fixed bug with alternate_corr flag

hm-ysjiang · Aug 28, 2020 · 13198c3 · 13198c3
1 parent 01ad964
commit 13198c3
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 31 deletions.
diff --git a/core/corr.py b/core/corr.py
@@ -60,26 +60,6 @@ def corr(fmap1, fmap2):
         return corr  / torch.sqrt(torch.tensor(dim).float())
 
 
-class CorrLayer(torch.autograd.Function):
-    @staticmethod
-    def forward(ctx, fmap1, fmap2, coords, r):
-        fmap1 = fmap1.contiguous()
-        fmap2 = fmap2.contiguous()
-        coords = coords.contiguous()
-        ctx.save_for_backward(fmap1, fmap2, coords)
-        ctx.r = r
-        corr, = correlation_cudaz.forward(fmap1, fmap2, coords, ctx.r)
-        return corr
-
-    @staticmethod
-    def backward(ctx, grad_corr):
-        fmap1, fmap2, coords = ctx.saved_tensors
-        grad_corr = grad_corr.contiguous()
-        fmap1_grad, fmap2_grad, coords_grad = \
-            correlation_cudaz.backward(fmap1, fmap2, coords, grad_corr, ctx.r)
-        return fmap1_grad, fmap2_grad, coords_grad, None
-
-
 class AlternateCorrBlock:
     def __init__(self, fmap1, fmap2, num_levels=4, radius=4):
         self.num_levels = num_levels
@@ -92,20 +72,20 @@ def __init__(self, fmap1, fmap2, num_levels=4, radius=4):
             self.pyramid.append((fmap1, fmap2))
 
     def __call__(self, coords):
-
         coords = coords.permute(0, 2, 3, 1)
         B, H, W, _ = coords.shape
+        dim = self.pyramid[0][0].shape[1]
 
         corr_list = []
         for i in range(self.num_levels):
             r = self.radius
-            fmap1_i = self.pyramid[0][0].permute(0, 2, 3, 1)
-            fmap2_i = self.pyramid[i][1].permute(0, 2, 3, 1)
+            fmap1_i = self.pyramid[0][0].permute(0, 2, 3, 1).contiguous()
+            fmap2_i = self.pyramid[i][1].permute(0, 2, 3, 1).contiguous()
 
             coords_i = (coords / 2**i).reshape(B, 1, H, W, 2).contiguous()
-            corr = alt_cuda_corr(fmap1_i, fmap2_i, coords_i, r)
+            corr, = alt_cuda_corr.forward(fmap1_i, fmap2_i, coords_i, r)
             corr_list.append(corr.squeeze(1))
 
         corr = torch.stack(corr_list, dim=1)
         corr = corr.reshape(B, -1, H, W)
-        return corr / 16.0
+        return corr / torch.sqrt(torch.tensor(dim).float())
diff --git a/core/raft.py b/core/raft.py
@@ -38,11 +38,11 @@ def __init__(self, args):
             args.corr_levels = 4
             args.corr_radius = 4
 
-        if 'dropout' not in args._get_kwargs():
-            args.dropout = 0
+        if 'dropout' not in self.args:
+            self.args.dropout = 0
 
-        if 'alternate_corr' not in args._get_kwargs():
-            args.alternate_corr = False
+        if 'alternate_corr' not in self.args:
+            self.args.alternate_corr = False
 
         # feature network, context network, and update block
         if args.small:
@@ -55,7 +55,6 @@ def __init__(self, args):
             self.cnet = BasicEncoder(output_dim=hdim+cdim, norm_fn='batch', dropout=args.dropout)
             self.update_block = BasicUpdateBlock(self.args, hidden_dim=hdim)
 
-
     def freeze_bn(self):
         for m in self.modules():
             if isinstance(m, nn.BatchNorm2d):
@@ -103,7 +102,7 @@ def forward(self, image1, image2, iters=12, flow_init=None, upsample=True, test_
         fmap1 = fmap1.float()
         fmap2 = fmap2.float()
         if self.args.alternate_corr:
-            corr_fn = CorrBlockAlternate(fmap1, fmap2, radius=self.args.corr_radius)
+            corr_fn = AlternateCorrBlock(fmap1, fmap2, radius=self.args.corr_radius)
         else:
             corr_fn = CorrBlock(fmap1, fmap2, radius=self.args.corr_radius)