Skip to content

Question about SFT data? #28

@hxdtest

Description

@hxdtest

According to https://huggingface.co/datasets/SimpleBerry/OpenLongCoT-SFT, there are five columns of the datasets, problem, response, ground truth, query and reasoning, how do you organize it for finetuning training?

For example, what dose <start_of_father_id>-1<end_of_father_id><start_of_local_id>0<end_of_local_id><start_of_thought> implies?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions