paddlers/docs/quick_start_en.md

简体中文 | English

Quick Start

Prerequisites

1. Install PaddlePaddle

   • Version requirements: PaddlePaddle>=2.2.0

2. Install PaddleRS

Check out releases of PaddleRS here. Download and extract the source code and run:

pip install -r requirements.txt
pip install .

The PaddleRS code will be updated as the development progresses. You can also install the develop branch to use the latest features as follows:

git clone https://github.com/PaddlePaddle/PaddleRS
cd PaddleRS
git checkout develop
pip install -r requirements.txt
pip install .

1. (Optional) Install GDAL

PaddleRS supports reading of various types of satellite data. To use the full data reading functionality of PaddleRS, you need to install GDAL as follows:

• Linux / MacOS

conda is recommended for installation:

conda install gdal

• Windows

Windows users can download GDAL wheels from this site. Please choose the wheel according to the Python version and the platform. Take GDAL‑3.3.3‑cp39‑cp39‑win_amd64.whl as an example, run the following command to install:

pip install GDAL‑3.3.3‑cp39‑cp39‑win_amd64.whl

We also provide a docker image for installation:

1. Pull from dockerhub:

   docker pull paddlepaddle/paddlers:1.0.0
   

Optionally, you can build the image from scratch. You can change the base images for different PaddlePaddle versions by setting PPTAG in Dockerfile.

git clone https://github.com/PaddlePaddle/PaddleRS
cd PaddleRS
docker build -t <imageName> .  # Default is PaddlePaddle-2.4.1-CPU
# docker build -t <imageName> . --build-arg PPTAG=2.4.1-gpu-cuda10.2-cudnn7.6-trt7.0  # Use a GPU version of PaddlePaddle
# For more tags, please refer to: https://hub.docker.com/r/paddlepaddle/paddle/tags

1. Start a container

   docker images  # View the ID of the image
   docker run -it <imageID>
   

Model Training

See here.

Model Evaluation

After the model training is completed, you can evaluate the model by executing the following snippet (take DeepLab V3+ as an example):

import paddlers as pdrs
from paddlers import transforms as T

# Load the trained model
model = pdrs.load_model('output/deeplabv3p/best_model')

# Combine data transformation operators
eval_transforms = [
    T.Resize(target_size=512),
    T.Normalize(
        mean=[0.5] * NUM_BANDS, std=[0.5] * NUM_BANDS),
    T.ReloadMask()
]

# Load the validation dataset
dataset = pdrs.datasets.SegDataset(
    data_dir='dataset',
    file_list='dataset/val/list.txt',
    label_list='dataset/labels.txt',
    transforms=eval_transforms)

# Do the evaluation
result = model.evaluate(dataset)

print(result)

Model Deployment

Model Exporting

Please refer to this document.

Deployment Using Python

Please refer to this document.