QuantPipe: Applying Adaptive Post-Training Quantization for Distributed Transformer Pipelines in Dynamic Edge Environments.
Haonan WangConnor ImesSouvik KunduPeter A. BeerelStephen P. CragoJohn Paul WaltersPublished in: CoRR (2022)
Keyphrases
- dynamic environments
- dynamically adjust
- distributed learning
- highly dynamic
- computing environments
- open systems
- edge detection
- supervised learning
- fuzzy logic
- cooperative
- distributed systems
- peer to peer
- test set
- distributed environment
- multi agent
- real world
- robot teams
- heterogeneous environments
- subband image coding
- training process
- wavelet transform
- computational complexity
- training data
- learning algorithm