Towards Optimal Trade-offs in Knowledge Distillation for CNNs and Vision Transformers at the Edge
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916328698281984 |
|---|---|
| author | Violos, John Papadopoulos, Symeon Kompatsiaris, Ioannis |
| author_facet | Violos, John Papadopoulos, Symeon Kompatsiaris, Ioannis |
| contents | This paper discusses four facets of the Knowledge Distillation (KD) process for Convolutional Neural Networks (CNNs) and Vision Transformer (ViT) architectures, particularly when executed on edge devices with constrained processing capabilities. First, we conduct a comparative analysis of the KD process between CNNs and ViT architectures, aiming to elucidate the feasibility and efficacy of employing different architectural configurations for the teacher and student, while assessing their performance and efficiency. Second, we explore the impact of varying the size of the student model on accuracy and inference speed, while maintaining a constant KD duration. Third, we examine the effects of employing higher resolution images on the accuracy, memory footprint and computational workload. Last, we examine the performance improvements obtained by fine-tuning the student model after KD to specific downstream tasks. Through empirical evaluations and analyses, this research provides AI practitioners with insights into optimal strategies for maximizing the effectiveness of the KD process on edge devices. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_12808 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Towards Optimal Trade-offs in Knowledge Distillation for CNNs and Vision Transformers at the Edge Violos, John Papadopoulos, Symeon Kompatsiaris, Ioannis Computer Vision and Pattern Recognition Artificial Intelligence This paper discusses four facets of the Knowledge Distillation (KD) process for Convolutional Neural Networks (CNNs) and Vision Transformer (ViT) architectures, particularly when executed on edge devices with constrained processing capabilities. First, we conduct a comparative analysis of the KD process between CNNs and ViT architectures, aiming to elucidate the feasibility and efficacy of employing different architectural configurations for the teacher and student, while assessing their performance and efficiency. Second, we explore the impact of varying the size of the student model on accuracy and inference speed, while maintaining a constant KD duration. Third, we examine the effects of employing higher resolution images on the accuracy, memory footprint and computational workload. Last, we examine the performance improvements obtained by fine-tuning the student model after KD to specific downstream tasks. Through empirical evaluations and analyses, this research provides AI practitioners with insights into optimal strategies for maximizing the effectiveness of the KD process on edge devices. |
| title | Towards Optimal Trade-offs in Knowledge Distillation for CNNs and Vision Transformers at the Edge |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2407.12808 |