- 09 Oct, 2020 2 commits
-
-
Mohammad Shoeybi authored
support for different query key sequence lengths See merge request ADLR/megatron-lm!151
-
Vijay Korthikanti authored
-
- 06 Oct, 2020 2 commits
-
-
Jared Casper authored
Avoid non deterministic arch order in compilation See merge request ADLR/megatron-lm!146
-
Vijay Korthikanti authored
-
- 05 Oct, 2020 2 commits
-
-
Vijay Korthikanti authored
-
Vijay Korthikanti authored
-
- 02 Oct, 2020 5 commits
-
-
Mohammad Shoeybi authored
pip 1.1.5 See merge request ADLR/megatron-lm!144
-
Evelina authored
Signed-off-by:Evelina <ebakhturina@nvidia.com>
-
Mohammad Shoeybi authored
1.1.4. pip version See merge request ADLR/megatron-lm!120
-
Mohammad Shoeybi authored
Fixed uninit communication gruop See merge request ADLR/megatron-lm!142
-
Mostofa Patwary authored
-
- 01 Oct, 2020 2 commits
-
-
Jared Casper authored
fix for nemo: do not initialize mpu if it is already initialized See merge request ADLR/megatron-lm!141
-
mohammad authored
-
- 30 Sep, 2020 3 commits
-
-
Mohammad Shoeybi authored
Vijay/weight mapping See merge request ADLR/megatron-lm!138
-
Vijay Korthikanti authored
-
Vijay Korthikanti authored
-
- 29 Sep, 2020 5 commits
-
-
Mohammad Shoeybi authored
Text generation fixes See merge request ADLR/megatron-lm!137
-
Vijay Korthikanti authored
-
Mostofa Patwary authored
-
Vijay Korthikanti authored
-
Mostofa Patwary authored
-
- 25 Sep, 2020 3 commits
-
-
Jared Casper authored
fixed logging bug See merge request ADLR/megatron-lm!134
-
mohammad authored
-
mohammad authored
-
- 24 Sep, 2020 2 commits
-
-
Jared Casper authored
scaling-generalized masking-softmax fusion kernel See merge request ADLR/megatron-lm!131
-
mohammad authored
-
- 22 Sep, 2020 1 commit
-
-
mohammad authored
-
- 18 Sep, 2020 1 commit
-
-
root authored
-
- 11 Sep, 2020 4 commits
-
-
Jared Casper authored
Update scaling numbers in README and other small tweaks. See merge request ADLR/megatron-lm!130
-
Jared Casper authored
-
Jared Casper authored
Various speed optimizations. See merge request ADLR/megatron-lm!124
-
Vijay Korthikanti authored
-
- 03 Sep, 2020 3 commits
-
-
Jared Casper authored
added splitting checkpointed activations across model parallel partitions See merge request ADLR/megatron-lm!121
-
Jared Casper authored
Fix cpu initialization for mp=1 and fp16 See merge request ADLR/megatron-lm!123
-
mohammad authored
-
- 02 Sep, 2020 2 commits
- 01 Sep, 2020 2 commits
-
-
Evelina authored
Signed-off-by:Evelina <ebakhturina@nvidia.com>
-
Mohammad Shoeybi authored
--onnx-safe made overrideable, -lazy_mpu_init act as _USE_CPU_INITIALIZATION See merge request ADLR/megatron-lm!117
-
- 28 Aug, 2020 1 commit
-
-
Boris Fomitchev authored
Signed-off-by:Boris Fomitchev <bfomitchev@nvidia.com>
-