Fixed unfused attn2d scale

tests/test_layers.py

@@ -1,7 +1,8 @@
+import pytest
 import torch
 import torch.nn as nn
 
-from timm.layers import create_act_layer, set_layer_config, get_act_layer, get_act_fn
+from timm.layers import create_act_layer, set_layer_config, get_act_layer, get_act_fn, Attention2d
 
 import importlib
 import os
@@ -119,3 +120,27 @@ def test_get_act_fn_none():
     assert get_act_fn(None) is None
     assert get_act_fn('') is None
 
+
+@pytest.mark.parametrize("bias", [True, False])
+@pytest.mark.parametrize("expand_first", [True, False])
+@pytest.mark.parametrize("head_first", [True, False])
+@pytest.mark.parametrize("attn_mask", [True, False])
+def test_attn2d(bias, expand_first, head_first, attn_mask):
+    x = torch.randn(1, 128, 32, 48)
+    attn = Attention2d(
+        128, 128, num_heads=4, bias=bias, expand_first=expand_first, head_first=head_first
+    )
+    
+    if attn_mask:
+        mask = torch.randint(0, 1, size=(32 * 48, 32 * 48), dtype=torch.float32)
+    else:
+        mask = None
+    
+    o1 = attn(x, mask)
+    attn.fused_attn = False
+    o2 = attn(x, mask)
+    
+    assert torch.allclose(o1, o2, atol=1e-5), f"{torch.abs(o1 - o2).max()}"
+
+    
+    

timm/layers/attention2d.py

@@ -312,7 +312,6 @@ class Attention2d(nn.Module):
         self.num_heads = num_heads
         self.dim_head = dim_attn // num_heads
         self.head_first = head_first
-        self.scale = num_heads ** -0.5
         self.fused_attn = use_fused_attn()
 
         self.qkv = nn.Conv2d(dim, dim_attn * 3, 1, bias=bias)
@@ -337,14 +336,15 @@ class Attention2d(nn.Module):
                 dropout_p=self.attn_drop.p if self.training else 0.,
             ).transpose(-1, -2).reshape(B, -1, H, W)
         else:
-            q = q * self.scale
-            attn = q.transpose(-2, -1) @ k
+            q = q.transpose(-1, -2)
+            v = v.transpose(-1, -2)
+            attn = q @ k * q.size(-1) ** -0.5
             if attn_mask is not None:
                 # NOTE: assumes mask is float and in correct shape
                 attn = attn + attn_mask
             attn = attn.softmax(dim=-1)
             attn = self.attn_drop(attn)
-            x = (v @ attn.transpose(-2, -1)).view(B, -1, H, W)
+            x = (attn @ v).transpose(-1, -2).reshape(B, -1, H, W)
 
         x = self.proj(x)
         x = self.proj_drop(x)