When using sdxl_train_network.py to train Lora, an error occurs. AttributeError: 'list' object has no attribute 'named_modules'.

Traceback (most recent call last):
File "D:\PycharmProjects\sd-scripts-sdxl\sdxl_train_network.py", line 189, in
trainer.train(args)
File "D:\PycharmProjects\sd-scripts-sdxl\train_network.py", line 290, in train
network = network_module.create_network(
File "D:\PycharmProjects\sd-scripts-sdxl\networks\dylora.py", line 190, in create_network
network = DyLoRANetwork(
File "D:\PycharmProjects\sd-scripts-sdxl\networks\dylora.py", line 310, in __init__
self.text_encoder_loras = create_modules(False, text_encoder, DyLoRANetwork.TEXT_ENCODER_TARGET_REPLACE_MODULE)
File "D:\PycharmProjects\sd-scripts-sdxl\networks\dylora.py", line 280, in create_modules
for name, module in root_module.named_modules():
AttributeError: 'list' object has no attribute 'named_modules'

When I debugged, I found that the Text encoder was a list containing two elements in sdxl_train_network.py. I think this is the cause of the error in training. Can anyone tell me how to solve this problem?

Start by reproducing the traceback with sdxl_train_network.py, then read the text-encoder setup around the reported entry point and networks/dylora.py:create_modules and DyLoRANetwork.__init__. Trace how the two text encoders are passed into create_modules and confirm the expected input shape. Done means SDXL LoRA training proceeds past this initialization without the named_modules error.

Written by the indexing model from the issue text.