Imagine-Data / README.md
Xnhyacinth's picture
Update removed-endpoint note: NQ/TQA research copies merged into this repository
175a415 verified
|
Raw History Blame Contribute Delete
15.4 kB
metadata
license: other
dataset_info:
  - config_name: NQ
    features:
      - name: id
        dtype: int64
      - name: question
        dtype: string
      - name: answers
        sequence: string
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
      - name: compressed_ctxs_1
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_5
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_10
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_20
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_50
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_100
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
    splits:
      - name: train
        num_bytes: 6106425112
        num_examples: 79168
      - name: eval
        num_bytes: 675422872
        num_examples: 8757
      - name: test
        num_bytes: 279441134
        num_examples: 3610
    download_size: 3923770995
    dataset_size: 7061289118
  - config_name: TQA
    features:
      - name: id
        dtype: int64
      - name: question
        dtype: string
      - name: answers
        sequence: string
      - name: target
        dtype: string
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
      - name: compressed_ctxs_1
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_5
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_10
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_20
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_50
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_100
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
    splits:
      - name: train
        num_bytes: 6116068815
        num_examples: 78785
      - name: eval
        num_bytes: 685921423
        num_examples: 8837
      - name: test
        num_bytes: 878592842
        num_examples: 11313
    download_size: 4430841051
    dataset_size: 7680583080
  - config_name: WQ
    features:
      - name: id
        dtype: string
      - name: question
        dtype: string
      - name: answers
        sequence: string
      - name: ctxs
        list:
          - name: hasanswer
            dtype: bool
          - name: id
            dtype: string
          - name: score
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
      - name: compressed_ctxs_5
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_10
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_20
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_50
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_100
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
    splits:
      - name: train
        num_bytes: 268644771
        num_examples: 3478
      - name: eval
        num_bytes: 23143123
        num_examples: 300
      - name: test
        num_bytes: 157146761
        num_examples: 2032
    download_size: 252874314
    dataset_size: 448934655
  - config_name: ctxs1
    features:
      - name: id
        dtype: int64
      - name: answers
        sequence: string
      - name: question
        dtype: string
      - name: compressed_prompt
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
    splits:
      - name: train
        num_bytes: 5212376970
        num_examples: 79168
      - name: eval
        num_bytes: 576466670
        num_examples: 8757
      - name: test
        num_bytes: 238448436
        num_examples: 3610
    download_size: 3327088129
    dataset_size: 6027292076
  - config_name: ctxs100
    features:
      - name: question
        dtype: string
      - name: compressed_prompt
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: answers
        sequence: string
      - name: id
        dtype: int64
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
    splits:
      - name: train
        num_bytes: 5316136567
        num_examples: 79168
      - name: eval
        num_bytes: 587931406
        num_examples: 8757
      - name: test
        num_bytes: 243224578
        num_examples: 3610
    download_size: 3406545161
    dataset_size: 6147292551
  - config_name: ctxs5
    features:
      - name: id
        dtype: int64
      - name: answers
        sequence: string
      - name: question
        dtype: string
      - name: compressed_prompt
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: score
            dtype: float64
          - name: text
            dtype: string
          - name: title
            dtype: string
    splits:
      - name: train
        num_bytes: 5379479670
        num_examples: 79168
      - name: eval
        num_bytes: 594986589
        num_examples: 8757
      - name: test
        num_bytes: 246104192
        num_examples: 3610
    download_size: 3401123245
    dataset_size: 6220570451
  - features:
      - name: id
        dtype: int64
      - name: question
        dtype: string
      - name: answers
        sequence: string
      - name: target
        dtype: string
      - name: ctxs
        list:
          - name: id
            dtype: string
          - name: text
            dtype: string
          - name: title
            dtype: string
      - name: compressed_ctxs_1
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
      - name: compressed_ctxs_100
        struct:
          - name: compressed_prompt
            dtype: string
          - name: compressed_tokens
            dtype: int64
          - name: origin_tokens
            dtype: int64
          - name: ratio
            dtype: string
          - name: saving
            dtype: string
    splits:
      - name: train
        num_bytes: 5395421489
        num_examples: 78785
      - name: eval
        num_bytes: 605118800
        num_examples: 8837
      - name: test
        num_bytes: 775128252
        num_examples: 11313
    download_size: 3905842508
    dataset_size: 6775668541
    config_name: tqa-raw
configs:
  - config_name: NQ
    data_files:
      - split: train
        path: NQ/train-*
      - split: eval
        path: NQ/eval-*
      - split: test
        path: NQ/test-*
  - config_name: TQA
    data_files:
      - split: train
        path: TQA/train-*
      - split: eval
        path: TQA/eval-*
      - split: test
        path: TQA/test-*
  - config_name: WQ
    data_files:
      - split: train
        path: WQ/train-*
      - split: eval
        path: WQ/eval-*
      - split: test
        path: WQ/test-*
  - config_name: ctxs1
    data_files:
      - split: train
        path: ctxs1/train-*
      - split: eval
        path: ctxs1/eval-*
      - split: test
        path: ctxs1/test-*
  - config_name: ctxs5
    data_files:
      - split: train
        path: ctxs5/train-*
      - split: eval
        path: ctxs5/eval-*
      - split: test
        path: ctxs5/test-*
  - config_name: ctxs100
    data_files:
      - split: train
        path: ctxs100/train-*
      - split: eval
        path: ctxs100/eval-*
      - split: test
        path: ctxs100/test-*
  - config_name: tqa-raw
    data_files:
      - split: train
        path: tqa-raw/train-*
      - split: eval
        path: tqa-raw/eval-*
      - split: test
        path: tqa-raw/test-*

Imagine

Question-answering tables for the Imagine / IAG-AAG research line, in one repository: the primary NQ, TQA, WQ views, the Natural-Questions context-count arms (ctxs1/ctxs5/ctxs100), and the original TQA splits (tqa-raw).

Subsets

Directory Config / data_dir Content Purpose
NQ/ NQ NQ question/answer, ctxs, compressed_ctxs_1..100 primary NQ view
TQA/ TQA TQA question/answer/target, ctxs, compressed_ctxs_1..100 primary TQA view
WQ/ WQ WQ question/answer, ctxs, compressed_ctxs_5..100 primary WQ view
ctxs1/ ctxs1 NQ with a single retrieved context NQ context-count arm (k=1)
ctxs5/ ctxs5 NQ with 5 retrieved contexts NQ context-count arm (k=5)
ctxs100/ ctxs100 NQ with 100 retrieved contexts NQ context-count arm (k=100)
tqa-raw/ tqa-raw original TQA splits as first published (the source repo's data/) raw TQA table

Every subset has train, eval and test splits.

from datasets import load_dataset

# primary views
nq  = load_dataset("Xnhyacinth/Imagine-Data", "NQ",  split="test")
tqa = load_dataset("Xnhyacinth/Imagine-Data", "TQA", split="test")
wq  = load_dataset("Xnhyacinth/Imagine-Data", "WQ",  split="test")

# Natural-Questions context-count arms
c1   = load_dataset("Xnhyacinth/Imagine-Data", "ctxs1",   split="test")
c5   = load_dataset("Xnhyacinth/Imagine-Data", "ctxs5",   split="test")
c100 = load_dataset("Xnhyacinth/Imagine-Data", "ctxs100", split="test")

# original TQA splits
raw = load_dataset("Xnhyacinth/Imagine-Data", "tqa-raw", split="test")

The NQ/TQA/WQ views were previously also published as separate endpoints (Xnhyacinth/NQ-Image-Research-Data, Xnhyacinth/TQA-Image-Research-Data); both were merged into this repository and removed on 2026-10-13.

Merged subsets: provenance and verification

ctxs1/, ctxs5/ and ctxs100/ were merged from Xnhyacinth/Imagine-NQ-Contexts at revision c8428a340444086df850693ee704938dd230db2c; tqa-raw/ was merged from Xnhyacinth/Imagine-TQA at revision 3ea2dc07bba59c7d05a7e65f3d5cb1e8958b02ba. Files were copied server-side on the Hub and every one of the 57 parquet files was verified byte-identical to its source by size, git blob_id and LFS sha256 before the source repositories were removed. Paths are unchanged except data/ -> tqa-raw/, and the original source cards are reproduced below.

ARTIFACTS-ctxs-nq.json and ARTIFACTS-tqa-raw.json retain the merged subsets' own reviewed-copy records (source revision, whole-file before/after hashes, substitution counts).

Reviewed research copy

This complete file-set derivative of Xnhyacinth/Image at 863f15e0aaf57edada8ef3d3ff1b69a8427d426c applies reviewed substitutions to selected privacy-relevant string spans. Original configurations, splits, schemas, record IDs and all other fields are retained. Formal references, source attribution and documented examples remain subject to the original source terms; this card introduces no new license.

The ctxs1/ctxs5/ctxs100 subsets are a complete file-set derivative of Xnhyacinth/NQ-Image at c77aa8aaf052e9a779884faa34c60e86733da0fc with the same kind of reviewed substitutions (6 changed string fields across 6 files). The tqa-raw subset is the same kind of derivative of Xnhyacinth/TQA-Image at 60a12d654d748ea7ee6f7e52ce08bd2909e53bf8 (4 changed string fields).

Historical token counts, compression ratios, scores and other measured numeric fields are retained as source observations and have not been recomputed after text edits. The edits do not certify unchanged benchmark performance. Dataset byte metadata changes only by the edited strings' UTF-8 size delta and the changed files' stored-size delta.

Review covers native text and supported static metadata. It does not include OCR of image pixels or analysis of model tensor contents. No source code, models, inference or experiments were executed to produce this copy. Some redacted links require navigation from the retained formal source page.

ARTIFACTS.json records source revision, whole-file before/after hashes and substitution counts for the NQ/TQA/WQ views. It excludes literal private values, individual-value fingerprints, record offsets and review contexts. Original histories are retained in the restricted source repositories.

License

Contains Natural Questions data (upstream CC-BY-SA-3.0, attribution required). Per-source terms apply elsewhere; see the ARTIFACTS/README sections above.