optimizing-attention-flash

github:zechenzhangAGI/claude-ai-research-skills · first committed 2025-11-07 · source · badge

Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.

Skill Health: not yet scored

Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.

Version history

6 commits to this file on GitHub (newest first).

2025-11-20 · commit 0a9ded4 · Standardize skill metadata: Add category tags and update authors
2025-11-20 · commit 72c8b2c · Trigger upload of remaining 26 skills
2025-11-20 · commit 51fc3a3 · Trigger upload of remaining 32 skills batch 2
2025-11-20 · commit 81c38fe · Upload all remaining 34 skills to Orchestra marketplace
2025-11-19 · commit bd4461f · Standardize YAML frontmatter for all 54 skills
2025-11-07 · commit b590f1f · Add Flash Attention and bitsandbytes optimization skills

Archived snapshots