how to run calm dataset?

[RyanAIResearch]

I use calm dataset
defaults:

• base

target: lobster.data.CalmLightningDataModule

root: ???
batch_size: 64
num_workers: 1
max_length: ${model.max_length}
tokenizer_dir: ${model.tokenizer_dir}
mlm: True

and use mlm but it has bugs.

and I check the dataset , batch is a list like this:
batch: ['ATGTTTCAAATTGGGGGCCTCATTGTCTTCTGTGGGCTGCTGACCCAGACCACAGCCCTGTTGGACGCCCTGCCCTTGCCCTTGGGTCAGGCCCAACCCTTGCCTCTGACTCCAGCCCTGGCCTCAAGTCCCACAGATCTTGCTGGAAGCCTGACAAGTGCTCTCAGCAATGGCCTGCTCTCTGGGGGTCTGTTGGATAGTCTCACAAACCTTCCACTCTTGGACATTCTGAAGACTGAAGGAGGCAATTCTGGTGGCCTGATTGGTGGCCTGCTTGGGAGAGTGACTTCAGTGATCCCTCTCCTGGACAACATCCTTGATGTGAAAGTCACTAAT']
and there are no tutorials guide me how to use cb model