sensAI

sensAI is a project to efficiently distribute ML.

sensAI paper is accepted by MLSys 2021.

Github: https://github.com/GuanhuaWang/sensAI

Previous short papers:

Computer Vision:
- sensAI: Fast ConvNets Serving on Live Data via Class Parallelism (MLOps Systems Workshop @ MLSys 2020)
Robustness:
- Robust Class Parallelism – Error Resilient Parallel Inference with Low Communication Cost (Asilomar 2020)

Guanhua Wang

guanhua@cs.berkeley.edu

Joseph Gonzalez

Joseph Gonzalez

jegonzal@EECS.Berkeley.EDU

Ion Stoica

istoica@EECS.Berkeley.EDU

Publications:

sensAI: ConvNets Decomposition via Class Parallelism for Fast Inference on Live Data
sensAI: Fast ConvNets Serving on Live Data via Class Parallelism
Robust Class Parallelism – Error Resilient Parallel Inference with Low Communication Cost