LANA: Latency Aware Network Acceleration

TL;DR

LANA is a fast, efficient neural network acceleration method that uses neural architecture search and optimization to produce models meeting specific latency constraints with improved accuracy and speed.

Contribution

LANA introduces a novel constrained ILP-based NAS approach that efficiently searches massive spaces and satisfies latency budgets, outperforming prior methods.

Findings

Achieves up to 3% accuracy improvement on compressed models.

Provides up to 5x speed-up with minimal accuracy loss.

Supports large search spaces and fast NAS within minutes.

Abstract

We introduce latency-aware network acceleration (LANA) - an approach that builds on neural architecture search techniques and teacher-student distillation to accelerate neural networks. LANA consists of two phases: in the first phase, it trains many alternative operations for every layer of the teacher network using layer-wise feature map distillation. In the second phase, it solves the combinatorial selection of efficient operations using a novel constrained integer linear optimization (ILP) approach. ILP brings unique properties as it (i) performs NAS within a few seconds to minutes, (ii) easily satisfies budget constraints, (iii) works on the layer-granularity, (iv) supports a huge search space $O(10^{100})$, surpassing prior search approaches in efficacy and efficiency. In extensive experiments, we show that LANA yields efficient and accurate models constrained by a target latency budget, while being significantly faster than other techniques. We analyze three popular network architectures: EfficientNetV1, EfficientNetV2 and ResNeST, and achieve accuracy improvement for all models (up to $3.0\%$) when compressing larger models to the latency level of smaller models. LANA achieves significant speed-ups (up to $5\times$) with minor to no accuracy drop on GPU and CPU. The code will be shared soon.