diff --git a/tutorials/train/semantic_segmentation/data/.gitignore b/tutorials/train/semantic_segmentation/data/.gitignore new file mode 100644 index 0000000..716e938 --- /dev/null +++ b/tutorials/train/semantic_segmentation/data/.gitignore @@ -0,0 +1,4 @@ +*.zip +*.tar.gz +rsseg/ +optic/ \ No newline at end of file diff --git a/tutorials/train/semantic_segmentation/deeplabv3p.py b/tutorials/train/semantic_segmentation/deeplabv3p.py new file mode 100644 index 0000000..965d3a8 --- /dev/null +++ b/tutorials/train/semantic_segmentation/deeplabv3p.py @@ -0,0 +1,91 @@ +#!/usr/bin/env python + +# 图像分割模型DeepLab V3+训练示例脚本 +# 执行此脚本前，请确认已正确安装PaddleRS库 + +import paddlers as pdrs +from paddlers import transforms as T + +# 下载文件存放目录 +DOWNLOAD_DIR = './data/rsseg/' +# 数据集存放目录 +DATA_DIR = './data/rsseg/remote_sensing_seg/' +# 训练集`file_list`文件路径 +TRAIN_FILE_LIST_PATH = './data/rsseg/remote_sensing_seg/train.txt' +# 验证集`file_list`文件路径 +EVAL_FILE_LIST_PATH = './data/rsseg/remote_sensing_seg/val.txt' +# 数据集类别信息文件路径 +LABEL_LIST_PATH = './data/rsseg/remote_sensing_seg/labels.txt' +# 实验目录，保存输出的模型权重和结果 +EXP_DIR = './output/deeplabv3p/' + +# 影像波段数量 +NUM_BANDS = 10 + +# 下载和解压多光谱地块分类数据集 +seg_dataset = 'https://paddleseg.bj.bcebos.com/dataset/remote_sensing_seg.zip' +pdrs.utils.download_and_decompress(seg_dataset, path=DOWNLOAD_DIR) + +# 定义训练和验证时使用的数据变换（数据增强、预处理等） +# 使用Compose组合多种变换方式。Compose中包含的变换将按顺序串行执行 +# API说明：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/docs/apis/transforms.md +train_transforms = T.Compose([ + # 将影像缩放到512x512大小 + T.Resize(target_size=512), + # 以50%的概率实施随机水平翻转 + T.RandomHorizontalFlip(prob=0.5), + # 将数据归一化到[-1,1] + T.Normalize( + mean=[0.5] * NUM_BANDS, std=[0.5] * NUM_BANDS), +]) + +eval_transforms = T.Compose([ + T.Resize(target_size=512), + # 验证阶段与训练阶段的数据归一化方式必须相同 + T.Normalize( + mean=[0.5] * NUM_BANDS, std=[0.5] * NUM_BANDS), +]) + +# 分别构建训练和验证所用的数据集 +train_dataset = pdrs.datasets.SegDataset( + data_dir=DATA_DIR, + file_list=TRAIN_FILE_LIST_PATH, + label_list=LABEL_LIST_PATH, + transforms=train_transforms, + num_workers=0, + shuffle=True) + +eval_dataset = pdrs.datasets.SegDataset( + data_dir=DATA_DIR, + file_list=EVAL_FILE_LIST_PATH, + label_list=LABEL_LIST_PATH, + transforms=eval_transforms, + num_workers=0, + shuffle=False) + +# 构建DeepLab V3+模型，使用ResNet-50作为backbone +# 目前已支持的模型请参考：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/docs/apis/model_zoo.md +# 模型输入参数请参考：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/paddlers/tasks/segmenter.py +model = pdrs.tasks.DeepLabV3P( + input_channel=NUM_BANDS, + num_classes=len(train_dataset.labels), + backbone='ResNet50_vd') + +# 执行模型训练 +model.train( + num_epochs=10, + train_dataset=train_dataset, + train_batch_size=4, + eval_dataset=eval_dataset, + save_interval_epochs=5, + # 每多少次迭代记录一次日志 + log_interval_steps=50, + save_dir=EXP_DIR, + # 初始学习率大小 + learning_rate=0.01, + # 是否使用early stopping策略，当精度不再改善时提前终止训练 + early_stop=False, + # 是否启用VisualDL日志功能 + use_vdl=True, + # 指定从某个检查点继续训练 + resume_checkpoint=None) diff --git a/tutorials/train/semantic_segmentation/deeplabv3p_resnet50_multi_channel.py b/tutorials/train/semantic_segmentation/deeplabv3p_resnet50_multi_channel.py deleted file mode 100644 index 3b6c0c2..0000000 --- a/tutorials/train/semantic_segmentation/deeplabv3p_resnet50_multi_channel.py +++ /dev/null @@ -1,54 +0,0 @@ -import os -os.environ['CUDA_VISIBLE_DEVICES'] = '0' - -import paddlers as pdrs -from paddlers import transforms as T - -# 下载和解压多光谱地块分类数据集 -dataset = 'https://paddleseg.bj.bcebos.com/dataset/remote_sensing_seg.zip' -pdrs.utils.download_and_decompress(dataset, path='./data') - -# 定义训练和验证时的transforms -channel = 10 -train_transforms = T.Compose([ - T.Resize(target_size=512), - T.RandomHorizontalFlip(), - T.Normalize( - mean=[0.5] * channel, std=[0.5] * channel), -]) - -eval_transforms = T.Compose([ - T.Resize(target_size=512), - T.Normalize( - mean=[0.5] * channel, std=[0.5] * channel), -]) - -# 定义训练和验证所用的数据集 -train_dataset = pdrs.datasets.SegDataset( - data_dir='./data/remote_sensing_seg', - file_list='./data/remote_sensing_seg/train.txt', - label_list='./data/remote_sensing_seg/labels.txt', - transforms=train_transforms, - num_workers=0, - shuffle=True) - -eval_dataset = pdrs.datasets.SegDataset( - data_dir='./data/remote_sensing_seg', - file_list='./data/remote_sensing_seg/val.txt', - label_list='./data/remote_sensing_seg/labels.txt', - transforms=eval_transforms, - num_workers=0, - shuffle=False) - -# 初始化模型，并进行训练 -# 可使用VisualDL查看训练指标 -num_classes = len(train_dataset.labels) -model = pdrs.tasks.DeepLabV3P(input_channel=channel, num_classes=num_classes, backbone='ResNet50_vd') - -model.train( - num_epochs=10, - train_dataset=train_dataset, - train_batch_size=4, - eval_dataset=eval_dataset, - learning_rate=0.01, - save_dir='output/deeplabv3p_r50vd') diff --git a/tutorials/train/semantic_segmentation/farseg_test.py b/tutorials/train/semantic_segmentation/farseg_test.py deleted file mode 100644 index 1762964..0000000 --- a/tutorials/train/semantic_segmentation/farseg_test.py +++ /dev/null @@ -1,58 +0,0 @@ -import os -os.environ['CUDA_VISIBLE_DEVICES'] = '0' - -import paddlers as pdrs -from paddlers import transforms as T - -# 下载和解压视盘分割数据集 -optic_dataset = 'https://bj.bcebos.com/paddlex/datasets/optic_disc_seg.tar.gz' -pdrs.utils.download_and_decompress(optic_dataset, path='./') - -# 定义训练和验证时的transforms -# API说明：https://github.com/PaddlePaddle/paddlers/blob/develop/docs/apis/transforms/transforms.md -train_transforms = T.Compose([ - T.Resize(target_size=512), - T.RandomHorizontalFlip(), - T.Normalize( - mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]), -]) - -eval_transforms = T.Compose([ - T.Resize(target_size=512), - T.Normalize( - mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]), -]) - -# 定义训练和验证所用的数据集 -# API说明：https://github.com/PaddlePaddle/paddlers/blob/develop/docs/apis/datasets.md -train_dataset = pdrs.datasets.SegDataset( - data_dir='optic_disc_seg', - file_list='optic_disc_seg/train_list.txt', - label_list='optic_disc_seg/labels.txt', - transforms=train_transforms, - num_workers=0, - shuffle=True) - -eval_dataset = pdrs.datasets.SegDataset( - data_dir='optic_disc_seg', - file_list='optic_disc_seg/val_list.txt', - label_list='optic_disc_seg/labels.txt', - transforms=eval_transforms, - num_workers=0, - shuffle=False) - -# 初始化模型，并进行训练 -# 可使用VisualDL查看训练指标，参考https://github.com/PaddlePaddle/paddlers/blob/develop/docs/visualdl.md -num_classes = len(train_dataset.labels) -model = pdrs.tasks.FarSeg(num_classes=num_classes) - -# API说明：https://github.com/PaddlePaddle/paddlers/blob/develop/docs/apis/models/semantic_segmentation.md -# 各参数介绍与调整说明：https://github.com/PaddlePaddle/paddlers/blob/develop/docs/parameters.md -model.train( - num_epochs=10, - train_dataset=train_dataset, - train_batch_size=4, - eval_dataset=eval_dataset, - learning_rate=0.01, - pretrain_weights=None, - save_dir='output/farseg') diff --git a/tutorials/train/semantic_segmentation/unet.py b/tutorials/train/semantic_segmentation/unet.py new file mode 100644 index 0000000..3a59784 --- /dev/null +++ b/tutorials/train/semantic_segmentation/unet.py @@ -0,0 +1,89 @@ +#!/usr/bin/env python + +# 图像分割模型UNet训练示例脚本 +# 执行此脚本前，请确认已正确安装PaddleRS库 + +import paddlers as pdrs +from paddlers import transforms as T + +# 下载文件存放目录 +DOWNLOAD_DIR = './data/rsseg/' +# 数据集存放目录 +DATA_DIR = './data/rsseg/remote_sensing_seg/' +# 训练集`file_list`文件路径 +TRAIN_FILE_LIST_PATH = './data/rsseg/remote_sensing_seg/train.txt' +# 验证集`file_list`文件路径 +EVAL_FILE_LIST_PATH = './data/rsseg/remote_sensing_seg/val.txt' +# 数据集类别信息文件路径 +LABEL_LIST_PATH = './data/rsseg/remote_sensing_seg/labels.txt' +# 实验目录，保存输出的模型权重和结果 +EXP_DIR = './output/unet/' + +# 影像波段数量 +NUM_BANDS = 10 + +# 下载和解压多光谱地块分类数据集 +seg_dataset = 'https://paddleseg.bj.bcebos.com/dataset/remote_sensing_seg.zip' +pdrs.utils.download_and_decompress(seg_dataset, path=DOWNLOAD_DIR) + +# 定义训练和验证时使用的数据变换（数据增强、预处理等） +# 使用Compose组合多种变换方式。Compose中包含的变换将按顺序串行执行 +# API说明：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/docs/apis/transforms.md +train_transforms = T.Compose([ + # 将影像缩放到512x512大小 + T.Resize(target_size=512), + # 以50%的概率实施随机水平翻转 + T.RandomHorizontalFlip(prob=0.5), + # 将数据归一化到[-1,1] + T.Normalize( + mean=[0.5] * NUM_BANDS, std=[0.5] * NUM_BANDS), +]) + +eval_transforms = T.Compose([ + T.Resize(target_size=512), + # 验证阶段与训练阶段的数据归一化方式必须相同 + T.Normalize( + mean=[0.5] * NUM_BANDS, std=[0.5] * NUM_BANDS), +]) + +# 分别构建训练和验证所用的数据集 +train_dataset = pdrs.datasets.SegDataset( + data_dir=DATA_DIR, + file_list=TRAIN_FILE_LIST_PATH, + label_list=LABEL_LIST_PATH, + transforms=train_transforms, + num_workers=0, + shuffle=True) + +eval_dataset = pdrs.datasets.SegDataset( + data_dir=DATA_DIR, + file_list=EVAL_FILE_LIST_PATH, + label_list=LABEL_LIST_PATH, + transforms=eval_transforms, + num_workers=0, + shuffle=False) + +# 构建UNet模型 +# 目前已支持的模型请参考：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/docs/apis/model_zoo.md +# 模型输入参数请参考：https://github.com/PaddleCV-SIG/PaddleRS/blob/develop/paddlers/tasks/segmenter.py +model = pdrs.tasks.UNet( + input_channel=NUM_BANDS, num_classes=len(train_dataset.labels)) + +# 执行模型训练 +model.train( + num_epochs=10, + train_dataset=train_dataset, + train_batch_size=4, + eval_dataset=eval_dataset, + save_interval_epochs=5, + # 每多少次迭代记录一次日志 + log_interval_steps=50, + save_dir=EXP_DIR, + # 初始学习率大小 + learning_rate=0.01, + # 是否使用early stopping策略，当精度不再改善时提前终止训练 + early_stop=False, + # 是否启用VisualDL日志功能 + use_vdl=True, + # 指定从某个检查点继续训练 + resume_checkpoint=None) diff --git a/tutorials/train/semantic_segmentation/unet_multi_channel.py b/tutorials/train/semantic_segmentation/unet_multi_channel.py deleted file mode 100644 index 525740b..0000000 --- a/tutorials/train/semantic_segmentation/unet_multi_channel.py +++ /dev/null @@ -1,55 +0,0 @@ -import os -os.environ['CUDA_VISIBLE_DEVICES'] = '0' - -import paddlers as pdrs -from paddlers import transforms as T - -# 下载和解压多光谱地块分类数据集 -dataset = 'https://paddleseg.bj.bcebos.com/dataset/remote_sensing_seg.zip' -pdrs.utils.download_and_decompress(dataset, path='./data') - -# 定义训练和验证时的transforms -channel = 10 -train_transforms = T.Compose([ - T.Resize(target_size=512), - T.RandomHorizontalFlip(), - T.Normalize( - mean=[0.5] * channel, std=[0.5] * channel), -]) - -eval_transforms = T.Compose([ - T.Resize(target_size=512), - T.Normalize( - mean=[0.5] * channel, std=[0.5] * channel), -]) - -# 定义训练和验证所用的数据集 -train_dataset = pdrs.datasets.SegDataset( - data_dir='./data/remote_sensing_seg', - file_list='./data/remote_sensing_seg/train.txt', - label_list='./data/remote_sensing_seg/labels.txt', - transforms=train_transforms, - num_workers=0, - shuffle=True) - -eval_dataset = pdrs.datasets.SegDataset( - data_dir='./data/remote_sensing_seg', - file_list='./data/remote_sensing_seg/val.txt', - label_list='./data/remote_sensing_seg/labels.txt', - transforms=eval_transforms, - num_workers=0, - shuffle=False) - -# 初始化模型，并进行训练 -# 可使用VisualDL查看训练指标 -num_classes = len(train_dataset.labels) -model = pdrs.tasks.UNet(input_channel=channel, num_classes=num_classes) - -model.train( - num_epochs=20, - train_dataset=train_dataset, - train_batch_size=4, - eval_dataset=eval_dataset, - learning_rate=0.01, - save_dir='output/unet', - use_vdl=True)