mjun0812/flash-attention-prebuild-wheels — GitHub trend

◂ BACK TO LEADERBOARD

FINDING #2802 · UNIT ID 878958395

mjun0812/flash-attention-prebuild-wheels

Provide with pre-build flash-attention 2 and 3 package wheels on Linux and Windows using GitHub Actions

[ PYTHON ][ GITHUB ↗ ]

SURPRISE SCORE

0.00

Score Breakdown

SURPRISE0.00987

ENGAGEMENT0.18

FRESHNESS1.38

SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE

SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)

5% OF STARS IN ARCHIVE

Growth Telemetry

VELOCITY /D

2.77

ACCEL

+0.00

RETENTION

39.6%

PEAK 2026-07-08 · FORK-RETENTION 0.0% · 83 STARS / WINDOW

Author Audience

AUDIENCE

240

FOLLOWERS

OWNER ★

1,704

Engagement Signals

FORKS

ISSUE AUTH

PR AUTH

UNIQUE STARGAZERS 83 / 83 (DIVERSITY 1.00)

Why This Is A Finding

mjun0812/flash-attention-prebuild-wheels собрал 83 звёзд за окно, тогда как у автора всего 70 подписчиков — эффективная аудитория ≈ 240. Это даёт surprise-индекс 0.00987 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 0.0% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация положительная — рост ещё не выдохся.

Related Findings

RANKS ABOVE 60% OF 6938 FINDINGS

#02shy3130/tickflow-stock-panel[ PYTHON ]0.45 #07Panniantong/Agent-Reach[ PYTHON ]0.28 #10sums001/Windows-Copilot-API[ PYTHON ]0.27 #12HUANGCHIHHUNGLeo/claude-real-video[ PYTHON ]0.24 #19MDX-Tom/gpt-5.6-instruct[ PYTHON ]0.20

METRICS IN CONTEXT

MEDIAN ACROSS ALL 6938 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW

METRICVALUEMEDIANΔ MEDPERCENTILE

SCORE0.000.00+0.00ABOVE 60%

VELOCITY2.774.13-1.37ABOVE 34%

RETENTION39.6%28.1%+11.5 PPABOVE 71%

FORKS7389-16ABOVE 45%

SURPRISE0.010.01-0.00ABOVE 46%