Datasets:
metadata
license: afl-3.0
language:
- en
- de
pretty_name: >-
Data Mix - DIRTY (uncleaned - input and Instruction is wrong, contains only
input and output field)
size_categories:
- 10M<n<100M
this dataset is a mix of of some datasets available on huggingface for text-to-text generation, it's not cleaned, it's not deduplicated.
It's not recommend to use it in it's current state.
as the higghest restriction counts, it's only for private and research usable - it's not allowed to create an competitive model to "whatever". Never use it for commercial purposes describes it for all Datasets in this repository.