HLC: A Hardware-friendly Quantization and Cache-based Accelerator for Transformer.
Xiangfeng SunYuanting ZhangYunchang JiangZheng LiBingjin HanJunyi MaiZhibin LuoEnyi YaoPublished in: AICAS (2024)
Keyphrases
- embedded processors
- parallel implementation
- memory hierarchy
- field programmable gate array
- low cost
- hardware and software
- multithreading
- memory access
- fuzzy logic
- hardware implementation
- real time
- power system
- data access
- fault diagnosis
- computer systems
- processor core
- image processing
- memory subsystem
- embedded systems
- model selection
- friendly interface
- prefetching
- hit rate
- quantization error
- computing power
- massively parallel
- query processing
- database systems
- tree models
- computing systems
- response time
- computational complexity
- neural network