Code

Gpu optimize and fix precision loss#30

Code

Merged

rozukke merged 5 commits intokachi-group:cuda-stagingkachi-group/ichida-algo:cuda-stagingfrom nhatdongdang:feat/optimize-paramnhatdongdang/ichida-algo:feat/optimize-paramCopy head branch name to clipboardJul 6, 2024

+104-62

Commits

Commits on Jul 6, 2024

• Optimize param

  nhatdongdang

  committed
  9402f01

  Copy full SHA for 9402f01

• Optimize flag

  nhatdongdang

  committed
  dc885f0

  Copy full SHA for dc885f0

• Fix softmax and argmax precision

  nhatdongdang

  committed
  9d479e7

  Copy full SHA for 9d479e7

• Further optimization using shared memory and local memory

  nhatdongdang

  committed
  d4b590e

  Copy full SHA for d4b590e

• Fix compilation flag

  nhatdongdang

  authored
  dd98f99

  Copy full SHA for dd98f99