Skip to content

Git conflict resolution headers remains in JSONLs #212

@vd-omermazig

Description

@vd-omermazig

Didn't use them yet, but I noticed:

https://github.com/NVIDIA/GenerativeAIExamples/blob/main/nemo/retriever-synthetic-data-generation/scripts/conf/config-nq.yaml
https://github.com/NVIDIA/GenerativeAIExamples/blob/main/nemo/retriever-synthetic-data-generation/scripts/conf/config-fiqa.yaml

Have some git headers that weren't resolved in the main branch:

easiness_filter:
  _target_: nemo_retriever_sdg.EasinessFilter
  filter_cfg:
<<<<<<< HEAD
    filter_threshold: 0.75
    embedding_model: "intfloat/e5-large-unsupervised"
=======
    filter_threshold: 0.95
    embedding_model: "BAAI/bge-large-en-v1.5"
>>>>>>> 3d49ef8 (added config-nq and config-fiqa)
    batch_size: 16

Which will of course make the jsonl unusable

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions