Henryjiepanli
commented
5 months ago 您好，这个应该是您标签哪里处理错了吧，标签的尺寸应该是（1536，1536），请仔细检查一下
-----原始邮件----- 发件人:yunxuan @.> 发送时间:2024-06-01 10:54:13 (星期六) 收件人: Henryjiepanli/Uncertainty-aware-Network @.> 抄送: Subscribed @.***> 主题: [Henryjiepanli/Uncertainty-aware-Network] 使用"massbuilding"数据时，遭遇参数计算错误。 (Issue #7)
您好！感谢您的贡献！
我在使用“Massachusetts Buildings Dataset_datasets”时，遇到如下错误： “”“ File "/root/autodl-tmp/BuildingExtraction/Uncertainty-aware-Network-master/tools/metric.py", line 67, in add_batch assert gt_image.shape == pre_image.shape, 'pre_image shape {}, gt_image shape {}'.format(pre_image.shape, AssertionError: pre_image shape (1536, 1536), gt_image shape (1536, 1536, 3) ”“”
这里显示预测结果与标签维度不匹配。加载的数据如下所示： “”“ train_dataset = MassBuildDataset(mosaic_ratio=0.25, transform=get_training_transform() train_loader = DataLoader(dataset=train_dataset, batch_size=12, num_workers=4, pin_memory=True, shuffle=True, drop_last=True).
for i, sample_batched in enumerate(train_loader): image, depth = sample_batched['img'], sample_batched['gt_semantic_seg'] print(image.shape) print(depth.shape) 输出： torch.Size([12, 3, 1024, 1024]) torch.Size([12, 1024, 1024, 3]) ...... ”“”
请问这种情况该如何解决？
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: @.***>
yunxuanDLRS
您好！感谢您的贡献！
我在使用“Massachusetts Buildings Dataset_datasets”时，遇到如下错误： “”“ File "/root/autodl-tmp/BuildingExtraction/Uncertainty-aware-Network-master/tools/metric.py", line 67, in add_batch assert gt_image.shape == pre_image.shape, 'pre_image shape {}, gt_image shape {}'.format(pre_image.shape, AssertionError: pre_image shape (1536, 1536), gt_image shape (1536, 1536, 3) ”“”
这里显示预测结果与标签维度不匹配。加载的数据如下所示： “”“ train_dataset = MassBuildDataset(mosaic_ratio=0.25, transform=get_training_transform() train_loader = DataLoader(dataset=train_dataset, batch_size=12, num_workers=4, pin_memory=True, shuffle=True, drop_last=True).
for i, sample_batched in enumerate(train_loader): image, depth = sample_batched['img'], sample_batched['gt_semantic_seg'] print(image.shape) print(depth.shape) 输出： torch.Size([12, 3, 1024, 1024]) torch.Size([12, 1024, 1024, 3]) ...... ”“”
请问这种情况该如何解决？