Silent `adaln` drop while anima lora training

Hello, recently I found adaln modules are not included in target lora module list due to this:

exclude_patterns.append(r".*(_modulation|_norm|_embedder|final_layer).*")

in networks/lora_anima.py:254

Was this intended? I guess other lora training scripts involve adaln module in target module list, or at least there should be a knob asking for whether to exclude or not.