# kimiyoung/transformer-xl

Repository: https://github.com/kimiyoung/transformer-xl
Canonical: https://ross.abutalabs.com/products/transformer-xl
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2022-09-21T06:22:01+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 2794, "days_push": 1442, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3714, forks 765 (observed 2026-08-28T04:08:15.306573+00:00)

## What it is
Official implementation of Transformer-XL, an attention-based language model architecture that extends context beyond a fixed length via segment-level recurrence. Includes PyTorch and TensorFlow code, training scripts, and pretrained models achieving state-of-the-art language modeling results.

## Use cases
- train a language model with long context beyond fixed length
- download and fine-tune pretrained Transformer-XL models
- reproduce state-of-the-art results on enwiki8 text8 and WikiText-103
- run char-level language modeling on PTB
- study segment-level recurrence attention implementation
- train language models on multiple GPUs or TPUs

## When to choose
- you need the original Transformer-XL architecture code for research or reproduction
- you want pretrained models for language modeling benchmarks
- you need both PyTorch and TensorFlow implementations

## When to avoid
- you want a maintained library integrated into modern frameworks - use Hugging Face Transformers instead
- you need production LLM inference or serving
- you need recent model architectures or active community support

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, llm-training, transformers
- domain: deep-learning, large-language-models
- platform: python
- tags: transformer-xl, language-modeling, pytorch, tensorflow, pretrained-models, research-code, natural-language-processing, gpu

## Member repositories
- kimiyoung/transformer-xl (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:15.306573+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 2026-08-29T18:29:23.813301+00:00, confidence not recorded.
  - readme: https://github.com/kimiyoung/transformer-xl (fetched 2026-08-28T04:08:15.306573+00:00, sha e858ccf15c8a)
- Data as of 2026-08-30T08:39:29.467469+00:00.