Dataset type conversion utilities

[qgallouedec]

System Info

Some things that are not really correct in the dataset type conversions in https://huggingface.co/docs/trl/main/en/dataset_formats#utilities-for-converting-dataset-types:

1. when converting a preference dataset to an unpaired preference dataset with unpair_preference_dataset(), we are converting from a relative ranking to an absolute ranking. In a preference dataset, despite having a "chosen" and a "rejected" example, both can be good or both bad, just one slightly better/worse. See the example below.
   So one should not convert a Preference dataset to an Unpaired Preference Dataset without keeping an eye on absolute ratings from e.g. a reward model.
   Suggestion: At least add a warning to the documentation and conversion code or even remove it
2. when converting from Unpaired preference or Stepwise supervision to anything un-labeled like Language modeling or Prompt-completion, only the good (label=True) examples should be used. Like when converting from a Preference dataset it only uses the chosen completions.
   Suggestion: Can easily fix that in the example conversion code

Information

• The official example scripts
• My own modified scripts

Tasks

• An officially supported task in the examples folder
• My own task or dataset (give details below)

Reproduction

from datasets import Dataset
dataset_dict = {
    "prompt": ["The sky is", "The sun is"]
    "chosen": [" blue.", " in our solar system"],
    "rejected": [" above.", " in the sky."]
}
dataset = Dataset.from_dict(dataset_dict)
dataset = unpair_preference_dataset(dataset)
dataset[1]

outputs:

e.g.

{'prompt': 'The sky is', 'completion': ' above.', 'label': False}

Expected behavior

{'prompt': 'The sky is', 'completion': ' blue.', 'label': True}
{'prompt': 'The sky is', 'completion': ' above.', 'label': True}

Checklist

• I have checked that my issue isn't already filed (see open issues)
• I have included my system information
• Any code provided is minimal, complete, and reproducible (more on MREs)
• Any code provided is properly formatted in code blocks, (no screenshot, more on code blocks)
• Any traceback provided is complete

[August-murr]

oops
didn't realize this fork doesn't have the same labels as TRL.
I'll add it and we'll try again