fix value residual

setup.py

@@ -6,7 +6,7 @@ with open('README.md') as f:
 setup(
   name = 'vit-pytorch',
   packages = find_packages(exclude=['examples']),
-  version = '1.8.6',
+  version = '1.8.7',
   license='MIT',
   description = 'Vision Transformer (ViT) - Pytorch',
   long_description=long_description,

vit_pytorch/simple_vit_with_value_residual.py

@@ -57,7 +57,7 @@ class Attention(Module):
         q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> b h n d', h = self.heads), qkv)
 
         if exists(value_residual):
-            v = v + value_residual
+            v = 0.5 * (v + value_residual)
 
         dots = torch.matmul(q, k.transpose(-1, -2)) * self.scale