Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Use global gradient clipping in FusedLAMB & add option for using NVLAMB
#819
opened May 1, 2020 by
kexinyu
fix cuda mem access error, which is described in: https://github.com/…
#689
opened Jan 24, 2020 by
chengmengli06
disable one_hot and gumbel_softmax with dim arg for Pytorch<1.1.0
#468
opened Sep 3, 2019 by
ptrblck
ProTip!
Add no:assignee to see everything that’s not assigned.