Leveraging Temporal Patterns: Automated Augmentation to Create Temporal Early Exit Networks for Efficient Edge AI
In embedded systems, efficient deep learning solutions are crucial to balance accuracy and
resource constraints. Early Exit Neural Networks offer a promising solution, but their manual …
resource constraints. Early Exit Neural Networks offer a promising solution, but their manual …
[PDF][PDF] PEAX-A Model Augmentation Framework for Adaptive Techniques and Embedded Applications
Deploying deep learning models to the edge is challenging due to limited available
resources of embedded devices. To address this, we introduce PEAX, a flexible and …
resources of embedded devices. To address this, we introduce PEAX, a flexible and …