基于 UnetPlusPlus 网络对水体分割实战【包含数据集、完整代码、训练好的结果、权重文件等等】

model: unetPlusPlus train device: cuda optimizer: Adam 【number of dataSet】trainSet:226,ValSet:56 0%| | 0/57 [00:00<?, ?it/s]visual trainSet sava successfully !!------>runs learning decay: lr -----------train----------- [epoch:1/5] 100%|██████████| 57/57 [01:37<00:00, 1.70s/it, dice=0.622, loss=0.563] 100%|██████████| 56/56 [00:04<00:00, 12.14it/s, dice=0.871, loss=0.266] 0%| | 0/57 [00:00<?, ?it/s]learning rate:0.00010000 train loss:0.4382 train dice:0.6075 val loss:0.4423 val dice:0.7491 [epoch:2/5] 100%|██████████| 57/57 [01:35<00:00, 1.68s/it, dice=0.787, loss=0.435] 100%|██████████| 56/56 [00:04<00:00, 12.14it/s, dice=0.884, loss=0.253] 0%| | 0/57 [00:00<?, ?it/s]learning rate:0.00010000 train loss:0.3764 train dice:0.7905 val loss:0.4443 val dice:0.7367 [epoch:3/5] 100%|██████████| 57/57 [01:35<00:00, 1.68s/it, dice=0.794, loss=0.326] 100%|██████████| 56/56 [00:04<00:00, 12.11it/s, dice=0.879, loss=0.28] learning rate:0.00010000 train loss:0.3664 train dice:0.7937 val loss:0.4436 val dice:0.7645 [epoch:4/5] 100%|██████████| 57/57 [01:35<00:00, 1.68s/it, dice=0.43, loss=1.22] 100%|██████████| 56/56 [00:04<00:00, 12.16it/s, dice=0.887, loss=0.237] 0%| | 0/57 [00:00<?, ?it/s]learning rate:0.00010000 train loss:0.3726 train dice:0.8152 val loss:0.4249 val dice:0.7457 [epoch:5/5] 100%|██████████| 57/57 [01:35<00:00, 1.68s/it, dice=0.919, loss=0.266] 100%|██████████| 56/56 [00:04<00:00, 12.14it/s, dice=0.903, loss=0.236] learning rate:0.00010000 train loss:0.3587 train dice:0.7925 val loss:0.4088 val dice:0.7724 val model: 100%|██████████| 56/56 [00:02<00:00, 19.80it/s] global correct: 0.8384 precision: 0.8596 recall: 0.8212 dice: 0.8286