Skip to content
This repository has been archived by the owner on Aug 5, 2022. It is now read-only.

Caffe_v1.0.1
Compare
Choose a tag to compare
@hshen14 hshen14 released this 07 Jul 06:45
· 826 commits to master since this release
1.0.1
9f0108f

This release includes:

  • Support configurable convolution algorithms (winograd and direct) under MKL-DNN
  • Support batch normal and scale layer fusion in convolution layer
  • Improve performance from MKL-DNN upgrade with optimizations on AVX512
  • Support multi-node training with model parallelism (experimental feature)
  • Support deconvolution layer under MKL2017
  • Integrate MLSL deployment in Makefile
Assets 3
Loading