AniFileBERT / datasets

Commit History

  1. fix(dataset): repair training-set alignment + relabel teneighty as GROUP; add Rust repair tool

    d8ddb0b

  1. Close schema v2 anime parser release

    67d8016

  1. Update AnimeName submodule pointer

    0fafc62

  1. Improve DMHY template labeling pipeline

    59042eb

  1. Add DMHY prefix graph annotation workflow

    33bb11c
  2. Export DMHY metadata list

    4248c69
  3. Add DMHY digit-dedup list export

    beb7665
  4. Update relabel artifacts and pathfix export

    5af16a8
  5. chore: update AnimeName submodule after relabel chunks

    a4c7b01
  6. chore: update AnimeName submodule after relabel chunks

    d535b94
  7. chore: update AnimeName submodule after relabel chunks

    7bb297e
  8. chore: update AnimeName submodule after relabel chunks

    0c430e6
  9. chore: update AnimeName submodule after relabel chunks

    832202f
  10. chore: update AnimeName submodule after relabel chunks

    a95994b
  11. chore: update AnimeName submodule after relabel chunks

    ac82693
  12. chore: update AnimeName submodule after relabel chunks

    7f2cd6b
  13. chore: update AnimeName submodule after relabel chunks

    29aea48
  14. chore: update AnimeName submodule after relabel chunks

    feea690

  1. chore: update AnimeName submodule after relabel chunks

    edb03a6
  2. chore: update AnimeName submodule after relabel chunks

    3dd0093
  3. chore: update AnimeName submodule after relabel chunks

    e7fc63a
  4. chore: update AnimeName submodule after relabel chunks

    87c7b59
  5. Advance AnimeName relabel submodule after additional batches

    fb5c9c2
  6. Advance AnimeName relabel submodule after next batches

    aaa7735
  7. Advance AnimeName relabel submodule batches

    ac7e795
  8. Add robust LLM relabel pipeline and enforce contiguous title

    fed9d99
  9. feat: add concurrency and token/cost telemetry for llm relabel runs

    7c5b41e
  10. feat: add tool-call based llm relabel pipeline and update dataset pointer

    e964ae5
  11. chore: update AnimeName submodule checkpoint

    1804249

  1. Train thin anime filename parser

    f712f4b

  1. Clean special code parsing

    e8412e3
  2. Fix GM-Team bilingual title parsing

    adf92db
  3. Improve anime filename parser model

    e63569d

  1. Add character-token DMHY training path

    0779202
  2. 完成整个数据集的整理

    f4f4e0e

  1. Add AniFileBERT model and training project⁠

    be5f706