increase the min lenghts for SDNQ atten

modules/attention.py

@@ -24,7 +24,7 @@ def set_sdnq_attention():
         sdpa_pre_sdnq_atten = torch.nn.functional.scaled_dot_product_attention
         @wraps(sdpa_pre_sdnq_atten)
         def sdpa_sdnq_atten(query: torch.FloatTensor, key: torch.FloatTensor, value: torch.FloatTensor, attn_mask: torch.Tensor | None = None, dropout_p: float = 0.0, is_causal: bool = False, scale: float | None = None, enable_gqa: bool = False, **kwargs) -> torch.FloatTensor:
-            if not is_causal and query.shape[-3] > 1: # VAE
+            if not is_causal and query.shape[-1] >= 32 and query.shape[-2] >= 256 and query.shape[-3] > 1:
                 return sdnq_triton_atten(
                     query=query, key=key, value=value,
                     attn_mask=attn_mask, scale=scale, enable_gqa=enable_gqa,