# allenai/bi-att-flow

Bi-directional Attention Flow (BiDAF) network is a multi-stage hierarchical process that represents context at different levels of granularity and uses a bi-directional attention flow mechanism to achieve a query-aware context representation without early summarization.

Repository: https://github.com/allenai/bi-att-flow
Canonical: https://ross.abutalabs.com/products/bi-att-flow
Homepage: http://allenai.github.io/bi-att-flow
Language: Python
License: Apache-2.0
License Family: permissive
Topics: squad, nlp, tensorflow, question-answering, bidaf
Last push: 2023-05-31T01:34:05+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 8, longevity 100
- inputs: {"age_days": 3724, "days_push": 1191, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1546, forks 667 (observed 2026-08-28T04:05:01.650665+00:00)

## What it is
Original implementation of the Bi-directional Attention Flow (BiDAF) network for machine comprehension on the SQuAD dataset. It is a multi-stage hierarchical deep learning model built in TensorFlow that produces query-aware context representations for extractive question answering.

## Use cases
- train a question answering model on SQuAD
- reproduce BiDAF results from the paper
- study bi-directional attention flow for reading comprehension
- run extractive QA on a passage given a question
- benchmark machine comprehension models

## When to choose
- you need the reference implementation of BiDAF for research or reproduction
- you want a classic baseline for extractive question answering on SQuAD

## When to avoid
- you need a modern, maintained QA system or transformer-based models
- you use TensorFlow 2.x or recent Python versions, since this targets TensorFlow r0.11
- you want production-ready question answering out of the box

## Facets
- artifact type: library
- maturity: abandoned
- function: machine-learning, nlp
- domain: machine-learning
- platform: python
- tags: question-answering, machine-comprehension, squad, bidaf, tensorflow, attention-mechanism, research-code, natural-language-processing, gpu

## Member repositories
- allenai/bi-att-flow (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:01.650665+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 2026-08-30T04:30:33.100009+00:00, confidence not recorded.
  - readme: https://github.com/allenai/bi-att-flow (fetched 2026-08-28T04:05:01.650665+00:00, sha dc6ccdbe0e9f)
  - homepage: http://allenai.github.io/bi-att-flow (fetched 2026-08-29T11:31:42.806268+00:00, sha cf7c8e336e05)
- Data as of 2026-08-30T08:39:29.467469+00:00.