remove attention abstraction (Comfy-Org#5324)

Author: contentis

comfy/ldm/modules/diffusionmodules/mmdit.py

@@ -5,7 +5,7 @@
 import numpy as np
 import torch
 import torch.nn as nn
-from .. import attention
+from ..attention import optimized_attention
 from einops import rearrange, repeat
 from .util import timestep_embedding
 import comfy.ops
@@ -266,8 +266,6 @@ def split_qkv(qkv, head_dim):
     qkv = qkv.reshape(qkv.shape[0], qkv.shape[1], 3, -1, head_dim).movedim(2, 0)
     return qkv[0], qkv[1], qkv[2]
 
-def optimized_attention(qkv, num_heads):
-    return attention.optimized_attention(qkv[0], qkv[1], qkv[2], num_heads)
 
 class SelfAttention(nn.Module):
     ATTENTION_MODES = ("xformers", "torch", "torch-hb", "math", "debug")
@@ -326,9 +324,9 @@ def post_attention(self, x: torch.Tensor) -> torch.Tensor:
         return x
 
     def forward(self, x: torch.Tensor) -> torch.Tensor:
-        qkv = self.pre_attention(x)
+        q, k, v = self.pre_attention(x)
         x = optimized_attention(
-            qkv, num_heads=self.num_heads
+            q, k, v, heads=self.num_heads
         )
         x = self.post_attention(x)
         return x
@@ -531,8 +529,8 @@ def forward(self, x: torch.Tensor, c: torch.Tensor) -> torch.Tensor:
         assert not self.pre_only
         qkv, intermediates = self.pre_attention(x, c)
         attn = optimized_attention(
-            qkv,
-            num_heads=self.attn.num_heads,
+            qkv[0], qkv[1], qkv[2],
+            heads=self.attn.num_heads,
         )
         return self.post_attention(attn, *intermediates)
 
@@ -557,8 +555,8 @@ def _block_mixing(context, x, context_block, x_block, c):
     qkv = tuple(o)
 
     attn = optimized_attention(
-        qkv,
-        num_heads=x_block.attn.num_heads,
+        qkv[0], qkv[1], qkv[2],
+        heads=x_block.attn.num_heads,
     )
     context_attn, x_attn = (
         attn[:, : context_qkv[0].shape[1]],
@@ -642,7 +640,7 @@ def __init__(self, dim, heads=8, dim_head=64, dtype=None, device=None, operation
     def forward(self, x):
         qkv = self.qkv(x)
         q, k, v = split_qkv(qkv, self.dim_head)
-        x = optimized_attention((q.reshape(q.shape[0], q.shape[1], -1), k, v), self.heads)
+        x = optimized_attention(q.reshape(q.shape[0], q.shape[1], -1), k, v, heads=self.heads)
         return self.proj(x)
 
 class ContextProcessorBlock(nn.Module):