Upload 2 files

best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c8ddab5b1dc614973fcb0f76a26972caccf6668aab945f8052cdf5b993613b1
+size 47230776

out.log

@@ -0,0 +1,613 @@
+2026-04-06 11:10:10,450 - INFO - Successfully copied delasemseg.py to output/log/20260406_111010/
+2026-04-06 11:10:10,450 - INFO - base
+2026-04-06 11:10:10,652 - INFO - ================================================================================
+CONFIG PARAMETERS
+================================================================================
+Training Parameters:
+  - Batch size: 8
+  - Learning rate: 0.006
+  - Epochs: 110
+  - Warmup: 10
+  - Label smoothing: 0.2
+S3DIS Dataset Parameters:
+  - k: [24, 32, 32, 32]
+  - grid_size: [0.04, 0.08, 0.16, 0.32]
+  - max_pts: 30000
+S3DIS Warmup Parameters:
+  - k: [24, 32, 32, 32]
+  - grid_size: [0.04, 3.5, 3.5, 3.5]
+  - max_pts: 30000
+Delaunay Segmentation Parameters:
+  - ks: [24, 32, 32, 32]
+  - depths: [4, 4, 8, 8]
+  - dims: [64, 128, 256, 512]
+  - head_dim: 256
+  - num_classes: 13
+  - drop_paths: [[0.0, 0.004347826354205608, 0.008695652708411217, 0.013043479062616825, 0.3], [0.017391305416822433, 0.021739132702350616, 0.02608695812523365, 0.030434783548116684, 0.3], [0.03478261083364487, 0.03913043811917305, 0.04347826540470123, 0.04782608896493912, 0.052173912525177, 0.056521736085414886, 0.06086956337094307, 0.06521739065647125, 0.3], [0.06956521421670914, 0.07391304522752762, 0.0782608687877655, 0.08260869979858398, 0.08695652335882187, 0.09130434691905975, 0.09565217792987823, 0.10000000149011612, 0.3]]
+  - head_drops: [0.0, 0.05000000074505806, 0.10000000894069672, 0.15000000596046448]
+  - bn_momentum: 0.02
+  - act: <class 'torch.nn.modules.activation.GELU'>
+  - mlp_ratio: 2
+  - use_cp: False
+  - cor_std: [1.6, 3.8, 7.6, 15.2]
+  - all_dist: tensor([ 0.0160,  0.0255,  0.0490,  ..., 11.5967, 12.8736, 14.9626])
+  - all_dist0: tensor([ 0.0404,  0.0425,  0.0451,  ..., 14.8801, 15.2674, 31.5054])
+  - cp_bn_momentum: 0.01005050633883342
+================================================================================
+2026-04-06 11:10:10,652 - INFO - MODEL STRUCTURE:
+2026-04-06 11:10:10,652 - INFO - ================================================================================
+2026-04-06 11:10:10,652 - INFO - torchsummary not installed, using simple model print
+2026-04-06 11:10:10,652 - INFO - Raw model structure:
+2026-04-06 11:10:10,653 - INFO - DelaSemSeg(
+  (stage): Stage(
+    (nbr_bn): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+    (nbr_proj): Sequential(
+      (0): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      (1): Linear(in_features=64, out_features=128, bias=True)
+      (2): GELU(approximate='none')
+      (3): Linear(in_features=128, out_features=64, bias=False)
+    )
+    (blk): Block(
+      (lfps): ModuleList(
+        (0-3): 4 x LFP(
+          (proj): Linear(in_features=64, out_features=64, bias=False)
+          (bn): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        )
+      )
+      (mlp): Mlp(
+        (mlp): Sequential(
+          (0): Linear(in_features=64, out_features=128, bias=True)
+          (1): GELU(approximate='none')
+          (2): Linear(in_features=128, out_features=64, bias=False)
+          (3): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        )
+      )
+      (mlps): ModuleList(
+        (0-1): 2 x Mlp(
+          (mlp): Sequential(
+            (0): Linear(in_features=64, out_features=128, bias=True)
+            (1): GELU(approximate='none')
+            (2): Linear(in_features=128, out_features=64, bias=False)
+            (3): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+        )
+      )
+      (drop_paths): ModuleList(
+        (0): DropPath(drop_prob=0.000)
+        (1): DropPath(drop_prob=0.004)
+        (2): DropPath(drop_prob=0.009)
+        (3): DropPath(drop_prob=0.013)
+        (4): DropPath(drop_prob=0.300)
+      )
+      (final_ctwc): CTWC_Block(
+        (drop): DropPath(drop_prob=0.100)
+        (mlp): Sequential(
+          (0): Linear(in_features=64, out_features=128, bias=True)
+          (1): GELU(approximate='none')
+          (2): Linear(in_features=128, out_features=64, bias=False)
+          (3): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        )
+        (net): Sequential(
+          (0): Linear(in_features=2, out_features=4, bias=True)
+          (1): ReLU(inplace=True)
+          (2): Linear(in_features=4, out_features=4, bias=True)
+          (3): ReLU(inplace=True)
+          (4): Linear(in_features=4, out_features=1, bias=True)
+          (5): Sigmoid()
+        )
+      )
+      (final_drop_path): DropPath(drop_prob=0.200)
+    )
+    (drop): DropPath(drop_prob=0.000)
+    (postproj): Sequential(
+      (0): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      (1): Linear(in_features=64, out_features=256, bias=False)
+    )
+    (cor_head): Sequential(
+      (0): Linear(in_features=64, out_features=32, bias=False)
+      (1): BatchNorm1d(32, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      (2): GELU(approximate='none')
+      (3): Linear(in_features=32, out_features=3, bias=False)
+    )
+    (sub_stage): Stage(
+      (nbr_bn): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      (nbr_proj): Sequential(
+        (0): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        (1): Linear(in_features=64, out_features=128, bias=True)
+        (2): GELU(approximate='none')
+        (3): Linear(in_features=128, out_features=128, bias=False)
+      )
+      (lfp): LFP(
+        (proj): Linear(in_features=64, out_features=128, bias=False)
+        (bn): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      )
+      (skip_proj): Sequential(
+        (0): Linear(in_features=64, out_features=128, bias=False)
+        (1): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+      )
+      (blk): Block(
+        (lfps): ModuleList(
+          (0-3): 4 x LFP(
+            (proj): Linear(in_features=128, out_features=128, bias=False)
+            (bn): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+        )
+        (mlp): Mlp(
+          (mlp): Sequential(
+            (0): Linear(in_features=128, out_features=256, bias=True)
+            (1): GELU(approximate='none')
+            (2): Linear(in_features=256, out_features=128, bias=False)
+            (3): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+        )
+        (mlps): ModuleList(
+          (0-1): 2 x Mlp(
+            (mlp): Sequential(
+              (0): Linear(in_features=128, out_features=256, bias=True)
+              (1): GELU(approximate='none')
+              (2): Linear(in_features=256, out_features=128, bias=False)
+              (3): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            )
+          )
+        )
+        (drop_paths): ModuleList(
+          (0): DropPath(drop_prob=0.017)
+          (1): DropPath(drop_prob=0.022)
+          (2): DropPath(drop_prob=0.026)
+          (3): DropPath(drop_prob=0.030)
+          (4): DropPath(drop_prob=0.300)
+        )
+        (final_ctwc): CTWC_Block(
+          (drop): DropPath(drop_prob=0.100)
+          (mlp): Sequential(
+            (0): Linear(in_features=128, out_features=256, bias=True)
+            (1): GELU(approximate='none')
+            (2): Linear(in_features=256, out_features=128, bias=False)
+            (3): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+          (net): Sequential(
+            (0): Linear(in_features=2, out_features=4, bias=True)
+            (1): ReLU(inplace=True)
+            (2): Linear(in_features=4, out_features=4, bias=True)
+            (3): ReLU(inplace=True)
+            (4): Linear(in_features=4, out_features=1, bias=True)
+            (5): Sigmoid()
+          )
+        )
+        (final_drop_path): DropPath(drop_prob=0.200)
+      )
+      (drop): DropPath(drop_prob=0.050)
+      (postproj): Sequential(
+        (0): BatchNorm1d(128, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        (1): Linear(in_features=128, out_features=256, bias=False)
+      )
+      (cor_head): Sequential(
+        (0): Linear(in_features=128, out_features=32, bias=False)
+        (1): BatchNorm1d(32, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        (2): GELU(approximate='none')
+        (3): Linear(in_features=32, out_features=3, bias=False)
+      )
+      (sub_stage): Stage(
+        (nbr_bn): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        (nbr_proj): Sequential(
+          (0): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          (1): Linear(in_features=64, out_features=128, bias=True)
+          (2): GELU(approximate='none')
+          (3): Linear(in_features=128, out_features=256, bias=False)
+        )
+        (lfp): LFP(
+          (proj): Linear(in_features=128, out_features=256, bias=False)
+          (bn): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        )
+        (skip_proj): Sequential(
+          (0): Linear(in_features=128, out_features=256, bias=False)
+          (1): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+        )
+        (blk): Block(
+          (lfps): ModuleList(
+            (0-7): 8 x LFP(
+              (proj): Linear(in_features=256, out_features=256, bias=False)
+              (bn): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            )
+          )
+          (mlp): Mlp(
+            (mlp): Sequential(
+              (0): Linear(in_features=256, out_features=512, bias=True)
+              (1): GELU(approximate='none')
+              (2): Linear(in_features=512, out_features=256, bias=False)
+              (3): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            )
+          )
+          (mlps): ModuleList(
+            (0-3): 4 x Mlp(
+              (mlp): Sequential(
+                (0): Linear(in_features=256, out_features=512, bias=True)
+                (1): GELU(approximate='none')
+                (2): Linear(in_features=512, out_features=256, bias=False)
+                (3): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+              )
+            )
+          )
+          (drop_paths): ModuleList(
+            (0): DropPath(drop_prob=0.035)
+            (1): DropPath(drop_prob=0.039)
+            (2): DropPath(drop_prob=0.043)
+            (3): DropPath(drop_prob=0.048)
+            (4): DropPath(drop_prob=0.052)
+            (5): DropPath(drop_prob=0.057)
+            (6): DropPath(drop_prob=0.061)
+            (7): DropPath(drop_prob=0.065)
+            (8): DropPath(drop_prob=0.300)
+          )
+          (final_ctwc): CTWC_Block(
+            (drop): DropPath(drop_prob=0.100)
+            (mlp): Sequential(
+              (0): Linear(in_features=256, out_features=512, bias=True)
+              (1): GELU(approximate='none')
+              (2): Linear(in_features=512, out_features=256, bias=False)
+              (3): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            )
+            (net): Sequential(
+              (0): Linear(in_features=2, out_features=4, bias=True)
+              (1): ReLU(inplace=True)
+              (2): Linear(in_features=4, out_features=4, bias=True)
+              (3): ReLU(inplace=True)
+              (4): Linear(in_features=4, out_features=1, bias=True)
+              (5): Sigmoid()
+            )
+          )
+          (final_drop_path): DropPath(drop_prob=0.200)
+        )
+        (drop): DropPath(drop_prob=0.100)
+        (postproj): Sequential(
+          (0): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          (1): Linear(in_features=256, out_features=256, bias=False)
+        )
+        (cor_head): Sequential(
+          (0): Linear(in_features=256, out_features=32, bias=False)
+          (1): BatchNorm1d(32, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          (2): GELU(approximate='none')
+          (3): Linear(in_features=32, out_features=3, bias=False)
+        )
+        (sub_stage): Stage(
+          (nbr_bn): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          (nbr_proj): Sequential(
+            (0): BatchNorm1d(64, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            (1): Linear(in_features=64, out_features=128, bias=True)
+            (2): GELU(approximate='none')
+            (3): Linear(in_features=128, out_features=512, bias=False)
+          )
+          (lfp): LFP(
+            (proj): Linear(in_features=256, out_features=512, bias=False)
+            (bn): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+          (skip_proj): Sequential(
+            (0): Linear(in_features=256, out_features=512, bias=False)
+            (1): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+          )
+          (blk): Block(
+            (lfps): ModuleList(
+              (0-7): 8 x LFP(
+                (proj): Linear(in_features=512, out_features=512, bias=False)
+                (bn): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+              )
+            )
+            (mlp): Mlp(
+              (mlp): Sequential(
+                (0): Linear(in_features=512, out_features=1024, bias=True)
+                (1): GELU(approximate='none')
+                (2): Linear(in_features=1024, out_features=512, bias=False)
+                (3): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+              )
+            )
+            (mlps): ModuleList(
+              (0-3): 4 x Mlp(
+                (mlp): Sequential(
+                  (0): Linear(in_features=512, out_features=1024, bias=True)
+                  (1): GELU(approximate='none')
+                  (2): Linear(in_features=1024, out_features=512, bias=False)
+                  (3): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+                )
+              )
+            )
+            (drop_paths): ModuleList(
+              (0): DropPath(drop_prob=0.070)
+              (1): DropPath(drop_prob=0.074)
+              (2): DropPath(drop_prob=0.078)
+              (3): DropPath(drop_prob=0.083)
+              (4): DropPath(drop_prob=0.087)
+              (5): DropPath(drop_prob=0.091)
+              (6): DropPath(drop_prob=0.096)
+              (7): DropPath(drop_prob=0.100)
+              (8): DropPath(drop_prob=0.300)
+            )
+            (final_ctwc): CTWC_Block(
+              (drop): DropPath(drop_prob=0.100)
+              (mlp): Sequential(
+                (0): Linear(in_features=512, out_features=1024, bias=True)
+                (1): GELU(approximate='none')
+                (2): Linear(in_features=1024, out_features=512, bias=False)
+                (3): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+              )
+              (net): Sequential(
+                (0): Linear(in_features=2, out_features=4, bias=True)
+                (1): ReLU(inplace=True)
+                (2): Linear(in_features=4, out_features=4, bias=True)
+                (3): ReLU(inplace=True)
+                (4): Linear(in_features=4, out_features=1, bias=True)
+                (5): Sigmoid()
+              )
+            )
+            (final_drop_path): DropPath(drop_prob=0.200)
+          )
+          (drop): DropPath(drop_prob=0.150)
+          (postproj): Sequential(
+            (0): BatchNorm1d(512, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            (1): Linear(in_features=512, out_features=256, bias=False)
+          )
+          (cor_head): Sequential(
+            (0): Linear(in_features=512, out_features=32, bias=False)
+            (1): BatchNorm1d(32, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+            (2): GELU(approximate='none')
+            (3): Linear(in_features=32, out_features=3, bias=False)
+          )
+        )
+      )
+    )
+  )
+  (head): Sequential(
+    (0): BatchNorm1d(256, eps=1e-05, momentum=0.02, affine=True, track_running_stats=True)
+    (1): GELU(approximate='none')
+    (2): Linear(in_features=256, out_features=13, bias=True)
+  )
+)
+2026-04-06 11:10:10,653 - INFO - ================================================================================
+2026-04-06 11:10:10,654 - INFO - MODEL PARAMETERS:
+  - Total parameters: 11,730,521
+  - Trainable parameters: 11,730,521
+  - Non-trainable parameters: 0
+  - Model size: 44.75 MB (FP32)
+================================================================================
+2026-04-06 11:12:08,580 - INFO - epoch:0/110 || loss: 1.367 || cls: 1.2886
+2026-04-06 11:12:11,691 - INFO - duration: 1:54 (mm:ss),train miou:0.4739,val miou:0.2725 best miou: 0
+2026-04-06 11:12:11,691 - INFO - ==============================  new best! epoch:0,train:0.4739,val:0.2725  ==============================
+2026-04-06 11:14:02,388 - INFO - epoch:1/110 || loss: 1.1982 || cls: 0.8202
+2026-04-06 11:14:04,867 - INFO - duration: 1:53 (mm:ss),train miou:0.6728,val miou:0.0078 best miou: 0.2725
+2026-04-06 11:15:55,881 - INFO - epoch:2/110 || loss: 1.1538 || cls: 0.7513
+2026-04-06 11:15:58,179 - INFO - duration: 1:53 (mm:ss),train miou:0.7284,val miou:0.2405 best miou: 0.2725
+2026-04-06 11:17:48,858 - INFO - epoch:3/110 || loss: 1.13 || cls: 0.6818
+2026-04-06 11:17:51,416 - INFO - duration: 1:53 (mm:ss),train miou:0.7582,val miou:0.1973 best miou: 0.2725
+2026-04-06 11:19:42,092 - INFO - epoch:4/110 || loss: 1.1173 || cls: 0.5636
+2026-04-06 11:19:44,606 - INFO - duration: 1:53 (mm:ss),train miou:0.7734,val miou:0.2793 best miou: 0.2725
+2026-04-06 11:19:44,606 - INFO - ==============================  new best! epoch:4,train:0.7734,val:0.2793  ==============================
+2026-04-06 11:21:35,344 - INFO - epoch:5/110 || loss: 1.1028 || cls: 0.4514
+2026-04-06 11:21:37,908 - INFO - duration: 1:53 (mm:ss),train miou:0.7942,val miou:0.6452 best miou: 0.2793
+2026-04-06 11:21:37,908 - INFO - ==============================  new best! epoch:5,train:0.7942,val:0.6452  ==============================
+2026-04-06 11:23:28,667 - INFO - epoch:6/110 || loss: 1.0968 || cls: 0.4074
+2026-04-06 11:23:31,199 - INFO - duration: 1:53 (mm:ss),train miou:0.7986,val miou:0.6002 best miou: 0.6452
+2026-04-06 11:25:21,953 - INFO - epoch:7/110 || loss: 1.0916 || cls: 0.3757
+2026-04-06 11:25:24,561 - INFO - duration: 1:53 (mm:ss),train miou:0.8054,val miou:0.5855 best miou: 0.6452
+2026-04-06 11:27:15,506 - INFO - epoch:8/110 || loss: 1.0912 || cls: 0.3647
+2026-04-06 11:27:17,758 - INFO - duration: 1:53 (mm:ss),train miou:0.8085,val miou:0.0091 best miou: 0.6452
+2026-04-06 11:29:08,651 - INFO - epoch:9/110 || loss: 1.0845 || cls: 0.3468
+2026-04-06 11:29:10,810 - INFO - duration: 1:52 (mm:ss),train miou:0.8148,val miou:0.3857 best miou: 0.6452
+2026-04-06 11:31:01,802 - INFO - epoch:10/110 || loss: 1.0783 || cls: 0.3385
+2026-04-06 11:31:03,970 - INFO - duration: 1:53 (mm:ss),train miou:0.8236,val miou:0.2654 best miou: 0.6452
+2026-04-06 11:32:54,917 - INFO - epoch:11/110 || loss: 1.0725 || cls: 0.3264
+2026-04-06 11:32:57,372 - INFO - duration: 1:53 (mm:ss),train miou:0.8314,val miou:0.143 best miou: 0.6452
+2026-04-06 11:34:48,193 - INFO - epoch:12/110 || loss: 1.0721 || cls: 0.3232
+2026-04-06 11:34:50,800 - INFO - duration: 1:53 (mm:ss),train miou:0.8315,val miou:0.1575 best miou: 0.6452
+2026-04-06 11:36:41,744 - INFO - epoch:13/110 || loss: 1.0676 || cls: 0.3161
+2026-04-06 11:36:44,526 - INFO - duration: 1:53 (mm:ss),train miou:0.8361,val miou:0.5703 best miou: 0.6452
+2026-04-06 11:38:35,972 - INFO - epoch:14/110 || loss: 1.0618 || cls: 0.3156
+2026-04-06 11:38:38,419 - INFO - duration: 1:53 (mm:ss),train miou:0.8431,val miou:0.6907 best miou: 0.6452
+2026-04-06 11:38:38,419 - INFO - ==============================  new best! epoch:14,train:0.8431,val:0.6907  ==============================
+2026-04-06 11:40:28,477 - INFO - epoch:15/110 || loss: 1.0572 || cls: 0.3096
+2026-04-06 11:40:31,048 - INFO - duration: 1:52 (mm:ss),train miou:0.8494,val miou:0.6781 best miou: 0.6907
+2026-04-06 11:42:21,985 - INFO - epoch:16/110 || loss: 1.0556 || cls: 0.3076
+2026-04-06 11:42:24,396 - INFO - duration: 1:53 (mm:ss),train miou:0.8522,val miou:0.6544 best miou: 0.6907
+2026-04-06 11:44:15,198 - INFO - epoch:17/110 || loss: 1.0523 || cls: 0.3061
+2026-04-06 11:44:17,891 - INFO - duration: 1:53 (mm:ss),train miou:0.8564,val miou:0.6781 best miou: 0.6907
+2026-04-06 11:46:09,150 - INFO - epoch:18/110 || loss: 1.0528 || cls: 0.3115
+2026-04-06 11:46:11,693 - INFO - duration: 1:53 (mm:ss),train miou:0.8543,val miou:0.69 best miou: 0.6907
+2026-04-06 11:48:02,717 - INFO - epoch:19/110 || loss: 1.0516 || cls: 0.3118
+2026-04-06 11:48:05,509 - INFO - duration: 1:53 (mm:ss),train miou:0.8584,val miou:0.6825 best miou: 0.6907
+2026-04-06 11:49:56,760 - INFO - epoch:20/110 || loss: 1.0471 || cls: 0.3137
+2026-04-06 11:49:59,086 - INFO - duration: 1:53 (mm:ss),train miou:0.8613,val miou:0.6854 best miou: 0.6907
+2026-04-06 11:51:50,678 - INFO - epoch:21/110 || loss: 1.0477 || cls: 0.3175
+2026-04-06 11:51:53,279 - INFO - duration: 1:54 (mm:ss),train miou:0.8614,val miou:0.6756 best miou: 0.6907
+2026-04-06 11:53:44,053 - INFO - epoch:22/110 || loss: 1.0421 || cls: 0.3109
+2026-04-06 11:53:46,791 - INFO - duration: 1:53 (mm:ss),train miou:0.8687,val miou:0.7078 best miou: 0.6907
+2026-04-06 11:53:46,791 - INFO - ==============================  new best! epoch:22,train:0.8687,val:0.7078  ==============================
+2026-04-06 11:55:37,444 - INFO - epoch:23/110 || loss: 1.0429 || cls: 0.317
+2026-04-06 11:55:40,058 - INFO - duration: 1:52 (mm:ss),train miou:0.8665,val miou:0.7015 best miou: 0.7078
+2026-04-06 11:57:30,704 - INFO - epoch:24/110 || loss: 1.0427 || cls: 0.3243
+2026-04-06 11:57:33,148 - INFO - duration: 1:52 (mm:ss),train miou:0.869,val miou:0.6963 best miou: 0.7078
+2026-04-06 11:59:23,215 - INFO - epoch:25/110 || loss: 1.0377 || cls: 0.319
+2026-04-06 11:59:25,567 - INFO - duration: 1:52 (mm:ss),train miou:0.8736,val miou:0.698 best miou: 0.7078
+2026-04-06 12:01:15,720 - INFO - epoch:26/110 || loss: 1.0406 || cls: 0.3312
+2026-04-06 12:01:18,305 - INFO - duration: 1:52 (mm:ss),train miou:0.87,val miou:0.7057 best miou: 0.7078
+2026-04-06 12:03:08,378 - INFO - epoch:27/110 || loss: 1.0338 || cls: 0.328
+2026-04-06 12:03:10,653 - INFO - duration: 1:52 (mm:ss),train miou:0.8795,val miou:0.689 best miou: 0.7078
+2026-04-06 12:05:00,509 - INFO - epoch:28/110 || loss: 1.0363 || cls: 0.3371
+2026-04-06 12:05:02,663 - INFO - duration: 1:51 (mm:ss),train miou:0.8743,val miou:0.6927 best miou: 0.7078
+2026-04-06 12:06:52,796 - INFO - epoch:29/110 || loss: 1.0326 || cls: 0.336
+2026-04-06 12:06:55,135 - INFO - duration: 1:52 (mm:ss),train miou:0.8818,val miou:0.724 best miou: 0.7078
+2026-04-06 12:06:55,136 - INFO - ==============================  new best! epoch:29,train:0.8818,val:0.724  ==============================
+2026-04-06 12:08:45,601 - INFO - epoch:30/110 || loss: 1.0327 || cls: 0.342
+2026-04-06 12:08:48,244 - INFO - duration: 1:52 (mm:ss),train miou:0.8806,val miou:0.7014 best miou: 0.724
+2026-04-06 12:10:39,187 - INFO - epoch:31/110 || loss: 1.031 || cls: 0.3484
+2026-04-06 12:10:41,566 - INFO - duration: 1:53 (mm:ss),train miou:0.8824,val miou:0.7031 best miou: 0.724
+2026-04-06 12:12:31,662 - INFO - epoch:32/110 || loss: 1.0275 || cls: 0.3426
+2026-04-06 12:12:34,211 - INFO - duration: 1:52 (mm:ss),train miou:0.8862,val miou:0.6985 best miou: 0.724
+2026-04-06 12:14:24,520 - INFO - epoch:33/110 || loss: 1.031 || cls: 0.3581
+2026-04-06 12:14:27,321 - INFO - duration: 1:52 (mm:ss),train miou:0.8826,val miou:0.7174 best miou: 0.724
+2026-04-06 12:16:17,513 - INFO - epoch:34/110 || loss: 1.0286 || cls: 0.3614
+2026-04-06 12:16:19,929 - INFO - duration: 1:52 (mm:ss),train miou:0.8852,val miou:0.6857 best miou: 0.724
+2026-04-06 12:18:10,228 - INFO - epoch:35/110 || loss: 1.0258 || cls: 0.3625
+2026-04-06 12:18:12,877 - INFO - duration: 1:52 (mm:ss),train miou:0.8882,val miou:0.7039 best miou: 0.724
+2026-04-06 12:20:03,216 - INFO - epoch:36/110 || loss: 1.0258 || cls: 0.3626
+2026-04-06 12:20:05,959 - INFO - duration: 1:52 (mm:ss),train miou:0.8888,val miou:0.7037 best miou: 0.724
+2026-04-06 12:21:56,389 - INFO - epoch:37/110 || loss: 1.0256 || cls: 0.3706
+2026-04-06 12:21:58,891 - INFO - duration: 1:52 (mm:ss),train miou:0.8892,val miou:0.6918 best miou: 0.724
+2026-04-06 12:23:49,604 - INFO - epoch:38/110 || loss: 1.0234 || cls: 0.374
+2026-04-06 12:23:51,917 - INFO - duration: 1:52 (mm:ss),train miou:0.8919,val miou:0.6994 best miou: 0.724
+2026-04-06 12:25:42,902 - INFO - epoch:39/110 || loss: 1.0233 || cls: 0.3836
+2026-04-06 12:25:45,244 - INFO - duration: 1:53 (mm:ss),train miou:0.8916,val miou:0.7127 best miou: 0.724
+2026-04-06 12:27:35,827 - INFO - epoch:40/110 || loss: 1.02 || cls: 0.3803
+2026-04-06 12:27:37,997 - INFO - duration: 1:52 (mm:ss),train miou:0.8955,val miou:0.7097 best miou: 0.724
+2026-04-06 12:29:27,934 - INFO - epoch:41/110 || loss: 1.0217 || cls: 0.3896
+2026-04-06 12:29:30,333 - INFO - duration: 1:52 (mm:ss),train miou:0.8935,val miou:0.6969 best miou: 0.724
+2026-04-06 12:31:20,979 - INFO - epoch:42/110 || loss: 1.0227 || cls: 0.4025
+2026-04-06 12:31:23,372 - INFO - duration: 1:52 (mm:ss),train miou:0.8898,val miou:0.6983 best miou: 0.724
+2026-04-06 12:33:13,814 - INFO - epoch:43/110 || loss: 1.0169 || cls: 0.3927
+2026-04-06 12:33:16,524 - INFO - duration: 1:53 (mm:ss),train miou:0.9002,val miou:0.7226 best miou: 0.724
+2026-04-06 12:35:06,906 - INFO - epoch:44/110 || loss: 1.0188 || cls: 0.4065
+2026-04-06 12:35:09,850 - INFO - duration: 1:53 (mm:ss),train miou:0.8978,val miou:0.711 best miou: 0.724
+2026-04-06 12:37:00,287 - INFO - epoch:45/110 || loss: 1.0175 || cls: 0.4087
+2026-04-06 12:37:02,462 - INFO - duration: 1:52 (mm:ss),train miou:0.9014,val miou:0.7323 best miou: 0.724
+2026-04-06 12:37:02,462 - INFO - ==============================  new best! epoch:45,train:0.9014,val:0.7323  ==============================
+2026-04-06 12:38:53,016 - INFO - epoch:46/110 || loss: 1.0154 || cls: 0.412
+2026-04-06 12:38:55,577 - INFO - duration: 1:52 (mm:ss),train miou:0.9001,val miou:0.726 best miou: 0.7323
+2026-04-06 12:40:45,688 - INFO - epoch:47/110 || loss: 1.0135 || cls: 0.4109
+2026-04-06 12:40:48,044 - INFO - duration: 1:52 (mm:ss),train miou:0.9046,val miou:0.7223 best miou: 0.7323
+2026-04-06 12:42:38,204 - INFO - epoch:48/110 || loss: 1.0167 || cls: 0.4279
+2026-04-06 12:42:40,817 - INFO - duration: 1:52 (mm:ss),train miou:0.9012,val miou:0.7349 best miou: 0.7323
+2026-04-06 12:42:40,817 - INFO - ==============================  new best! epoch:48,train:0.9012,val:0.7349  ==============================
+2026-04-06 12:44:31,079 - INFO - epoch:49/110 || loss: 1.0154 || cls: 0.4263
+2026-04-06 12:44:33,506 - INFO - duration: 1:52 (mm:ss),train miou:0.9026,val miou:0.7224 best miou: 0.7349
+2026-04-06 12:46:23,712 - INFO - epoch:50/110 || loss: 1.011 || cls: 0.4275
+2026-04-06 12:46:26,224 - INFO - duration: 1:52 (mm:ss),train miou:0.9079,val miou:0.7003 best miou: 0.7349
+2026-04-06 12:48:16,384 - INFO - epoch:51/110 || loss: 1.0105 || cls: 0.4329
+2026-04-06 12:48:18,766 - INFO - duration: 1:52 (mm:ss),train miou:0.9083,val miou:0.6976 best miou: 0.7349
+2026-04-06 12:50:09,086 - INFO - epoch:52/110 || loss: 1.0102 || cls: 0.4446
+2026-04-06 12:50:11,283 - INFO - duration: 1:52 (mm:ss),train miou:0.91,val miou:0.7302 best miou: 0.7349
+2026-04-06 12:52:01,438 - INFO - epoch:53/110 || loss: 1.0091 || cls: 0.4412
+2026-04-06 12:52:03,907 - INFO - duration: 1:52 (mm:ss),train miou:0.9103,val miou:0.7193 best miou: 0.7349
+2026-04-06 12:53:54,208 - INFO - epoch:54/110 || loss: 1.0093 || cls: 0.4552
+2026-04-06 12:53:56,583 - INFO - duration: 1:52 (mm:ss),train miou:0.9086,val miou:0.7284 best miou: 0.7349
+2026-04-06 12:55:46,570 - INFO - epoch:55/110 || loss: 1.0053 || cls: 0.4457
+2026-04-06 12:55:48,966 - INFO - duration: 1:52 (mm:ss),train miou:0.9135,val miou:0.7138 best miou: 0.7349
+2026-04-06 12:57:39,378 - INFO - epoch:56/110 || loss: 1.0049 || cls: 0.454
+2026-04-06 12:57:41,854 - INFO - duration: 1:52 (mm:ss),train miou:0.915,val miou:0.7382 best miou: 0.7349
+2026-04-06 12:57:41,855 - INFO - ==============================  new best! epoch:56,train:0.915,val:0.7382  ==============================
+2026-04-06 12:59:32,131 - INFO - epoch:57/110 || loss: 1.0044 || cls: 0.4581
+2026-04-06 12:59:34,393 - INFO - duration: 1:52 (mm:ss),train miou:0.9148,val miou:0.7329 best miou: 0.7382
+2026-04-06 13:01:24,976 - INFO - epoch:58/110 || loss: 1.0028 || cls: 0.4632
+2026-04-06 13:01:27,286 - INFO - duration: 1:52 (mm:ss),train miou:0.9164,val miou:0.7291 best miou: 0.7382
+2026-04-06 13:03:17,488 - INFO - epoch:59/110 || loss: 1.0053 || cls: 0.4766
+2026-04-06 13:03:20,147 - INFO - duration: 1:52 (mm:ss),train miou:0.9157,val miou:0.712 best miou: 0.7382
+2026-04-06 13:05:10,451 - INFO - epoch:60/110 || loss: 1.0018 || cls: 0.473
+2026-04-06 13:05:12,838 - INFO - duration: 1:52 (mm:ss),train miou:0.9192,val miou:0.7492 best miou: 0.7382
+2026-04-06 13:05:12,838 - INFO - ==============================  new best! epoch:60,train:0.9192,val:0.7492  ==============================
+2026-04-06 13:07:03,011 - INFO - epoch:61/110 || loss: 1.0007 || cls: 0.472
+2026-04-06 13:07:05,297 - INFO - duration: 1:52 (mm:ss),train miou:0.9209,val miou:0.7148 best miou: 0.7492
+2026-04-06 13:08:55,258 - INFO - epoch:62/110 || loss: 1.0004 || cls: 0.4853
+2026-04-06 13:08:57,747 - INFO - duration: 1:52 (mm:ss),train miou:0.9195,val miou:0.7215 best miou: 0.7492
+2026-04-06 13:10:48,047 - INFO - epoch:63/110 || loss: 1.0006 || cls: 0.486
+2026-04-06 13:10:50,688 - INFO - duration: 1:52 (mm:ss),train miou:0.9206,val miou:0.7215 best miou: 0.7492
+2026-04-06 13:12:40,777 - INFO - epoch:64/110 || loss: 0.9956 || cls: 0.4706
+2026-04-06 13:12:43,000 - INFO - duration: 1:52 (mm:ss),train miou:0.9285,val miou:0.7007 best miou: 0.7492
+2026-04-06 13:14:33,314 - INFO - epoch:65/110 || loss: 0.9969 || cls: 0.498
+2026-04-06 13:14:35,895 - INFO - duration: 1:52 (mm:ss),train miou:0.9257,val miou:0.735 best miou: 0.7492
+2026-04-06 13:16:25,972 - INFO - epoch:66/110 || loss: 0.9947 || cls: 0.4834
+2026-04-06 13:16:28,287 - INFO - duration: 1:52 (mm:ss),train miou:0.9285,val miou:0.7211 best miou: 0.7492
+2026-04-06 13:18:18,173 - INFO - epoch:67/110 || loss: 0.997 || cls: 0.5033
+2026-04-06 13:18:20,615 - INFO - duration: 1:52 (mm:ss),train miou:0.9238,val miou:0.7417 best miou: 0.7492
+2026-04-06 13:20:10,829 - INFO - epoch:68/110 || loss: 0.9929 || cls: 0.4897
+2026-04-06 13:20:13,249 - INFO - duration: 1:52 (mm:ss),train miou:0.9297,val miou:0.7474 best miou: 0.7492
+2026-04-06 13:22:03,265 - INFO - epoch:69/110 || loss: 0.9941 || cls: 0.5012
+2026-04-06 13:22:05,566 - INFO - duration: 1:52 (mm:ss),train miou:0.929,val miou:0.726 best miou: 0.7492
+2026-04-06 13:23:55,584 - INFO - epoch:70/110 || loss: 0.9912 || cls: 0.4921
+2026-04-06 13:23:57,936 - INFO - duration: 1:52 (mm:ss),train miou:0.9322,val miou:0.7466 best miou: 0.7492
+2026-04-06 13:25:48,276 - INFO - epoch:71/110 || loss: 0.9912 || cls: 0.4927
+2026-04-06 13:25:50,660 - INFO - duration: 1:52 (mm:ss),train miou:0.9325,val miou:0.7388 best miou: 0.7492
+2026-04-06 13:27:40,603 - INFO - epoch:72/110 || loss: 0.9909 || cls: 0.5017
+2026-04-06 13:27:43,184 - INFO - duration: 1:52 (mm:ss),train miou:0.9329,val miou:0.7306 best miou: 0.7492
+2026-04-06 13:29:33,169 - INFO - epoch:73/110 || loss: 0.9895 || cls: 0.5036
+2026-04-06 13:29:35,719 - INFO - duration: 1:52 (mm:ss),train miou:0.9356,val miou:0.7209 best miou: 0.7492
+2026-04-06 13:31:25,865 - INFO - epoch:74/110 || loss: 0.9894 || cls: 0.5116
+2026-04-06 13:31:28,122 - INFO - duration: 1:52 (mm:ss),train miou:0.9346,val miou:0.7415 best miou: 0.7492
+2026-04-06 13:33:18,033 - INFO - epoch:75/110 || loss: 0.9858 || cls: 0.4985
+2026-04-06 13:33:20,654 - INFO - duration: 1:52 (mm:ss),train miou:0.9398,val miou:0.7504 best miou: 0.7492
+2026-04-06 13:33:20,654 - INFO - ==============================  new best! epoch:75,train:0.9398,val:0.7504  ==============================
+2026-04-06 13:35:10,861 - INFO - epoch:76/110 || loss: 0.9868 || cls: 0.5072
+2026-04-06 13:35:13,318 - INFO - duration: 1:52 (mm:ss),train miou:0.9378,val miou:0.7406 best miou: 0.7504
+2026-04-06 13:37:03,236 - INFO - epoch:77/110 || loss: 0.9851 || cls: 0.5086
+2026-04-06 13:37:05,690 - INFO - duration: 1:52 (mm:ss),train miou:0.939,val miou:0.7303 best miou: 0.7504
+2026-04-06 13:38:55,836 - INFO - epoch:78/110 || loss: 0.9843 || cls: 0.5037
+2026-04-06 13:38:58,613 - INFO - duration: 1:52 (mm:ss),train miou:0.9415,val miou:0.7352 best miou: 0.7504
+2026-04-06 13:40:48,628 - INFO - epoch:79/110 || loss: 0.9845 || cls: 0.5109
+2026-04-06 13:40:51,254 - INFO - duration: 1:52 (mm:ss),train miou:0.9415,val miou:0.7362 best miou: 0.7504
+2026-04-06 13:42:41,641 - INFO - epoch:80/110 || loss: 0.982 || cls: 0.5003
+2026-04-06 13:42:44,090 - INFO - duration: 1:52 (mm:ss),train miou:0.9439,val miou:0.7517 best miou: 0.7504
+2026-04-06 13:42:44,090 - INFO - ==============================  new best! epoch:80,train:0.9439,val:0.7517  ==============================
+2026-04-06 13:44:34,347 - INFO - epoch:81/110 || loss: 0.9825 || cls: 0.5099
+2026-04-06 13:44:36,622 - INFO - duration: 1:52 (mm:ss),train miou:0.9444,val miou:0.742 best miou: 0.7517
+2026-04-06 13:46:26,774 - INFO - epoch:82/110 || loss: 0.9803 || cls: 0.5036
+2026-04-06 13:46:29,118 - INFO - duration: 1:52 (mm:ss),train miou:0.946,val miou:0.7397 best miou: 0.7517
+2026-04-06 13:48:19,135 - INFO - epoch:83/110 || loss: 0.9784 || cls: 0.4934
+2026-04-06 13:48:21,552 - INFO - duration: 1:52 (mm:ss),train miou:0.949,val miou:0.7328 best miou: 0.7517
+2026-04-06 13:50:11,724 - INFO - epoch:84/110 || loss: 0.9793 || cls: 0.5028
+2026-04-06 13:50:14,166 - INFO - duration: 1:52 (mm:ss),train miou:0.9475,val miou:0.7599 best miou: 0.7517
+2026-04-06 13:50:14,166 - INFO - ==============================  new best! epoch:84,train:0.9475,val:0.7599  ==============================
+2026-04-06 13:52:04,146 - INFO - epoch:85/110 || loss: 0.9789 || cls: 0.5107
+2026-04-06 13:52:06,778 - INFO - duration: 1:52 (mm:ss),train miou:0.9479,val miou:0.7416 best miou: 0.7599
+2026-04-06 13:53:56,972 - INFO - epoch:86/110 || loss: 0.9777 || cls: 0.4988
+2026-04-06 13:53:59,452 - INFO - duration: 1:52 (mm:ss),train miou:0.9494,val miou:0.7571 best miou: 0.7599
+2026-04-06 13:55:49,746 - INFO - epoch:87/110 || loss: 0.9766 || cls: 0.4946
+2026-04-06 13:55:52,315 - INFO - duration: 1:52 (mm:ss),train miou:0.9513,val miou:0.7413 best miou: 0.7599
+2026-04-06 13:57:42,600 - INFO - epoch:88/110 || loss: 0.9758 || cls: 0.4943
+2026-04-06 13:57:44,961 - INFO - duration: 1:52 (mm:ss),train miou:0.9518,val miou:0.7495 best miou: 0.7599
+2026-04-06 13:59:35,151 - INFO - epoch:89/110 || loss: 0.9752 || cls: 0.4893
+2026-04-06 13:59:37,520 - INFO - duration: 1:52 (mm:ss),train miou:0.953,val miou:0.7385 best miou: 0.7599
+2026-04-06 14:01:28,149 - INFO - epoch:90/110 || loss: 0.9745 || cls: 0.4862
+2026-04-06 14:01:30,672 - INFO - duration: 1:53 (mm:ss),train miou:0.9534,val miou:0.7524 best miou: 0.7599
+2026-04-06 14:03:20,970 - INFO - epoch:91/110 || loss: 0.9745 || cls: 0.4924
+2026-04-06 14:03:23,378 - INFO - duration: 1:52 (mm:ss),train miou:0.9537,val miou:0.752 best miou: 0.7599
+2026-04-06 14:05:13,750 - INFO - epoch:92/110 || loss: 0.9733 || cls: 0.4858
+2026-04-06 14:05:16,159 - INFO - duration: 1:52 (mm:ss),train miou:0.9554,val miou:0.7449 best miou: 0.7599
+2026-04-06 14:07:06,010 - INFO - epoch:93/110 || loss: 0.9724 || cls: 0.4829
+2026-04-06 14:07:08,273 - INFO - duration: 1:51 (mm:ss),train miou:0.9562,val miou:0.7505 best miou: 0.7599
+2026-04-06 14:08:58,254 - INFO - epoch:94/110 || loss: 0.9724 || cls: 0.4838
+2026-04-06 14:09:00,635 - INFO - duration: 1:52 (mm:ss),train miou:0.9564,val miou:0.7519 best miou: 0.7599
+2026-04-06 14:10:50,622 - INFO - epoch:95/110 || loss: 0.9712 || cls: 0.4835
+2026-04-06 14:10:53,086 - INFO - duration: 1:52 (mm:ss),train miou:0.958,val miou:0.7495 best miou: 0.7599
+2026-04-06 14:12:43,127 - INFO - epoch:96/110 || loss: 0.9711 || cls: 0.4805
+2026-04-06 14:12:45,907 - INFO - duration: 1:52 (mm:ss),train miou:0.9584,val miou:0.7498 best miou: 0.7599
+2026-04-06 14:14:36,031 - INFO - epoch:97/110 || loss: 0.9706 || cls: 0.4762
+2026-04-06 14:14:38,657 - INFO - duration: 1:52 (mm:ss),train miou:0.9584,val miou:0.7552 best miou: 0.7599
+2026-04-06 14:16:28,999 - INFO - epoch:98/110 || loss: 0.9704 || cls: 0.4752
+2026-04-06 14:16:31,646 - INFO - duration: 1:52 (mm:ss),train miou:0.9587,val miou:0.7565 best miou: 0.7599
+2026-04-06 14:18:21,852 - INFO - epoch:99/110 || loss: 0.9698 || cls: 0.4748
+2026-04-06 14:18:24,221 - INFO - duration: 1:52 (mm:ss),train miou:0.9596,val miou:0.7549 best miou: 0.7599
+2026-04-06 14:20:14,442 - INFO - epoch:100/110 || loss: 0.9696 || cls: 0.4734
+2026-04-06 14:20:17,018 - INFO - duration: 1:52 (mm:ss),train miou:0.9601,val miou:0.7549 best miou: 0.7599
+2026-04-06 14:22:07,169 - INFO - epoch:101/110 || loss: 0.9696 || cls: 0.471
+2026-04-06 14:22:09,902 - INFO - duration: 1:52 (mm:ss),train miou:0.96,val miou:0.7539 best miou: 0.7599
+2026-04-06 14:23:59,926 - INFO - epoch:102/110 || loss: 0.9693 || cls: 0.4713
+2026-04-06 14:24:02,352 - INFO - duration: 1:52 (mm:ss),train miou:0.9607,val miou:0.7575 best miou: 0.7599
+2026-04-06 14:25:52,658 - INFO - epoch:103/110 || loss: 0.9689 || cls: 0.4724
+2026-04-06 14:25:55,316 - INFO - duration: 1:52 (mm:ss),train miou:0.961,val miou:0.7535 best miou: 0.7599
+2026-04-06 14:27:45,841 - INFO - epoch:104/110 || loss: 0.9688 || cls: 0.4708
+2026-04-06 14:27:48,390 - INFO - duration: 1:52 (mm:ss),train miou:0.9606,val miou:0.7579 best miou: 0.7599
+2026-04-06 14:29:38,685 - INFO - epoch:105/110 || loss: 0.9686 || cls: 0.4694
+2026-04-06 14:29:41,546 - INFO - duration: 1:53 (mm:ss),train miou:0.961,val miou:0.7547 best miou: 0.7599
+2026-04-06 14:31:32,226 - INFO - epoch:106/110 || loss: 0.9685 || cls: 0.4693
+2026-04-06 14:31:34,669 - INFO - duration: 1:52 (mm:ss),train miou:0.9614,val miou:0.7551 best miou: 0.7599
+2026-04-06 14:33:25,164 - INFO - epoch:107/110 || loss: 0.9684 || cls: 0.4692
+2026-04-06 14:33:27,732 - INFO - duration: 1:52 (mm:ss),train miou:0.9611,val miou:0.7581 best miou: 0.7599
+2026-04-06 14:35:17,892 - INFO - epoch:108/110 || loss: 0.9683 || cls: 0.4691
+2026-04-06 14:35:20,202 - INFO - duration: 1:52 (mm:ss),train miou:0.9617,val miou:0.7566 best miou: 0.7599
+2026-04-06 14:37:10,594 - INFO - epoch:109/110 || loss: 0.9684 || cls: 0.4674
+2026-04-06 14:37:12,968 - INFO - duration: 1:52 (mm:ss),train miou:0.9612,val miou:0.7566 best miou: 0.7599
+2026-04-06 14:37:13,134 - INFO - ================================================================================
+2026-04-06 14:37:13,134 - INFO - Training completed! Starting automatic testing with best model...
+2026-04-06 14:37:13,134 - INFO - ================================================================================
+2026-04-06 14:37:48,404 - INFO - FINAL TEST RESULTS:
+2026-04-06 14:37:48,404 - INFO - Best model test accuracy: 0.9361
+2026-04-06 14:37:48,404 - INFO - Best model test mAcc: 0.8222
+2026-04-06 14:37:48,404 - INFO - Best model test mIoU: 0.773
+2026-04-06 14:37:48,404 - INFO - ================================================================================
+2026-04-06 14:37:48,404 - INFO - Training and testing process completed!
+2026-04-06 14:37:48,404 - INFO - ================================================================================