Closed pfxuan closed 8 months ago
This PR adds and re-uses the cuda/libtorch cache layers across linux build matrix. The workflow run time of each build is reduced from 9 minutes to 3 minutes 🚀.
That's awesome! Thanks @pfxuan
This PR adds and re-uses the cuda/libtorch cache layers across linux build matrix. The workflow run time of each build is reduced from 9 minutes to 3 minutes 🚀.