提交 · 2c0251814cecd4fc02f6da543bdfebab8d9b9dee · PaddlePaddle / Paddle-Lite

05 1月, 2019 1 次提交
- H
  
  Optimize int8 5x5 depthwise conv, add aarch64 macros to make compilation no problem · 91b8d2be
  由 hjchen2 提交于 1月 05, 2019
  
  91b8d2be
30 12月, 2018 3 次提交
- H
  
  Optimize pooling which efficiency has increased by 30% for googlenet, Fix pooling3x3 for stride 2 · 310b1dbd
  由 hjchen2 提交于 12月 30, 2018
  
  310b1dbd
- H
  
  Fix pooling3x3 for stride 2 · 3b4e3ff7
  由 hjchen2 提交于 12月 30, 2018
  
  3b4e3ff7
- H
  
  Add pooling2x2 neon implementation · dbffe4ad
  由 hjchen2 提交于 12月 30, 2018
  
  dbffe4ad