Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LLM Inference Scheduling on Alpaca

0.1576P90 Per-token Latency (s/token)

Oracle SJF

-0.2904862.734165.7588058.78345Sep 25, 2025Oct 26, 2025Nov 26, 2025Dec 28, 2025Jan 28, 2026Feb 28, 2026Apr 1, 2026
Updated 27d ago

Evaluation Results

MethodLinks
2025.09
0.1576---
2025.09
0.18---
2025.09
0.208---
2025.09
0.2155---
2025.09
0.2298---
2025.09
0.2306---
2025.09
0.2581---
2025.09
0.3334---
2025.09
0.3868---
2025.09
0.7052---
2025.09
0.8787---
2026.04
0.930.5254.18123.02
2026.04
1.470.7162.93140.8
2026.04
1.760.8365.33142.17
2025.09
2.2257---
2026.04
2.811.54146.14359.08
2026.04
3.621.4583.65159.72
2026.04
4.322.06171.14395.78
2026.04
5.142.36174.31394.1
2026.04
11.364.52235.64444.52