formatting fix

torchrs/datasets/eurosat.py

@@ -11,7 +11,7 @@ class EuroSATRGB(ImageFolder):
     """ Sentinel-2 RGB Land Cover Classification dataset from 'EuroSAT: A Novel Dataset
     and Deep Learning Benchmark for Land Use and Land Cover Classification', Helber at al. (2017)
     https://arxiv.org/abs/1709.00029
-    
+
     'We present a novel dataset based on Sentinel-2 satellite images covering 13 spectral
     bands and consisting out of 10 classes with in total 27,000 labeled and geo-referenced images.'
 
@@ -32,7 +32,7 @@ class EuroSATMS(ImageFolder):
     """ Sentinel-2 RGB Land Cover Classification dataset from 'EuroSAT: A Novel Dataset
     and Deep Learning Benchmark for Land Use and Land Cover Classification', Helber at al. (2017)
     https://arxiv.org/abs/1709.00029
-    
+
     'We present a novel dataset based on Sentinel-2 satellite images covering 13 spectral
     bands and consisting out of 10 classes with in total 27,000 labeled and geo-referenced images.'
 

torchrs/datasets/fair1m.py

@@ -1,7 +1,7 @@
 import os
 from glob import glob
 from xml.etree import ElementTree
-from typing import List, Dict, Tuple
+from typing import List, Dict
 
 import torch
 import numpy as np
@@ -74,6 +74,7 @@ class FAIR1M(torch.utils.data.Dataset):
         "Intersection":     {"id": 35, "category": "Road"},
         "Bridge":           {"id": 36, "category": "Road"}
     }
+
     def __init__(
         self,
         root: str = ".data/fair1m",

torchrs/models/__init__.py

@@ -4,5 +4,5 @@
 
 
 __all__ = [
-    "RAMS", "EarlyFusion", "FCEF", "FCSiamConc", "FCSiamDiff"
+    "RAMS", "EarlyFusion", "Siam", "FCEF", "FCSiamConc", "FCSiamDiff"
 ]

torchrs/models/fc_cd.py

@@ -25,7 +25,6 @@ def forward(self, x: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
         return self.pool(x), x
 
 
-
 class DeConvBlock(nn.Sequential):
 
     def __init__(self, filters: List[int], kernel_size: int = 3, dropout: float = 0.2):
@@ -154,7 +153,7 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
 
         # Concat skips
         skips = [rearrange(skip, "(b t) c h w -> b (t c) h w", t=t) for skip in skips]
-        
+
         # Only first input encoding is passed directly to decoder
         x = rearrange(x, "(b t) c h w -> b t c h w", t=t)
         x = x[:, 0, ...]

torchrs/models/oscd.py

@@ -26,7 +26,6 @@ class EarlyFusion(nn.Module):
     def __init__(self, channels: int = 3, t: int = 2, num_classes: int = 2):
         super().__init__()
         filters = [channels * t, 32, 32, 64, 64, 128, 128]
-        kernel_size = 3
         dropout = 0.2
         self.encoder = nn.Sequential(
             *[ConvBlock(filters[i-1], filters[i]) for i in range(1, len(filters))],
@@ -35,9 +34,9 @@ def __init__(self, channels: int = 3, t: int = 2, num_classes: int = 2):
         )
         self.mlp = nn.Sequential(
             nn.Linear(128, 8),
-            nn.BatchNorm1d(8), 
+            nn.BatchNorm1d(8),
             nn.ReLU(),
-            nn.Dropout2d(dropout), 
+            nn.Dropout2d(dropout),
             nn.Linear(8, num_classes)
         )
 
@@ -60,17 +59,16 @@ class Siam(nn.Module):
     def __init__(self, channels: int = 3, t: int = 2, num_classes: int = 2):
         super().__init__()
         filters = [channels, 64, 64, 128]
-        kernel_size = 3
         dropout = 0.2
         self.encoder = nn.Sequential(
             *[ConvBlock(filters[i-1], filters[i]) for i in range(1, len(filters))],
             ConvBlock(filters[-1], 128, dropout=0.0),
         )
         self.mlp = nn.Sequential(
             nn.Linear(t*128*7*7, 64),
-            nn.BatchNorm1d(64), 
+            nn.BatchNorm1d(64),
             nn.ReLU(),
-            nn.Dropout2d(dropout), 
+            nn.Dropout2d(dropout),
             nn.Linear(64, num_classes)
         )
 

torchrs/models/rams.py

@@ -1,6 +1,4 @@
 """ Referenced from official TF implementation https://github.com/EscVM/RAMS/blob/master/utils/network.py """
-import math
-
 import torch
 import torch.nn as nn
 from einops import rearrange

torchrs/transforms.py

@@ -49,7 +49,7 @@ def __call__(self, x: np.ndarray) -> torch.Tensor:
 
 
 class ToDtype(object):
-    
+
     def __init__(self, dtype: torch.dtype):
         self.dtype = dtype