TIDE
Dynamic per-token early exit for LLM inference. Skip layers tokens don't need
- 35
- Python
- Apache-2.0
- 2 mo ago
Dynamic per-token early exit for LLM inference. Skip layers tokens don't need
The blazing-fast, multimodal AI decision engine. Uses adaptive early-exit computing and attention rollout for sub-15ms explainable zero-shot classification
Code for paper "EdgeKE: An On-Demand Deep Learning IoT System for Cognitive Big Data on Industrial Edge Devices"
A deep learning framework that implements Early Exit strategies in Convolutional Neural Networks (CNNs) using Deep Q-Learning (DQN). This project enhances computational efficiency by dynamically determining the optimal exit point in a neural network for image classification tasks on CIFAR-10.