ALCOP: Automatic Load-Compute Pipelining in Deep Learning Compiler for AI-GPUs.
Guyue HuangYang BaiLiu LiuYuke WangBei YuYufei DingYuan XiePublished in: MLSys (2023)
Keyphrases
- deep learning
- machine learning
- parallel processing
- unsupervised learning
- unsupervised feature learning
- artificial intelligence
- general purpose
- restricted boltzmann machine
- mental models
- parallel programming
- deep architectures
- weakly supervised
- expert systems
- multi class
- supervised learning
- multiscale
- decision trees
- learning algorithm