STACKQUADRANT

feifeibear/long-context-attention

Fine-tuning Tools

USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference

5.7
GitHub Metrics
Stars
696
Forks
87
Open Issues
15
Watchers
5
Contributors
27
Weekly Commits
0
Language
Python
License
Apache-2.0
Last Commit
Sep 16, 2026
Created
Mar 27, 2024
Latest Release
0.6.4
Release Date
Jan 13, 2026
Synced: Oct 17, 2026
Quality Scores
Documentation Qualityw: 20%
4.8

No dedicated docs site. Description: 119 chars. Stars signal: 696. Contributors: 27. Score: 4.8/10

Community Healthw: 20%
5.1

Stars: 696. Contributors: 27. Watchers: 5. Forks: 87. Issue ratio: 2.2%. Score: 5.1/10

Maintenance Velocityw: 15%
5.6

Last commit: 31d ago. Weekly commits: 0. Latest release: 0.6.4. Maturity bonus: 2.6y old. Score: 5.6/10

API Design & DXw: 20%
6.4

Stars/issues ratio: 46. Dynamic language: Python. No dedicated API docs. Permissive license: Apache-2.0. Popularity signal: 696 stars. Score: 6.4/10

Production Readinessw: 15%
6.2

Battle-tested: 696 stars. Peer review: 27 contributors. Versioned: 0.6.4. Licensed: Apache-2.0. Age: 2.6 years. Maintenance: last commit 31d ago. Score: 6.2/10

Ecosystem Integrationw: 10%
6.6

Fork interest: 87. Major ecosystem: Python. Integration-friendly: Apache-2.0. Adoption: 696 stars. Score: 6.6/10

Tags
attention-is-all-you-needdeepspeed-ulyssesllm-inferencellm-trainingpytorchring-attention
Radar
Documentation Quality
Community Health
Maintenance Velocity
API Design & DX
Production Readiness
Ecosystem Integration