- 21 Jan, 2021 1 commit
-
-
Jeff Daily authored
use __launch_bounds__(1024) for multi_tensor_apply, re-enable skipped tests
-
- 18 Jan, 2021 1 commit
-
-
Jeff Daily authored
-
- 31 Dec, 2020 2 commits
-
-
lcskrishna authored
-
lcskrishna authored
-
- 01 Dec, 2020 1 commit
-
-
Kexin Yu authored
DistributedFusedAdam Model Parallelism Support (Megatron) Co-authored-by:
Kexin Yu <kexiny@nvidia.com> Co-authored-by:
Kexin Yu <kexinznzn@gmail.com>
-
- 05 Aug, 2020 1 commit
-
-
ngimel authored
* add device guards to the optimizers * add untracked file * set deviceGuard in multi_tensor_apply * address review comments; fix lamb * indent * typo
-
- 07 Jul, 2020 1 commit
-
-
lcskrishna authored
-
- 23 Jun, 2020 3 commits
- 26 May, 2020 1 commit
-
-
rohithkrn authored
-
- 14 May, 2020 1 commit
-
-
Andrew Tulloch authored
-
- 03 Sep, 2019 1 commit
-
-
Deyu Fu authored
* move import of amp_C to __init__() * make fp16/32 separate lists to support mixed param types, disable double test * make zero_grad consistent between adam/novograd/lamb
-
- 17 Aug, 2019 1 commit
-
-
Deyu Fu authored
-
- 13 Aug, 2019 1 commit
-
-
Deyu Fu authored
FusedSGD now work as before FusedAdam now work with o1/o2, no longer fuse scaling and casting Removed special backend handling for FusedAdam Moved and updated test for FusedAdam into run_optimizers Removed legacy tests for optimizers.FP16_optimizer and FusedAdam in run_mixed_adam
-
- 12 Aug, 2019 1 commit
-
-
Deyu Fu authored
-
