FINDING #4695 · UNIT ID 790916393
NVIDIA/Model-Optimizer
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc. to optimize inference speed.
SURPRISE SCORE
0.00
Score Breakdown
SCORE = SURPRISE × ENGAGEMENT^0.7 × FRESHNESS × VISIBILITY × CONFIDENCE
SURPRISE = WINDOW STARS / DAYS / (AUDIENCE + FLOOR)
1% OF STARS IN ARCHIVE
Growth Telemetry
VELOCITY /D
13.33
ACCEL
-1.50
RETENTION
30.4%
PEAK 2026-10-16 · FORK-RETENTION 12.5% · 40 STARS / WINDOW
Author Audience
AUDIENCE
567,194
FOLLOWERS
31,458
OWNER ★
441,205
Engagement Signals
FORKS
752
ISSUE AUTH
0
PR AUTH
0
UNIQUE STARGAZERS 40 / 40 (DIVERSITY 1.00)
Why This Is A Finding
NVIDIA/Model-Optimizer собрал 40 звёзд за окно, тогда как у автора всего 31,458 подписчиков — эффективная аудитория ≈ 567,194. Это даёт surprise-индекс 0.0000235 (звёзды относительно охвата автора, а не в абсолюте). Удержание форков 12.5% и 0 внешних контрибьюторов отделяют реальный инструмент от разовой вспышки. Акселерация отрицательная — внимание остывает после пика.
Related Findings
RANKS ABOVE 6% OF 4978 FINDINGS
METRICS IN CONTEXT
MEDIAN ACROSS ALL 4978 FINDINGS · Δ vs MEDIAN · PERCENTILE = SHARE RANKED BELOW
METRICVALUEMEDIANΔ MEDPERCENTILE
SCORE0.000.00-0.00ABOVE 6%
VELOCITY13.335.33+8.00ABOVE 78%
RETENTION30.4%0.0%+30.4 PPABOVE 63%
FORKS752346+406ABOVE 64%
SURPRISE0.000.00-0.00ABOVE 2%