# google-research/tapas

End-to-end neural table-text understanding models.

Repository: https://github.com/google-research/tapas
Canonical: https://ross.abutalabs.com/products/tapas
Language: Python
License: Apache-2.0
License Family: permissive
Topics: nlp-machine-learning, question-answering, table-parsing, tensorflow
Archived: true
Last push: 2024-07-22T13:51:30+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2346, "days_push": 772, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1202, forks 214 (observed 2026-08-28T04:03:58.489830+00:00)

## What it is
TAPAS is Google Research's implementation of transformer-based models for question answering over tables, including pretrained checkpoints and training code. It also covers related models for table retrieval, open-domain QA over tables, table entailment (TabFact), and efficient table attention variants (MATE, DoT).

## Use cases
- answer natural language questions against tabular data
- parse tables with a neural question answering model
- retrieve relevant tables for open-domain question answering
- classify whether a statement is entailed by a table (TabFact)
- fine-tune a table QA model on a custom dataset
- run efficient transformer attention over large tables

## When to choose
- you need state-of-the-art neural table question answering with pretrained checkpoints
- you are doing research on table-text understanding or table retrieval
- you want TensorFlow-based training code for table parsing models

## When to avoid
- you need a production-ready QA service rather than research code
- you prefer PyTorch or want to use the models via Hugging Face Transformers instead
- you need general document or text QA not grounded in tables

## Facets
- artifact type: library
- maturity: maintenance
- function: nlp, machine-learning, deep-learning, search-engine
- domain: machine-learning, deep-learning
- platform: python
- tags: table-question-answering, table-parsing, transformers, tensorflow, research-code, question-answering, table-retrieval, natural-language-processing, search, linux, gpu

## Member repositories
- google-research/tapas (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:58.489830+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:20:35.988736+00:00, confidence not recorded.
  - readme: https://github.com/google-research/tapas (fetched 2026-08-28T04:03:58.489830+00:00, sha 2d17760c6bab)
- Data as of 2026-08-30T08:39:29.467469+00:00.