Skip to content

Shifu 0.2.7 Distributed LR Algorithm Improvement (Experimental)

Jump to bottom Edit New page
Zhang Pengshan (David) edited this page Aug 31, 2015 · 1 revision

  1. Distributed LR is added if Set Algorithm in Train Step to 'LR'

  2. 'LearningRate' and 'RegularizedConstant' are two parameters in LR Configurations.

  3. Whole process in Shifu, LR is Supported except 'export' Step.

  4. TODO: Performance Comparison.

Shifu: A Distributed Model Training Framework on Hadoop

DOWNLOAD

Clone this wiki locally