Pith. sign in

REVIEW 1 cited by

Relax: Composable Abstractions for End-to-End Dynamic Machine Learning

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2311.02103 v2 pith:54TZC6SA submitted 2023-11-01 cs.LG cs.AIcs.PL

Relax: Composable Abstractions for End-to-End Dynamic Machine Learning

classification cs.LG cs.AIcs.PL
keywords dynamicrelaxmodelsshapeacrossemergingend-to-endlearning
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Dynamic shape computations have become critical in modern machine learning workloads, especially in emerging large language models. The success of these models has driven the demand for their universal deployment across a diverse set of backend environments. In this paper, we present Relax, a compiler abstraction for optimizing end-to-end dynamic machine learning workloads. Relax introduces a cross-level abstraction that encapsulates computational graphs, loop-level tensor programs, and external library calls in a single representation. Relax also introduces first-class symbolic shape annotations to track dynamic shape computations globally across the program, enabling dynamic shape-aware cross-level optimizations. We build an end-to-end compilation framework using the proposed approach to optimize dynamic shape models. Experimental results on LLMs show that Relax delivers performance competitive with state-of-the-art systems across various GPUs and enables deployment of emerging models to a broader set of emerging environments, including mobile phones, embedded devices, and web browsers.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. ODMA: On-Demand Memory Allocation Strategy for LLM Serving on LPDDR-Class Accelerators

    cs.AR 2025-12 unverdicted novelty 5.0

    ODMA raises KV-cache utilization by up to 19.25% and throughput by 23-27% on Cambricon MLU accelerators by dynamically adjusting prediction buckets and using a safety pool for LLM serving.