Better memory estimation for ROCm that support mem efficient attention.

There is no way to check if the card actually supports it so it assumes
that it does if you use --use-pytorch-cross-attention with yours.

comfy/model_management.py

@@ -909,6 +909,8 @@ def pytorch_attention_flash_attention():
             return True
         if is_ascend_npu():
             return True
+        if is_amd():
+            return True #if you have pytorch attention enabled on AMD it probably supports at least mem efficient attention
     return False
 
 def mac_version():