微调IP_adapter，请问如何生成对应的训练数据集？按照普通的全参微调生成的数据进行微调，报错。

[RobinHan24]

ValueError: group_length must be a dict, but got <class 'list'>
You may using a vanilla Index V2 file. Try ArrowIndexV2 instead.

[zml-ai]

It seems that you didn't create index v2 first and then make multiple buckets.
It's recommended to create the index file for a single resolution first. You can refer to the code like this: idk base -c dataset/yamls/porcelain.yaml -t dataset/porcelain/jsons/porcelain.json. And in the training script, cancel the instruction of --multireso for multiple buckets.