Commits · 4a4b0e57831aca4294ee1b2924375ed587bd345b · chenpangpang / transformers

14 Dec, 2018 2 commits
- remove logging. basicConfig from library code · 4a4b0e57
  thomwolf authored Dec 14, 2018
  
  4a4b0e57
- set encoding to 'utf-8' in calls to open · ae88eb88
  thomwolf authored Dec 14, 2018
  
  ae88eb88
13 Dec, 2018 17 commits
- no fp16 on evaluation · e1eab59a
  thomwolf authored Dec 13, 2018
  
  e1eab59a
- fix reloading model for evaluation in examples · 087798b7
  thomwolf authored Dec 13, 2018
  
  087798b7
- fix swag example for work with apex · 0f544625
  thomwolf authored Dec 13, 2018
  
  0f544625
- make examples work without apex · 0cf88ff0
  thomwolf authored Dec 13, 2018
  
  0cf88ff0
- clean up apex integration · 52c53f39
  thomwolf authored Dec 13, 2018
  
  52c53f39
- run_swag example in readme · 4946c2c5
  thomwolf authored Dec 13, 2018
  
  4946c2c5
- model loading apex modification · d23eed85
  thomwolf authored Dec 13, 2018
  
  d23eed85
- include version number + comment in setup.py · 1cbb32a5
  thomwolf authored Dec 11, 2018
  
  1cbb32a5
- added version in __init__.py · ce521776
  thomwolf authored Dec 11, 2018
  
  ce521776
- add saving and loading model in examples · d3fcec1a
  thomwolf authored Dec 11, 2018
  
  d3fcec1a
- add pretrained loading from state_dict · 93f335ef
  thomwolf authored Dec 11, 2018
  
  93f335ef
- adding save checkpoint and loading in examples · b3caec5a
  thomwolf authored Dec 09, 2018
  
  b3caec5a
- compatibility PT 1.0 and 0.4.1 · 85fff78c
  thomwolf authored Dec 09, 2018
  
  85fff78c
- fixing Adam weights skip in TF convert script · 13bf0d46
  thomwolf authored Dec 09, 2018
  
  13bf0d46
- Merge pull request #116 from FDecaYed/deyuf/fp16_with_apex · 91aab2a6
  Thomas Wolf authored Dec 13, 2018
```
Change to use apex for better fp16 and multi-gpu support
```
  91aab2a6
- Merge pull request #113 from hzhwcmhf/master · 32a227f5
  Thomas Wolf authored Dec 13, 2018
```
fix compatibility with python 3.5.2
```
  32a227f5
- Merge pull request #96 from rodgzilla/multiple-choice-code · ffe9075f
  Thomas Wolf authored Dec 13, 2018
```
BertForMultipleChoice and Swag dataset example.
```
  ffe9075f
12 Dec, 2018 3 commits
- add fallback path for apex used in modeling.py · 3b0a14b7
  Deyu Fu authored Dec 12, 2018
  
  3b0a14b7
- Swag example readme section update with gradient accumulation run. · dcb50eaa
  Grégory Châtel authored Dec 12, 2018
  
  dcb50eaa
- change to apex for better fp16 and multi-gpu support · c8ea2860
  Deyu Fu authored Dec 05, 2018
  
  c8ea2860
11 Dec, 2018 11 commits
- add pathlib support for file_utils.py on python 3.5 · 485adde7
  hzhwcmhf authored Dec 11, 2018
  
  485adde7
- fix compatibility with python 3.5.2; convert path to str · bc659f86
  hzhwcmhf authored Dec 11, 2018
  
  bc659f86
- Merge branch 'fourth-release' of... · 1df6f262
  thomwolf authored Dec 11, 2018
```
Merge branch 'fourth-release' of https://github.com/huggingface/pytorch-pretrained-BERT into fourth-release
```
  1df6f262
- include version number + comment in setup.py · 770f805a
  thomwolf authored Dec 11, 2018
  
  770f805a
- added version in __init__.py · ed3b62cd
  thomwolf authored Dec 11, 2018
  
  ed3b62cd
- Merge branch 'master' into fourth-release · 632f2d2d
  Thomas Wolf authored Dec 11, 2018
  
  632f2d2d
- add saving and loading model in examples · b13abfa9
  thomwolf authored Dec 11, 2018
  
  b13abfa9
- add pretrained loading from state_dict · 270fa2f2
  thomwolf authored Dec 11, 2018
  
  270fa2f2
- Bump up requirements to Python 3.6 · a3a3180c
  Thomas Wolf authored Dec 11, 2018
  
  a3a3180c
- Merge pull request #107 from lliimsft/master · e7c0a8dd
  Thomas Wolf authored Dec 11, 2018
```
Fix optimizer to work with horovod
```
  e7c0a8dd
- Merge pull request #91 from rodgzilla/convert-examples-code-improvement · e622790a
  Thomas Wolf authored Dec 11, 2018
```
run_classifier.py improvements
```
  e622790a
10 Dec, 2018 3 commits
- Removing the dependency to pandas and using the csv module to load data. · df34f228
  Grégory Châtel authored Dec 10, 2018
  
  df34f228
- Change to the README file to add SWAG results. · 0876b77f
  Grégory Châtel authored Dec 10, 2018
  
  0876b77f
- Fix optimizer to work with horovod · 81e1e248
  Li Li authored Dec 10, 2018
  
  81e1e248
09 Dec, 2018 4 commits
- adding save checkpoint and loading in examples · 174cdbcc
  thomwolf authored Dec 09, 2018
  
  174cdbcc
- compatibility PT 1.0 and 0.4.1 · 1db916b5
  thomwolf authored Dec 09, 2018
  
  1db916b5
- fixing Adam weights skip in TF convert script · 68f77303
  thomwolf authored Dec 09, 2018
  
  68f77303
- Merge pull request #101 from davidefiocco/patch-1 · a2b6918a
  Thomas Wolf authored Dec 09, 2018
```
Adding --do_lower_case for all uncased BERTs examples
```
  a2b6918a