Comprehensive Evaluation of CNN-Based Audio Tagging Models on Resource-Constrained Devices

2025-09-17

AI Summary

Deploying CNN-based audio tagging models on resource-constrained edge devices (e.g., Raspberry Pi) faces challenges including high inference latency, thermal runaway, and poor long-term operational stability. This work systematically evaluates six model families—PANNs (1D/2D), ConvNeXt, MobileNetV3, and two newly designed lightweight architectures (CNN9 and CNN13)—under 24-hour continuous inference, measuring both accuracy and thermal behavior. We introduce, for the first time, a standardized long-term stability evaluation protocol and unify deployment via ONNX to ensure cross-platform compatibility and efficiency. Results show that CNN9 achieves the optimal trade-off among accuracy, latency, and power consumption: its average latency variation remains below 1.2% over 24 hours, with temperature rise stabilized at ≤15°C. ONNX conversion improves deployment portability and delivers up to 1.8× inference speedup. This study provides a reproducible, sustainable, and lightweight deployment framework for edge audio perception.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Learning & Optimization for CV
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Convolutional Neural Networks (CNNs) have demonstrated exceptional performance in audio tagging tasks. However, deploying these models on resource-constrained devices like the Raspberry Pi poses challenges related to computational efficiency and thermal management. In this paper, a comprehensive evaluation of multiple convolutional neural network (CNN) architectures for audio tagging on the Raspberry Pi is conducted, encompassing all 1D and 2D models from the Pretrained Audio Neural Networks (PANNs) framework, a ConvNeXt-based model adapted for audio classification, as well as MobileNetV3 architectures. In addition, two PANNs-derived networks, CNN9 and CNN13, recently proposed, are also evaluated. To enhance deployment efficiency and portability across diverse hardware platforms, all models are converted to the Open Neural Network Exchange (ONNX) format. Unlike previous works that focus on a single model, our analysis encompasses a broader range of architectures and involves continuous 24-hour inference sessions to assess performance stability. Our experiments reveal that, with appropriate model selection and optimization, it is possible to maintain consistent inference latency and manage thermal behavior effectively over extended periods. These findings provide valuable insights for deploying audio tagging models in real-world edge computing scenarios.

Problem

• Evaluating CNN audio tagging models on Raspberry Pi
• Assessing computational efficiency and thermal management challenges
• Testing multiple architectures for long-term performance stability

Innovation

• Evaluated multiple CNN architectures on Raspberry Pi
• Converted models to ONNX format for portability
• Conducted 24-hour inference for stability assessment