2022-04-14 12:37:30 [INFO]	[TRAIN] epoch: 19, iter: 970/1000, loss: 0.5418, lr: 0.000439, batch_cost: 0.6037, reader_cost: 0.00178, ips: 53.0053 samples/sec | ETA 00:00:18
2022-04-14 12:37:36 [INFO]	[TRAIN] epoch: 19, iter: 980/1000, loss: 0.5613, lr: 0.000309, batch_cost: 0.5929, reader_cost: 0.00152, ips: 53.9685 samples/sec | ETA 00:00:11
2022-04-14 12:37:42 [INFO]	[TRAIN] epoch: 20, iter: 990/1000, loss: 0.5241, lr: 0.000173, batch_cost: 0.5946, reader_cost: 0.01715, ips: 53.8205 samples/sec | ETA 00:00:05
2022-04-14 12:37:48 [INFO]	[TRAIN] epoch: 20, iter: 1000/1000, loss: 0.5681, lr: 0.000020, batch_cost: 0.5928, reader_cost: 0.00015, ips: 53.9844 samples/sec | ETA 00:00:00
2022-04-14 12:37:48 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
420/420 [==============================] - 60s 144ms/step - batch_cost: 0.1434 - reader cost: 5.1340e-
2022-04-14 12:38:48 [INFO]	[EVAL] #Images: 420 mIoU: 0.6804 Acc: 0.8398 Kappa: 0.8132 
2022-04-14 12:38:48 [INFO]	[EVAL] Class IoU: 
[0.     0.6279 0.5664 0.7303 0.6736 0.5809 0.8024 0.5217 0.963  0.6403
 0.6745 0.8123 0.6961 0.921  0.699  0.7392 0.7469 0.8508]
2022-04-14 12:38:48 [INFO]	[EVAL] Class Acc: 
[0.     0.7598 0.7189 0.842  0.8211 0.775  0.8507 0.7527 1.     0.782
 0.7552 0.8984 0.8332 0.9858 0.7765 0.804  0.8528 0.936 ]
2022-04-14 12:38:49 [INFO]	[EVAL] The model with the best validation mIoU (0.6823) was saved at iter 900.
<class 'paddle.nn.layer.conv.Conv2D'>'s flops has been counted
Customize Function has been applied to <class 'paddle.nn.layer.norm.SyncBatchNorm'>
Cannot find suitable count function for <class 'paddleseg.models.ocrnet.SpatialGatherBlock'>. Treat it as zero FLOPs.
Cannot find suitable count function for <class 'paddle.nn.layer.common.Dropout2D'>. Treat it as zero FLOPs.
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/tensor/creation.py:125: DeprecationWarning: `np.object` is a deprecated alias for the builtin `object`. To silence this warning, use `object` by itself. Doing this will not modify any behavior and is safe. 
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  if data.dtype == np.object:
Total Flops: 1757012736     Total Params: 12122446
       

模型测试

参考paddleSeg

In [9]
# 模型测试, aug_eval可以开启增强测试模式# !python val.py \#        --config configs/unet/unet_dlrsd.yml \#        --model_path output/best_model/model.pdparams \#        --aug_eval \#        --scales 0.75 1.0 1.25 \#        --flip_horizontal# 使用ocrnet进行测试!python val.py \
       --config ocrnet_dlrsd.yml \
       --model_path output/best_model/model.pdparams \
       --aug_eval \
       --scales 0.75 1.0 1.25 \
       --flip_horizontal
       
2022-04-14 12:38:53 [INFO]
---------------Config Information---------------
batch_size: 2
iters: 160000
loss:
  coef:
  - 1
  - 0.4
  types:
  - type: CrossEntropyLoss
  - type: CrossEntropyLoss
lr_scheduler:
  end_lr: 0
  learning_rate: 0.01
  power: 0.9
  type: PolynomialDecay
model:
  backbone:
    pretrained: https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
    type: HRNet_W18
  backbone_indices:
  - 0
  num_classes: 18
  type: OCRNet
optimizer:
  momentum: 0.9
  type: sgd
  weight_decay: 4.0e-05
train_dataset:
  dataset_root: dataset/dlrsd
  mode: train
  num_classes: 18
  train_path: dataset/dlrsd/train.txt
  transforms:
  - max_scale_factor: 2.0
    min_scale_factor: 0.5
    scale_step_size: 0.25
    type: ResizeStepScaling
  - crop_size:
    - 256
    - 256
    type: RandomPaddingCrop
  - type: RandomHorizontalFlip
  - type: Normalize
  type: Dataset
val_dataset:
  dataset_root: dataset/dlrsd
  mode: val
  num_classes: 18
  transforms:
  - type: Normalize
  - target_size:
    - 256
    - 256
    type: Resize
  type: Dataset
  val_path: dataset/dlrsd/val.txt
------------------------------------------------
W0414 12:38:53.728370  4958 device_context.cc:404] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.0, Runtime API Version: 10.1
W0414 12:38:53.728415  4958 device_context.cc:422] device: 0, cuDNN Version: 7.6.
2022-04-14 12:38:57 [INFO]	Loading pretrained model from https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
2022-04-14 12:38:58 [INFO]	There are 1525/1525 variables loaded into HRNet.
2022-04-14 12:38:58 [INFO]	Loading pretrained model from output/best_model/model.pdparams
2022-04-14 12:38:58 [INFO]	There are 1583/1583 variables loaded into OCRNet.
2022-04-14 12:38:58 [INFO]	Loaded trained params of model successfully
2022-04-14 12:38:58 [INFO]	Start evaluating (total_samples: 420, total_iters: 420)...
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py:239: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.int32, but right dtype is paddle.bool, the right dtype will convert to paddle.int32
  format(lhs_dtype, rhs_dtype, lhs_dtype))
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py:239: UserWarning: The dtype of left and right variables are not the same, left dtype is paddle.int64, but right dtype is paddle.bool, the right dtype will convert to paddle.int64
  format(lhs_dtype, rhs_dtype, lhs_dtype))
420/420 [==============================] - 335s 797ms/step - batch_cost: 0.7971 - reader cost: 8.0393e-
2022-04-14 12:44:33 [INFO]	[EVAL] #Images: 420 mIoU: 0.6839 Acc: 0.8405 Kappa: 0.8143 
2022-04-14 12:44:33 [INFO]	[EVAL] Class IoU: 
[0.     0.6325 0.563  0.7203 0.6722 0.6052 0.8156 0.5139 0.9655 0.6445
 0.6614 0.814  0.7313 0.9119 0.7007 0.7548 0.7473 0.8563]
2022-04-14 12:44:33 [INFO]	[EVAL] Class Acc: 
[0.     0.7304 0.7056 0.8246 0.7997 0.7365 0.872  0.8375 1.     0.8088
 0.7385 0.9016 0.829  0.9841 0.7464 0.8043 0.8618 0.9425]
       

模型推理

参考paddleSeg

In [10]
# 模型推理, image_path可以设置为单个文件或者文件夹# !python predict.py \#        --config configs/unet/unet_dlrsd.yml \#        --model_path output/best_model/model.pdparams \#        --image_path dataset/dlrsd/image/agricultural00.tif \#        --save_dir output/result# 使用ocrnet模型推理!python predict.py \
       --config ocrnet_dlrsd.yml \
       --model_path output/best_model/model.pdparams \
       --image_path dataset/dlrsd/image/agricultural00.tif \
       --save_dir output/result
       
2022-04-14 12:44:36 [INFO]
---------------Config Information---------------
batch_size: 2
iters: 160000
loss:
  coef:
  - 1
  - 0.4
  types:
  - type: CrossEntropyLoss
  - type: CrossEntropyLoss
lr_scheduler:
  end_lr: 0
  learning_rate: 0.01
  power: 0.9
  type: PolynomialDecay
model:
  backbone:
    pretrained: https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
    type: HRNet_W18
  backbone_indices:
  - 0
  num_classes: 18
  type: OCRNet
optimizer:
  momentum: 0.9
  type: sgd
  weight_decay: 4.0e-05
train_dataset:
  dataset_root: dataset/dlrsd
  mode: train
  num_classes: 18
  train_path: dataset/dlrsd/train.txt
  transforms:
  - max_scale_factor: 2.0
    min_scale_factor: 0.5
    scale_step_size: 0.25
    type: ResizeStepScaling
  - crop_size:
    - 256
    - 256
    type: RandomPaddingCrop
  - type: RandomHorizontalFlip
  - type: Normalize
  type: Dataset
val_dataset:
  dataset_root: dataset/dlrsd
  mode: val
  num_classes: 18
  transforms:
  - type: Normalize
  - target_size:
    - 256
    - 256
    type: Resize
  type: Dataset
  val_path: dataset/dlrsd/val.txt
------------------------------------------------
W0414 12:44:36.429344  6170 device_context.cc:404] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 11.0, Runtime API Version: 10.1
W0414 12:44:36.429426  6170 device_context.cc:422] device: 0, cuDNN Version: 7.6.
2022-04-14 12:44:39 [INFO]	Loading pretrained model from https://bj.bcebos.com/paddleseg/dygraph/hrnet_w18_ssld.tar.gz
2022-04-14 12:44:40 [INFO]	There are 1525/1525 variables loaded into HRNet.
2022-04-14 12:44:40 [INFO]	Number of predict images = 1
2022-04-14 12:44:40 [INFO]	Loading pretrained model from output/best_model/model.pdparams
2022-04-14 12:44:41 [INFO]	There are 1583/1583 variables loaded into OCRNet.
2022-04-14 12:44:41 [INFO]	Start to predict...
1/1 [==============================] - 0s 173ms/step
       

推理结果展示

基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例 - php中文网        

基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例 - php中文网        

基于paddleSeg的自定义遥感数据语义分割——以DLRSD为例 - php中文网        

相关文章

workbuddy部署资源要求_workbuddy资源配置说明【说明】

workbuddy部署环境配置_workbuddy部署前配置介绍【介绍】

OpenClaw单机部署_OpenClaw单机部署详解【详解】

OpenClaw卸载命令是什么 OpenClaw CLI删除工具详细教程

Linux上OpenClaw怎么安装_OpenClawLinux安装技巧【技巧】

相关标签:

linux python git ai cos yy red igs AI编程开发

本站声明：本文内容由网友自发贡献，版权归原作者所有，本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容，请联系admin@php.cn

上一篇：趣味体验馆踩坑指南 下一篇：【AI达人创造营第二期】电机异音AI诊断

作者最新文章

【AI达人特训营】基于飞桨实现系统认证风险预测-异常检测

2025-08-01 17:14

我在AI Studio学编程

2025-08-01 17:25

【复现赛 NLP】 基于 Paddle 实现 Canine 模型

2025-08-01 17:27

基于PaddlePaddle2.2的数据建模研究助手

2025-08-01 17:30

给你的神经网络降低一个八度：八度卷积飞桨版本复现

2025-08-01 17:32

【PaddlePaddle+OpenVINO】用PaddeHub的ONNX模型

2025-08-01 17:36

飞桨常规赛：点击反欺诈预测 - 12月第7名方案

2025-08-01 17:39

PaddleHub DEMO项目合集(运行版)

2025-08-01 17:40

【PaddlePaddle】基础理论教程 - 前馈神经网络概论

2025-08-01 17:43

基于 Paddle2.0 实现 HarDNet 模型

2025-08-01 17:51

热门AI工具

更多
DeepSeek
DeepSeek

幻方量化公司旗下的开源大模型平台

AI编程开发AI聊天问答
豆包大模型
豆包大模型

字节跳动自主研发的一系列大型语言模型

AI编程开发AI大模型
WorkBuddy
WorkBuddy

腾讯云推出的AI原生桌面智能体工作台

AI办公学习Agent智能体
腾讯元宝
腾讯元宝

腾讯混元平台推出的AI助手

文档处理Excel 表格
文心一言
文心一言

文心一言是百度开发的AI聊天机器人，通过对话可以生成各种形式的内容。

AI编程开发AI文本写作
讯飞写作
讯飞写作

基于讯飞星火大模型的AI写作工具，可以快速生成新闻稿件、品宣文案、工作总结、心得体会等各种文文稿

AI文本写作中文写作
即梦AI
即梦AI

一站式AI创作平台，免费AI图片和视频生成。

图片拼接图画生成
ChatGPT
ChatGPT

最最强大的AI聊天机器人程序，ChatGPT不单是聊天机器人，还能进行撰写邮件、视频脚本、文案、翻译、代码等任务。

AI编程开发AI文本写作
智谱清言 - 免费全能的AI助手
智谱清言 - 免费全能的AI助手

智谱清言 - 免费全能的AI助手

AI编程开发Agent智能体

相关专题

更多
TypeScript类型系统进阶与大型前端项目实践
TypeScript类型系统进阶与大型前端项目实践

本专题围绕 TypeScript 在大型前端项目中的应用展开，深入讲解类型系统设计与工程化开发方法。内容包括泛型与高级类型、类型推断机制、声明文件编写、模块化结构设计以及代码规范管理。通过真实项目案例分析，帮助开发者构建类型安全、结构清晰、易维护的前端工程体系，提高团队协作效率与代码质量。

26

2026.03.13

Python异步编程与Asyncio高并发应用实践
Python异步编程与Asyncio高并发应用实践

本专题围绕 Python 异步编程模型展开，深入讲解 Asyncio 框架的核心原理与应用实践。内容包括事件循环机制、协程任务调度、异步 IO 处理以及并发任务管理策略。通过构建高并发网络请求与异步数据处理案例，帮助开发者掌握 Python 在高并发场景中的高效开发方法，并提升系统资源利用率与整体运行性能。

46

2026.03.12

C# ASP.NET Core微服务架构与API网关实践
C# ASP.NET Core微服务架构与API网关实践

本专题围绕 C# 在现代后端架构中的微服务实践展开，系统讲解基于 ASP.NET Core 构建可扩展服务体系的核心方法。内容涵盖服务拆分策略、RESTful API 设计、服务间通信、API 网关统一入口管理以及服务治理机制。通过真实项目案例，帮助开发者掌握构建高可用微服务系统的关键技术，提高系统的可扩展性与维护效率。

178

2026.03.11

Go高并发任务调度与Goroutine池化实践
Go高并发任务调度与Goroutine池化实践

本专题围绕 Go 语言在高并发任务处理场景中的实践展开，系统讲解 Goroutine 调度模型、Channel 通信机制以及并发控制策略。内容包括任务队列设计、Goroutine 池化管理、资源限制控制以及并发任务的性能优化方法。通过实际案例演示，帮助开发者构建稳定高效的 Go 并发任务处理系统，提高系统在高负载环境下的处理能力与稳定性。

51

2026.03.10

Kotlin Android模块化架构与组件化开发实践
Kotlin Android模块化架构与组件化开发实践

本专题围绕 Kotlin 在 Android 应用开发中的架构实践展开，重点讲解模块化设计与组件化开发的实现思路。内容包括项目模块拆分策略、公共组件封装、依赖管理优化、路由通信机制以及大型项目的工程化管理方法。通过真实项目案例分析，帮助开发者构建结构清晰、易扩展且维护成本低的 Android 应用架构体系，提升团队协作效率与项目迭代速度。

92

2026.03.09

JavaScript浏览器渲染机制与前端性能优化实践
JavaScript浏览器渲染机制与前端性能优化实践

本专题围绕 JavaScript 在浏览器中的执行与渲染机制展开，系统讲解 DOM 构建、CSSOM 解析、重排与重绘原理，以及关键渲染路径优化方法。内容涵盖事件循环机制、异步任务调度、资源加载优化、代码拆分与懒加载等性能优化策略。通过真实前端项目案例，帮助开发者理解浏览器底层工作原理，并掌握提升网页加载速度与交互体验的实用技巧。

102

2026.03.06

Rust内存安全机制与所有权模型深度实践
Rust内存安全机制与所有权模型深度实践

本专题围绕 Rust 语言核心特性展开，深入讲解所有权机制、借用规则、生命周期管理以及智能指针等关键概念。通过系统级开发案例，分析内存安全保障原理与零成本抽象优势，并结合并发场景讲解 Send 与 Sync 特性实现机制。帮助开发者真正理解 Rust 的设计哲学，掌握在高性能与安全性并重场景中的工程实践能力。

227

2026.03.05

PHP高性能API设计与Laravel服务架构实践
PHP高性能API设计与Laravel服务架构实践

本专题围绕 PHP 在现代 Web 后端开发中的高性能实践展开，重点讲解基于 Laravel 框架构建可扩展 API 服务的核心方法。内容涵盖路由与中间件机制、服务容器与依赖注入、接口版本管理、缓存策略设计以及队列异步处理方案。同时结合高并发场景，深入分析性能瓶颈定位与优化思路，帮助开发者构建稳定、高效、易维护的 PHP 后端服务体系。

532

2026.03.04

AI安装教程大全
AI安装教程大全

2026最全AI工具安装教程专题：包含各版本AI绘图、AI视频、智能办公软件的本地化部署手册。全篇零基础友好，附带最新模型下载地址、一键安装脚本及常见报错修复方案。每日更新，收藏这一篇就够了，让AI安装不再报错！

171

2026.03.04

热门下载

更多
网站特效
/
网站源码
/
网站素材
/
前端模板

相关下载

更多
php商城系统
淘源码商城PHP淘宝查信誉
PHP房产程序[BBWPS]
PHP简约自动发卡平台个人版
ERMEB域名PHP离线网络授权系统
Difeye-敏捷的轻量级PHP框架
大泉州汽车网PHP整站程序

精品课程

更多
相关推荐
/
热门推荐
/
最新课程
PostgreSQL 教程
PostgreSQL 教程

共48课时 | 10.7万人学习

Git 教程
Git 教程

共21课时 | 4.2万人学习

Linux运维网站架构，知识点全面总结
Linux运维网站架构，知识点全面总结

共16课时 | 1.1万人学习

最新文章

更多
WorkBuddy怎么查看技能列表_WorkBuddySkills管理界面使用介绍
通义千问回答重复内容怎么解决_通义千问回答优化方法【技巧】
QClaw更新后界面变化怎么适应_QClaw新版界面使用指南【指南】
Perplexity和ChatGPT哪个更好用_AI搜索引擎对比分析【详解】
Notion AI自动整理任务清单怎么设置_操作方法是什么【操作】
WorkBuddy怎么批量重命名文件？高效整理电脑文件夹方法【指南】
WorkBuddy怎么自动抓取小红书爆款标题_WorkBuddy自媒体运营辅助教程【进阶】
Perplexity多账号切换方法_Perplexity个人与工作环境隔离教程【详解】
Workbuddy连接平台数量有限制吗_Workbuddy平台接入规则说明【介绍】
QClaw执行失败时能不能自动重试_QClaw错误重试机制与设置【操作】
关于我们 免责申明 举报中心 意见反馈 讲师合作 广告合作 最新更新
php中文网：公益在线php培训，帮助PHP学习者快速成长！
关注服务号 技术交流群
PHP中文网订阅号
每天精选资源文章推送

Copyright 2014-2026 https://www.php.cn/ All Rights Reserved | php.cn | 湘ICP备2023035733号

  • PHP学习

  • 技术支持

  • 返回顶部