carabiner-tools[mpl,pd]>=0.0.4
datasets>=3.1.0
openpyxl==3.1.0
pandas
schemist>=0.0.4.post1
tensorboard
tokenizers
transformers[torch]
umap-learn
