- 25 Jul, 2024 1 commit
-
-
zhanggzh authored
-
- 11 Sep, 2023 1 commit
-
-
Rick Ho authored
-
- 24 Jul, 2023 1 commit
-
-
Rick Ho authored
-
- 04 Jul, 2023 1 commit
-
-
Jiezhong Qiu authored
* support megatron v3.0.2 * keep num_experts for lower version of megatron
-
- 19 May, 2023 3 commits
-
-
Arturo Ghinassi authored
-
Rick Ho authored
-
Arturo Ghinassi authored
AMP support
-
- 18 May, 2023 1 commit
-
-
Arturo Ghinassi authored
When using CUDA AMP FMoE Linear throws type error as input is half() and weights are float()
-
- 21 Mar, 2023 1 commit
-
-
zms1999 authored
-
- 20 Mar, 2023 1 commit
-
-
zms1999 authored
-
- 13 Feb, 2023 1 commit
-
-
Fragile-azalea authored
-
- 14 Dec, 2022 2 commits
- 04 Aug, 2022 1 commit
-
-
Fragile-azalea authored
-
- 10 Jun, 2022 1 commit
-
-
Rick Ho authored
-
- 01 Jun, 2022 3 commits
- 26 May, 2022 1 commit
-
-
Rick Ho authored
-
- 30 Apr, 2022 4 commits
- 02 Apr, 2022 1 commit
-
-
Rick Ho authored
-
- 01 Apr, 2022 1 commit
-
-
Rick Ho authored
-
- 31 Mar, 2022 1 commit
-
-
Rick Ho authored
-
- 30 Mar, 2022 2 commits
- 29 Mar, 2022 3 commits
- 28 Mar, 2022 3 commits
- 29 Nov, 2021 1 commit
-
-
Rick Ho authored
-
- 23 Nov, 2021 4 commits
-
-
Jiezhong Qiu authored
-
Jiezhong Qiu authored
-
Jiezhong Qiu authored
-
Jiezhong Qiu authored
-
- 08 Nov, 2021 1 commit
-
-
Rick Ho authored
-