Commits · d805e1a014ebdaa6a8cb09714d6a4ffad785c8f0 · OpenDAS / Megatron-LM

20 Apr, 2020 4 commits
- Merge branch 'doc-vocab-links' into 'master' · d805e1a0
  Mohammad Shoeybi authored Apr 20, 2020
```
Doc vocab links

See merge request ADLR/megatron-lm!70
```
  d805e1a0
- Move links to download checkpoints section · b46e28bd
  Steven Steinke authored Apr 20, 2020
  
  b46e28bd
- Links to vocab in Data Preprocessing; added "WordPiece" to description · 8e4cb2a6
  Steven Steinke authored Apr 20, 2020
  
  8e4cb2a6
- Links to vocabs and merge table in Data Preprocessing · bdc3d5ba
  Steven Steinke authored Apr 20, 2020
  
  bdc3d5ba
16 Apr, 2020 30 commits
- Merge branch 'staging' into 'master' · 22c0e300
  Jared Casper authored Apr 16, 2020
```
April 2020 Release

See merge request ADLR/megatron-lm!69
```
  22c0e300
- Merge branch 'doc-updates-final' into 'staging' · acf8780b
  Mohammad Shoeybi authored Apr 16, 2020
```
Doc updates final

See merge request ADLR/megatron-lm!68
```
  acf8780b
- remove wikitext comment · 7ec42d95
  Steven Steinke authored Apr 16, 2020
  
  7ec42d95
- Fix embeddings and seq length · 475a6b99
  Steven Steinke authored Apr 16, 2020
  
  475a6b99
- Fix finetine -> tune · 5cb2d1ea
  Steven Steinke authored Apr 16, 2020
  
  5cb2d1ea
- fix typos · 50391a0b
  Steven Steinke authored Apr 16, 2020
  
  50391a0b
- RACE explained · 4f770782
  Steven Steinke authored Apr 16, 2020
  
  4f770782
- Merge branch 'staging_license' into 'staging' · 13d90b36
  Raul Puri authored Apr 16, 2020
```
Update license with Facebook's MIT code and add some clarification.

See merge request ADLR/megatron-lm!66
```
  13d90b36
- changed licence 2019 to 2020 · 463d1257
  Mohammad authored Apr 16, 2020
  
  463d1257
- make minor modifications to licnese wording · 39af7956
  Mohammad authored Apr 16, 2020
  
  39af7956
- Update README.md · d976f011
  Steven Steinke authored Apr 16, 2020
  
  d976f011
- Fixed minor inconsistencies in scripts, added distributed comment · 2e0b3fca
  Steven Steinke authored Apr 16, 2020
  
  2e0b3fca
- Update license with Facebook's MIT code and add some clarification. · 0c6f0f2d
  Jared Casper authored Apr 16, 2020
  
  0c6f0f2d
- Merge branch 'staging_docs' into 'staging' · 050c24d5
  Steven Steinke authored Apr 16, 2020
```
Staging docs

See merge request ADLR/megatron-lm!64
```
  050c24d5
- looks great · 3df23c72
  Mohammad authored Apr 16, 2020
  
  3df23c72
- latest · 0d8f1243
  Mohammad authored Apr 16, 2020
  
  0d8f1243
- latest · 1f45adcc
  Mohammad authored Apr 16, 2020
  
  1f45adcc
- updated scripts · ca19a8b5
  Mohammad authored Apr 16, 2020
  
  ca19a8b5
- made monir changes to the first two paras · fe40745e
  Mohammad authored Apr 16, 2020
  
  fe40745e
- Merge branch 'staging' into staging_docs · add35c17
  Mohammad authored Apr 16, 2020
  
  add35c17
- Merge branch 'staging_dataset_error_reporting' into 'staging' · d928dd51
  Mohammad Shoeybi authored Apr 16, 2020
```
Better error reporting when a dataset doesn't exist.

See merge request ADLR/megatron-lm!63
```
  d928dd51
- moved steves branch · 78cb1781
  Mohammad authored Apr 16, 2020
  
  78cb1781
- Merge branch 'staging_fix_task_main_imports' into 'staging' · c6b5c137
  Raul Puri authored Apr 16, 2020
```
Staging fix task main imports

See merge request ADLR/megatron-lm!62
```
  c6b5c137
- Better error reporting when a dataset doesn't exist. · f2779b18
  Jared Casper authored Apr 16, 2020
  
  f2779b18
- fixed tasks/main imports · 4d39aba4
  Mohammad authored Apr 16, 2020
  
  4d39aba4
- Merge branch 'staging_compile_helpers' into 'staging' · f1f9fa0a
  Jared Casper authored Apr 16, 2020
```
runtime compilation of helpers

See merge request ADLR/megatron-lm!61
```
  f1f9fa0a
- Check that the dataset helpers make runs successfully. · 9073c4bd
  Jared Casper authored Apr 16, 2020
  
  9073c4bd
- added runtime compilation for helpers · c6882114
  Mohammad authored Apr 15, 2020
  
  c6882114
- Merge branch 'staging_optimization_level_2' into 'staging' · 8134313d
  Jared Casper authored Apr 15, 2020
```
Staging optimization level 2

See merge request ADLR/megatron-lm!55
```
  8134313d
- Merge branch 'staging_optimization' into 'staging' · 97913a1f
  Jared Casper authored Apr 15, 2020
```
Staging optimization level 1

See merge request ADLR/megatron-lm!54
```
  97913a1f
15 Apr, 2020 6 commits
- addressed jareds comment · a4489ffb
  Mohammad authored Apr 15, 2020
  
  a4489ffb
- addressed jareds comments · b6e73188
  Mohammad authored Apr 15, 2020
  
  b6e73188
- added faster L2 grad clipping and new torch gelu · 3ea5491e
  Mohammad authored Apr 14, 2020
  
  3ea5491e
- added multi-tensor-apply to fp16 · 99410264
  Mohammad authored Apr 14, 2020
  
  99410264
- Merge branch 'staging' into staging_optimization · 2467ae15
  Mohammad authored Apr 14, 2020
  
  2467ae15
- Merge branch 'staging_optional_args' into 'staging' · be5300b7
  Neel Kant authored Apr 14, 2020
```
made size arguments optional so they can be set from input function

See merge request ADLR/megatron-lm!53
```
  be5300b7