Commits · 928a200c08dd5eea39a5faa7ee6e0c95456e6a9e · OpenDAS / Megatron-LM

21 Jul, 2022 1 commit
- Remove deprecated destination argument to state_dict functions and make all... · 928a200c
  Jared Casper authored Jul 21, 2022
```
Remove deprecated destination argument to state_dict functions and make all arguments keyword to avoid warnings.
```
  928a200c
24 May, 2022 4 commits
- resolved review comments · 739cb43d
  Vijay Korthikanti authored May 24, 2022
  
  739cb43d
- fixed for new jit warmup changes. · cf1c7848
  Lawrence McAfee authored May 24, 2022
  
  cf1c7848
- preallocating global buffer to avoid memory fragmentation · 9dc3c42a
  Vijay Korthikanti authored May 24, 2022
  
  9dc3c42a
- fix for sequence parallelism in bert pooling · 8474e6e5
  Vijay Korthikanti authored May 24, 2022
  
  8474e6e5
20 May, 2022 2 commits
- avoiding sequence parallelism on the pooler · 3f91f09b
  Vijay Korthikanti authored May 20, 2022
  
  3f91f09b
- bert regression fixes · 6aaafee6
  Vijay Korthikanti authored May 20, 2022
  
  6aaafee6
18 May, 2022 1 commit
- t5 regression fixes · 5d647381
  Vijay Korthikanti authored May 18, 2022
  
  5d647381
17 May, 2022 3 commits
- address review comments · d19e8ad7
  Vijay Korthikanti authored May 16, 2022
  
  d19e8ad7
- address review comments · bbab79f8
  Vijay Korthikanti authored May 16, 2022
  
  bbab79f8
- address review comments · 356eb36a
  Vijay Korthikanti authored May 16, 2022
  
  356eb36a
16 May, 2022 1 commit
- cleanup imports in transformer.py · 28b6ffce
  Vijay Korthikanti authored May 16, 2022
  
  28b6ffce
09 May, 2022 1 commit
- (1)support pyt20.12 compatibility, (2) arg naming update · f602ac56
  Sangkug Lym authored Apr 11, 2022
```
change dummy_handler to nullcontext
```
  f602ac56
28 Apr, 2022 1 commit
- address review comments · cfd2e216
  Vijay Korthikanti authored Apr 28, 2022
  
  cfd2e216
30 Mar, 2022 1 commit
- sequence parallelism for embedding dropout and last linear layer + memory optimizations · eec218d8
  Vijay Korthikanti authored Mar 30, 2022
  
  eec218d8
25 Mar, 2022 1 commit
- jit functions warmpups to (1) match fprop and recompute results and (2) remove grad_enable · c7fef593
  Sangkug Lym authored Mar 24, 2022
  
  c7fef593
15 Mar, 2022 2 commits
- fixed. · 65e6bc32
  Lawrence McAfee authored Mar 15, 2022
  
  65e6bc32
- debugging. · 4c598f9d
  Lawrence McAfee authored Mar 15, 2022
  
  4c598f9d
14 Mar, 2022 1 commit
- cleaned distributed.py. · 14e60427
  Lawrence McAfee authored Mar 14, 2022
  
  14e60427
11 Mar, 2022 1 commit
- checkpoint support for core attention · 53718d4c
  Vijay Korthikanti authored Mar 11, 2022
  
  53718d4c
10 Mar, 2022 1 commit
- small fixes. · 862d70fc
  Lawrence McAfee authored Mar 10, 2022
  
  862d70fc
09 Mar, 2022 2 commits
- padded ddp's grad_buffer to multiple of data parallel world size · 55695f81
  Lawrence McAfee authored Mar 09, 2022
  
  55695f81
- bug fixes · dd96d402
  Vijay Korthikanti authored Mar 08, 2022
  
  dd96d402
07 Mar, 2022 2 commits
- fixes to main merge · 269f28f7
  Vijay Korthikanti authored Mar 07, 2022
  
  269f28f7
- refactor to help merge with main · 0d77c0e9
  Vijay Korthikanti authored Mar 07, 2022
  
  0d77c0e9
03 Mar, 2022 1 commit
- get sequence parallelism to work with pipeline parallelism · 6658158b
  Vijay Korthikanti authored Mar 02, 2022
  
  6658158b
02 Mar, 2022 1 commit
- layernorm grad sync + name chnages · c0f10643
  Vijay Korthikanti authored Mar 02, 2022
  
  c0f10643
01 Mar, 2022 1 commit
- Fourth phase vision merge: classification and segmentation tasks · ca7a2e34
  Vijay Korthikanti authored Feb 28, 2022
  
  ca7a2e34
19 Feb, 2022 2 commits
- tensor model parallelism memory optmization · 5d4689c4
  Vijay Korthikanti authored Feb 18, 2022
  
  5d4689c4
- Making loading arguments from checkpoint cleaner and available more broadly. · 942c402d
  Jared Casper authored Feb 18, 2022
  
  942c402d
18 Feb, 2022 1 commit
- support fp32 training and fix embedding update · b5726555
  Sangkug Lym authored Feb 18, 2022
  
  b5726555
16 Feb, 2022 1 commit

Sangkug Lym authored Feb 13, 2022

remove redundant linear layer class definition

add fuse_gradient_accumulation attribute to weights for simple targetting

reflect feedback and clean up the codes

arg change

83b1e42f

15 Feb, 2022 4 commits
- adress review comments · 488f8c02
  Vijay Korthikanti authored Feb 11, 2022
  
  488f8c02
- more minor fixes · 1cd3650d
  Vijay Korthikanti authored Feb 01, 2022
  
  1cd3650d
- minor fixes · 6f3bf9c0
  Vijay Korthikanti authored Feb 01, 2022
  
  6f3bf9c0
- vision third phase merge: pretraining methods + mit,swin backbones · 48c2a144
  Vijay Korthikanti authored Jan 31, 2022
  
  48c2a144
14 Feb, 2022 4 commits
- todo; align shards with model's contiguous buffer · a3f3c3ad
  Lawrence McAfee authored Feb 14, 2022
  
  a3f3c3ad
- tweaked slice index naming convention · 1215c420
  Lawrence McAfee authored Feb 14, 2022
  
  1215c420
- map param to originating virtual model; eventually move this to constructor · c5f93269
  Lawrence McAfee authored Feb 14, 2022
  
  c5f93269
- collect param offsets for contiguous grad buffer · f7232502
  Lawrence McAfee authored Feb 14, 2022
  
  f7232502