Make last PR not crash comfy on old pytorch. (#9324)

comfy/ldm/hunyuan3d/vae.py

@@ -178,7 +178,7 @@ class FourierEmbedder(nn.Module):
 
 class CrossAttentionProcessor:
     def __call__(self, attn, q, k, v):
-        out = ops.scaled_dot_product_attention(q, k, v)
+        out = comfy.ops.scaled_dot_product_attention(q, k, v)
         return out
 
 

comfy/ldm/modules/attention.py

@@ -448,7 +448,7 @@ def attention_pytorch(q, k, v, heads, mask=None, attn_precision=None, skip_resha
             mask = mask.unsqueeze(1)
 
     if SDP_BATCH_LIMIT >= b:
-        out = ops.scaled_dot_product_attention(q, k, v, attn_mask=mask, dropout_p=0.0, is_causal=False)
+        out = comfy.ops.scaled_dot_product_attention(q, k, v, attn_mask=mask, dropout_p=0.0, is_causal=False)
         if not skip_output_reshape:
             out = (
                 out.transpose(1, 2).reshape(b, -1, heads * dim_head)
@@ -461,7 +461,7 @@ def attention_pytorch(q, k, v, heads, mask=None, attn_precision=None, skip_resha
                 if mask.shape[0] > 1:
                     m = mask[i : i + SDP_BATCH_LIMIT]
 
-            out[i : i + SDP_BATCH_LIMIT] = ops.scaled_dot_product_attention(
+            out[i : i + SDP_BATCH_LIMIT] = comfy.ops.scaled_dot_product_attention(
                 q[i : i + SDP_BATCH_LIMIT],
                 k[i : i + SDP_BATCH_LIMIT],
                 v[i : i + SDP_BATCH_LIMIT],

comfy/ldm/modules/diffusionmodules/model.py

@@ -285,7 +285,7 @@ def pytorch_attention(q, k, v):
     )
 
     try:
-        out = ops.scaled_dot_product_attention(q, k, v, attn_mask=None, dropout_p=0.0, is_causal=False)
+        out = comfy.ops.scaled_dot_product_attention(q, k, v, attn_mask=None, dropout_p=0.0, is_causal=False)
         out = out.transpose(2, 3).reshape(orig_shape)
     except model_management.OOM_EXCEPTION:
         logging.warning("scaled_dot_product_attention OOMed: switched to slice attention")

comfy/ops.py

@@ -23,18 +23,32 @@ from comfy.cli_args import args, PerformanceFeature
 import comfy.float
 import comfy.rmsnorm
 import contextlib
-from torch.nn.attention import SDPBackend, sdpa_kernel
+
+
+def scaled_dot_product_attention(q, k, v, *args, **kwargs):
+    return torch.nn.functional.scaled_dot_product_attention(q, k, v, *args, **kwargs)
+
+
+try:
+    if torch.cuda.is_available():
+        from torch.nn.attention import SDPBackend, sdpa_kernel
+
+        SDPA_BACKEND_PRIORITY = [
+            SDPBackend.FLASH_ATTENTION,
+            SDPBackend.EFFICIENT_ATTENTION,
+            SDPBackend.MATH,
+        ]
+
+        SDPA_BACKEND_PRIORITY.insert(0, SDPBackend.CUDNN_ATTENTION)
+
+        @sdpa_kernel(backends=SDPA_BACKEND_PRIORITY, set_priority=True)
+        def scaled_dot_product_attention(q, k, v, *args, **kwargs):
+            return torch.nn.functional.scaled_dot_product_attention(q, k, v, *args, **kwargs)
+except (ModuleNotFoundError, TypeError):
+    logging.warning("Could not set sdpa backend priority.")
 
 cast_to = comfy.model_management.cast_to #TODO: remove once no more references
 
-SDPA_BACKEND_PRIORITY = [
-    SDPBackend.FLASH_ATTENTION,
-    SDPBackend.EFFICIENT_ATTENTION,
-    SDPBackend.MATH,
-]
-if torch.cuda.is_available():
-    SDPA_BACKEND_PRIORITY.insert(0, SDPBackend.CUDNN_ATTENTION)
-
 def cast_to_input(weight, input, non_blocking=False, copy=True):
     return comfy.model_management.cast_to(weight, input.dtype, input.device, non_blocking=non_blocking, copy=copy)
 
@@ -258,10 +272,6 @@ class disable_weight_init:
         else:
             raise ValueError(f"unsupported dimensions: {dims}")
 
-    @staticmethod
-    @sdpa_kernel(backends=SDPA_BACKEND_PRIORITY, set_priority=True)
-    def scaled_dot_product_attention(q, k, v, *args, **kwargs):
-        return torch.nn.functional.scaled_dot_product_attention(q, k, v, *args, **kwargs)
 
 class manual_cast(disable_weight_init):
     class Linear(disable_weight_init.Linear):