Fix the bug that joint_attention_kwargs is not passed to the FLUX's transformer attention processors
#9517
+8
−2