Skip to content

Instantly share code, notes, and snippets.

@dcslin

dcslin/15jul.md

Last active July 15, 2020 01:47
Show Gist options
  • Save dcslin/f700f716786b6d5b86c8eb437b289a85 to your computer and use it in GitHub Desktop.
Save dcslin/f700f716786b6d5b86c8eb437b289a85 to your computer and use it in GitHub Desktop.
Download ZIP
15jul.md
Raw
15jul.md

qabot:

  1. fix cos sim bug, value same as pytorch, ok now
  2. lstm+cos sim+margin loss, loss ok, top1 accuracy 4%~10%, http://ncrs/:8888/notebooks/singa-etc/notebook/singa-qabot-train.ipynb
  3. lstm+cos sim+pooling+margin, loss not ok, top1 accuracy not ok
  4. pytorch: lstm+cos sim+pooling+margin, loss ok

kint:

  1. follow numpy convension, if int tensor + int tensor, return int tensor, else return float
  2. check input types and cast to float in the GenTensorScalarFn when necessary
  3. pr merged

c++ print trace stack when crash:

  1. tried https://github.com/bombela/backward-cpp
  2. pros: easy to integrate(add 3 lines to cmake), pretty result, result example: https://gist.github.com/dcslin/e3c36d546e7610a61b10ac7df4a6339b
  3. cons: need to check how to work with SWIG
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment