- 21 Nov, 2019 1 commit
-
-
Mohammad Shoeybi authored
-
- 20 Nov, 2019 2 commits
-
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
- 19 Nov, 2019 1 commit
-
-
Jared Casper authored
-
- 13 Nov, 2019 1 commit
-
-
Jared Casper authored
Some cleanup of types, etc. in the helpers cpp code.
-
- 11 Nov, 2019 2 commits
-
-
Mohammad Shoeybi authored
Merge branch 'albert_data_loader' of ssh://gitlab-master.nvidia.com:12051/ADLR/megatron-lm into albert_data_loader
-
Mohammad Shoeybi authored
-
- 09 Nov, 2019 2 commits
-
-
Jared Casper authored
-
Jared Casper authored
Also removed the fix_lua_indexing stuff from fairseq.
-
- 08 Nov, 2019 5 commits
-
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
Jared Casper authored
-
- 07 Nov, 2019 2 commits
-
-
Jared Casper authored
-
Mohammad Shoeybi authored
-
- 04 Nov, 2019 2 commits
-
-
Mohammad Shoeybi authored
Fixed edge case with multiple end of sequence in one sequence See merge request ADLR/megatron-lm!12
-
Raul Puri authored
-
- 01 Nov, 2019 1 commit
-
-
Jared Casper authored
refactored for code reuse See merge request ADLR/megatron-lm!11
-
- 31 Oct, 2019 2 commits
-
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
- 29 Oct, 2019 2 commits
-
-
Raul Puri authored
Major refactoring, combining gpt2 and bert See merge request ADLR/megatron-lm!8
-
Mohammad Shoeybi authored
-
- 09 Oct, 2019 5 commits
-
-
Raul Puri authored
Move vocab padding to utils See merge request ADLR/megatron-lm!6
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
-
Raul Puri authored
fixed deserializing issue with old checkpoint See merge request ADLR/megatron-lm!5
-
Mohammad Shoeybi authored
-
- 08 Oct, 2019 6 commits
-
-
Jared Casper authored
moved few common elements between bert and gpt to utils See merge request ADLR/megatron-lm!4
-
Mohammad Shoeybi authored
-
Jared Casper authored
created megatron package See merge request ADLR/megatron-lm!3
-
Mohammad Shoeybi authored
-
Mohammad Shoeybi authored
Use DDP command line argument instead of source flag in pretrain_bert.py. See merge request ADLR/megatron-lm!2
-
Jared Casper authored
Note that there is currently an issue with bert using Torch DDP.
-
- 07 Oct, 2019 2 commits
-
-
Jared Casper authored
Updating public repo with latest changes. See merge request ADLR/megatron-lm!1
-
Jared Casper authored
-
- 04 Oct, 2019 1 commit
-
-
Jared Casper authored
-
- 12 Sep, 2019 2 commits
-
-
Raul Puri authored
-
Jared Casper authored
Fixes #7.
-
- 13 Aug, 2019 1 commit
-
-
Raul Puri authored
-
