Avoid using flashinfer_allreduce_fusion when dp attention is enabled. (#11632)

This commit is contained in:
Elfie Guo
2025-10-26 12:31:14 -07:00
committed by GitHub
parent a8023891f6
commit a1d5bc4cce

View File

@@ -337,6 +337,7 @@ class LayerCommunicator:
static_conditions_met = (
(not self.is_last_layer)
and (self._context.tp_size > 1)
and not is_dp_attention_enabled()
and get_global_server_args().enable_flashinfer_allreduce_fusion
and _is_flashinfer_available
)