Trendoday

AutoCompress: Automated Hardware-Guided Model Compression

LLM pruning must preserve quality while meeting hardware- and workload-specific constraints on latency, throughput, and memory. Existing pruning workflows often target parameter counts, requiring prac

  • Upvotes6
–Rank now
#17Peak rank
2Days on list
128Score · 7d
Nov 9, 2026First seen
Rank · 7 days

About this paper

AutoCompress is a hardware-guided compression framework that uses differentiable architecture search and serving-engine cost models to produce heterogeneous LLM architectures meeting latency, throughput, and memory constraints while preserving accuracy.

LLM pruning must preserve quality while meeting hardware- and workload-specific constraints on latency, throughput, and memory. Existing pruning workflows often target parameter counts, requiring practitioners to manually select model sizes that meet latency, throughput, and GPU memory limits. However, parameter count alone does not predict these metrics, which depend on the model architecture, workload, hardware, and serving engine. We introduce AutoCompress, a hardware-guided compression framework that directly incorporates such constraints into differentiable architecture search. AutoCompress jointly optimizes global and per-layer structure to produce heterogeneous architectures, guided…

LLM pruninghardware-guided compressiondifferentiable architecture searchheterogeneous architecturesglobal and per-layer structureserving-engine cost modelsend-to-end latencythroughputGPU memory limitslong-context local coding
Upvotes
6
Authors
7
Published
Nov 5, 2026

Summary and abstract excerpt via Hugging Face Papers ↗ · arXiv.

Trending history

AutoCompress: Automated Hardware-Guided Model Compression was first seen on Hugging Face · Daily papers on Nov 9, 2026. Since then it has appeared in 16 checks on 2 different days. Its best position was #17 of 30, reached on Nov 10, 15:28 UTC. It dropped off the list after Nov 10, 19:28 UTC.

Rank history

Position on Hugging Face · Daily papers at every check (lower is better).

#1#11#20#30 13:2820:5804:2811:5819:28 Nov 9, 13:28 UTC · #27Nov 9, 15:28 UTC · #25Nov 9, 17:28 UTC · #23Nov 9, 19:28 UTC · #25Nov 9, 21:28 UTC · #25Nov 9, 23:28 UTC · #26Nov 10, 01:28 UTC · #24Nov 10, 03:28 UTC · #24Nov 10, 05:28 UTC · #23Nov 10, 07:28 UTC · #20Nov 10, 09:28 UTC · #23Nov 10, 11:28 UTC · #20Nov 10, 13:28 UTC · #18Nov 10, 15:28 UTC · #17Nov 10, 17:28 UTC · #18Nov 10, 19:28 UTC · #17

Last 14 days

Best rank per day (UTC).

Nov 3today

Metric history

Platform metric as reported by Hugging Face.

04711 13:2820:5804:2811:5819:28 Nov 9, 13:28 UTC · 6Nov 9, 15:28 UTC · 7Nov 9, 17:28 UTC · 8Nov 9, 19:28 UTC · 8Nov 9, 21:28 UTC · 8Nov 9, 23:28 UTC · 8Nov 10, 01:28 UTC · 8Nov 10, 03:28 UTC · 8Nov 10, 05:28 UTC · 9Nov 10, 07:28 UTC · 9Nov 10, 09:28 UTC · 9Nov 10, 11:28 UTC · 9Nov 10, 13:28 UTC · 9Nov 10, 15:28 UTC · 9Nov 10, 17:28 UTC · 9Nov 10, 19:28 UTC · 11

Questions

Is AutoCompress: Automated Hardware-Guided Model Compression trending on HF Daily Papers right now?

Not at the moment. It was last on the list on Nov 10, 19:28 UTC, after 2 days of trending.

What is the highest rank AutoCompress: Automated Hardware-Guided Model Compression reached?

#17 on Hugging Face · Daily papers, on Nov 10, 15:28 UTC. Trendoday checks this list every 120 minutes.

How long has AutoCompress: Automated Hardware-Guided Model Compression been trending?

It was first seen on Nov 9, 2026 and has been on the list on 2 days (16 checks) since.