Changes (#53)

* fix bug in cutmix (again)

* support output_stride in BiFPN

* fixed cutmix bug connected to `prev_data` shape

Author: bonlime

pytorch_tools/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.0"
+__version__ = "0.1.1"
 
 from . import fit_wrapper
 from . import losses

pytorch_tools/fit_wrapper/callbacks.py

@@ -526,11 +526,11 @@ def mixup(self, data, target):
             target_one_hot = target
         if not self.state.is_train or np.random.rand() > self.prob:
             return data, target_one_hot
-        prev_data, prev_target = data, target_one_hot if self.prev_input is None else self.prev_input
+        prev_data, prev_target = (data, target_one_hot) if self.prev_input is None else self.prev_input
+        self.prev_input = data, target_one_hot
         c = self.tb.sample()
         md = c * data + (1 - c) * prev_data
         mt = c * target_one_hot + (1 - c) * prev_target
-        self.prev_input = data, target_one_hot
         return md, mt
 
 
@@ -569,16 +569,17 @@ def cutmix(self, data, target):
             target_one_hot = target
         if not self.state.is_train or np.random.rand() > self.prob:
             return data, target_one_hot
-        prev_data, prev_target = data, target_one_hot if self.prev_input is None else self.prev_input
-        _, _, H, W = data.size()
+        prev_data, prev_target = (data, target_one_hot) if self.prev_input is None else self.prev_input
+        self.prev_input = data, target_one_hot
+        # prev_data shape can be different from current. so need to take min
+        H, W = min(data.size(2), prev_data.size(2)), min(data.size(3), prev_data.size(3))
         lam = self.tb.sample()
         lam = min([lam, 1 - lam])
         bbh1, bbw1, bbh2, bbw2 = self.rand_bbox(H, W, lam)
         # real lambda may be diffrent from sampled. adjust for it
         lam = (bbh2 - bbh1) * (bbw2 - bbw1) / (H * W)
-        data[:, bbh1:bbh2, bbw1:bbw2] = prev_data[:, bbh1:bbh2, bbw1:bbw2]
+        data[:, :, bbh1:bbh2, bbw1:bbw2] = prev_data[:, :, bbh1:bbh2, bbw1:bbw2]
         mixed_target = (1 - lam) * target_one_hot + lam * prev_target
-        self.prev_input = data, target_one_hot
         return data, mixed_target
 
     @staticmethod
@@ -607,12 +608,12 @@ def __init__(self, alpha=1.0, prob=0.5):
     def cutmix(self, data, target):
         if not self.state.is_train or np.random.rand() > self.prob:
             return data, target
-        prev_data, prev_target = data, target if self.prev_input is None else self.prev_input
-        _, _, H, W = data.size()
+        prev_data, prev_target = (data, target) if self.prev_input is None else self.prev_input
+        self.prev_input = data, target
+        H, W = min(data.size(2), prev_data.size(2)), min(data.size(3), prev_data.size(3))
         lam = self.tb.sample()
         lam = min([lam, 1 - lam])
         bbh1, bbw1, bbh2, bbw2 = self.rand_bbox(H, W, lam)
         data[:, :, bbh1:bbh2, bbw1:bbw2] = prev_data[:, :, bbh1:bbh2, bbw1:bbw2]
         target[:, :, bbh1:bbh2, bbw1:bbw2] = prev_target[:, :, bbh1:bbh2, bbw1:bbw2]
-        self.prev_input = data, target
         return data, target

pytorch_tools/modules/bifpn.py

@@ -41,13 +41,15 @@ class BiFPNLayer(nn.Module):
         p_out: features processed by 1 layer of BiFPN
     """
 
-    def __init__(self, channels=64, upsample_mode="nearest", **bn_args):
+    def __init__(self, channels=64, output_stride=32, upsample_mode="nearest", **bn_args):
         super(BiFPNLayer, self).__init__()
 
         self.up = nn.Upsample(scale_factor=2, mode=upsample_mode)
+        self.first_up = self.up if output_stride == 32 else nn.Identity()
+        last_stride = 2 if output_stride == 32 else 1
         self.down_p2 = DepthwiseSeparableConv(channels, channels, stride=2, **bn_args)
         self.down_p3 = DepthwiseSeparableConv(channels, channels, stride=2, **bn_args)
-        self.down_p4 = DepthwiseSeparableConv(channels, channels, stride=2, **bn_args)
+        self.down_p4 = DepthwiseSeparableConv(channels, channels, stride=last_stride, **bn_args)
 
         ## TODO (jamil) 11.02.2020 Rewrite this using list comprehensions
         self.fuse_p4_td = FastNormalizedFusion(in_nodes=2)
@@ -75,7 +77,7 @@ def forward(self, features):
         p5_inp, p4_inp, p3_inp, p2_inp, p1_inp = features
 
         # Top-down pathway
-        p4_td = self.p4_td(self.fuse_p4_td(p4_inp, self.up(p5_inp)))
+        p4_td = self.p4_td(self.fuse_p4_td(p4_inp, self.first_up(p5_inp)))
         p3_td = self.p3_td(self.fuse_p3_td(p3_inp, self.up(p4_td)))
         p2_out = self.p2_td(self.fuse_p2_td(p2_inp, self.up(p3_td)))
 
@@ -134,6 +136,7 @@ def __init__(
         encoder_channels,
         pyramid_channels=64,
         num_layers=1,
+        output_stride=32,
         **bn_args,
     ):
         super(BiFPN, self).__init__()
@@ -142,7 +145,7 @@ def __init__(
 
         bifpns = []
         for _ in range(num_layers):
-            bifpns.append(BiFPNLayer(pyramid_channels, **bn_args))
+            bifpns.append(BiFPNLayer(pyramid_channels, output_stride, **bn_args))
         self.bifpn = nn.Sequential(*bifpns)
 
     def forward(self, features):

pytorch_tools/segmentation_models/segm_fpn.py

@@ -97,6 +97,7 @@ def __init__(
            self.encoder.out_shapes,
            pyramid_channels=pyramid_channels,
            num_layers=num_fpn_layers,
+           output_stride=output_stride,
            **bn_args,
         )