PIT: Optimization of Dynamic Sparse Deep Learning Models via Permutation Invariant Transformation.
Ningxin ZhengHuiqiang JiangQuanlu ZhangZhenhua HanLingxiao MaYuqing YangFan YangChengruidong ZhangLili QiuMao YangLidong ZhouPublished in: SOSP (2023)
Keyphrases
- learning models
- sparse metric learning
- loss function
- learning tasks
- semi supervised learning
- machine learning
- high dimensional
- machine learning algorithms
- learning problems
- classification models
- learning algorithm
- sparse representation
- conditional random fields
- group sparsity
- probabilistic model
- optimization method
- data sets