lance-format/squad-v2-lance

SQuAD v2 (Lance Format)

Lance-formatted version of SQuAD v2 — Stanford Question Answering Dataset, version 2 — with MiniLM sentence embeddings stored inline alongside the questions, contexts, and answers.

Why this version?

• One self-contained Lance dataset with 130k+ Wikipedia-grounded questions and reference answers.
• Pre-computed text embeddings (sentence-transformers/all-MiniLM-L6-v2, 384-dim, L2-normalized) on the question column with an IVF_PQ index — instant semantic question retrieval.
• Full-text inverted indices on both question and context for keyword search.
• BITMAP on is_impossible for fast filtering between answerable and unanswerable questions.

Splits

Split	Rows
train.lance	130,319
validation.lance	11,873

Schema

Column	Type	Notes
id	string	SQuAD question id
title	string	Wikipedia article title
context	string	Paragraph the question was generated from
question	string	The question text
answers	list<string>	Accepted answer spans (empty for impossible questions)
answer_starts	list<int32>	Character offsets of each answer within context
is_impossible	bool	true for SQuAD 2.0 unanswerable questions
question_emb	fixed_size_list<float32, 384>	MiniLM embedding of question (cosine-normalized)

Pre-built indices

• IVF_PQ on question_emb — metric=cosine
• INVERTED on question and context
• BTREE on id and title
• BITMAP on is_impossible

Quick start

import lance

ds = lance.dataset("hf://datasets/lance-format/squad-v2-lance/data/validation.lance")
print(ds.count_rows(), ds.schema.names, ds.list_indices())

Semantic question retrieval

import lance
import pyarrow as pa
from sentence_transformers import SentenceTransformer

encoder = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2", device="cuda")
q_vec = encoder.encode(["what year was the eiffel tower built?"], normalize_embeddings=True)[0]

ds = lance.dataset("hf://datasets/lance-format/squad-v2-lance/data/train.lance")
emb_field = ds.schema.field("question_emb")
query = pa.array([q_vec.tolist()], type=emb_field.type)

hits = ds.scanner(
    nearest={"column": "question_emb", "q": query[0], "k": 10, "nprobes": 16, "refine_factor": 30},
    columns=["id", "title", "question", "answers"],
).to_table().to_pylist()

Full-text search on contexts

ds = lance.dataset("hf://datasets/lance-format/squad-v2-lance/data/train.lance")
hits = ds.scanner(
    full_text_query="great pyramid of giza",
    columns=["title", "question", "context"],
    limit=5,
).to_table().to_pylist()

Filter answerable vs impossible questions

ds = lance.dataset("hf://datasets/lance-format/squad-v2-lance/data/validation.lance")
impossible = ds.scanner(filter="is_impossible = true", columns=["question"], limit=5).to_table()

Why Lance?

• One dataset carries questions + contexts + answers + embeddings + indices — no sidecar files.
• On-disk vector and full-text indices live next to the data, so search works on local copies and on the Hub.
• Schema evolution: add columns (alternate embeddings, model predictions, task labels) without rewriting the data.

Source & license

Converted from rajpurkar/squad_v2. SQuAD v2 is released under CC BY-SA 4.0.

Citation

@article{rajpurkar2018know,
  title={Know What You Don't Know: Unanswerable Questions for SQuAD},
  author={Rajpurkar, Pranav and Jia, Robin and Liang, Percy},
  journal={Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Short Papers)},
  year={2018},
}