feifeibear/long-context-attention
Fine-tuning ToolsUSP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference
No dedicated docs site. Description: 119 chars. Stars signal: 672. Contributors: 24. Score: 4.7/10
Stars: 672. Contributors: 24. Watchers: 5. Forks: 79. Issue ratio: 1.8%. Score: 4.9/10
Last commit: 13d ago. Weekly commits: 0. Latest release: 0.6.4. Maturity bonus: 2.2y old. Score: 6.8/10
Stars/issues ratio: 56. Dynamic language: Python. No dedicated API docs. Permissive license: Apache-2.0. Popularity signal: 672 stars. Score: 6.7/10
Battle-tested: 672 stars. Peer review: 24 contributors. Versioned: 0.6.4. Licensed: Apache-2.0. Age: 2.2 years. Maintenance: last commit 13d ago. Score: 6.3/10
Fork interest: 79. Major ecosystem: Python. Integration-friendly: Apache-2.0. Adoption: 672 stars. Score: 6.5/10