nv-kkudrynski
2592d5a02c
Merge pull request #931 from roclark/update-glue-downloader
...
[BERT/PyT] Update GLUE Download Tool
2021-11-04 15:08:15 +01:00
Jan Golda
43df08994f
Removed pip upgrade from Dockerfiles
2021-10-28 16:59:30 +02:00
Sharath Turuvekere Sreenivas
26d8955cc5
[PyT/TF/TF2] remove install html2text requirement for NLP models
2021-10-08 01:30:31 -07:00
Sharath Turuvekere Sreenivas
78ec680207
Add GPus/Node in readme
2021-08-27 14:17:30 +00:00
Szymon Migacz
ef98b2cef9
[TXL/PyT] update: ( #989 )
...
* changed API calls to torch.einsum
* added export OMP_NUM_THREADS=1 to all launcher scripts
* additional runtime checks to ensure that launch configuration is valid
2021-08-20 08:39:12 -07:00
Andrei Shumak
6a642837c4
Revert changes to FastPitch and BERT on PyTorch
2021-08-13 16:19:34 +02:00
Krzysztof Kudrynski
a860701665
[BART/PyT] Initial release
2021-08-11 17:08:16 +00:00
Andrei Shumak
9c5766102f
[PyT] BERT: new CI configurations
2021-08-06 15:18:29 +00:00
Krzysztof Kudrynski
49e23b4597
Adding links to performance benchmark page
2021-07-21 14:39:48 +02:00
Robert Clark
1982f5bacd
[BERT/PyT] Update GLUE Download Tool
...
The linked GLUE downloader script has several issues which prevent it
from downloading the MRPC components. Creating a new fork of the linked
gist allows these items to be sorted out.
Signed-Off-By: Robert Clark <roclark@nvidia.com>
2021-04-30 16:43:30 -05:00
nv-kkudrynski
0a3a4d0916
Merge pull request #901 from szmigacz/txl-update3
...
[TXL/PyT] Update for PyT Transformer-XL:
2021-04-06 12:00:46 +02:00
Sharath TS
499fb1c5ad
[BERT/PyT] fix typos
...
* fix typos
* fix typo
2021-04-06 11:46:25 +02:00
Szymon Migacz
09a17a0f33
[TXL/PyT] Update for PyT Transformer-XL:
...
* WAR for issues with logging in distributed setting
* improved CPU-GPU affinity
* minor optimizations for the model
2021-04-02 22:07:48 -07:00
Sharath TS
dd6b8ca2bb
reflect default paths
...
Due to server issues with BookCorpus, the data download script skips downloading BookCorpus by default.
2021-03-09 16:23:55 -08:00
Kazuhiro Yamasaki
c33299b322
Update URLs for GLUE downloader.
2021-03-07 13:27:22 +09:00
Sharath TS
008fa2e19d
[BERT/PyT] Update paths ( #845 )
...
* Update run_pretraining.sh
* fix paths
2021-03-01 13:43:49 -08:00
Sharath TS
04988752a8
[BERT/PyT][BERT/TF] Switch back to the original server for data download
...
* update - wiki download
2021-02-25 14:13:53 -08:00
Sharath TS
3459f97d58
[BERT/PyT][BERT/TF] Use mirror and wget ( #833 )
...
* Use mirror and wget
Mirror speedup download by 10x
wget fixes the stall with urllib
* add comment
* Update wikidownloader of bert tf1 to use mirror and wget
Co-authored-by: Swetha Mandava <sweth.mandava@gmail.com>
2021-02-17 14:49:39 -08:00
Shashank Verma
313c3a589a
Add PyT Language Modeling README
...
Signed-off-by: Shashank Verma <shashank3959@gmail.com>
2020-12-02 00:28:58 -08:00
Szymon Migacz
533f7444ae
[TXL/PyT] Fixed issue with AMP training together with gradient accumulation ( #720 )
2020-10-23 15:46:58 -07:00
Sharath T S
a74236afd4
[BERT/PyT] remove redundant section ( #690 )
2020-09-16 17:06:29 -07:00
Sharath T S
482fe9ac8a
[BERT/PyT] fix onnx export ( #689 )
2020-09-15 12:44:10 -07:00
Szymon Migacz
6b82d3acb3
[TXL/PyT] Minor update for PyTorch Transformer-XL ( #688 )
2020-09-14 14:42:57 -07:00
nv-kkudrynski
152d0c0344
Merge pull request #684 from gpauloski/bert_pytorch_fix
...
[BERT/PyT] Fix dataloader typo
2020-09-11 18:40:26 +02:00
nv-kkudrynski
cf54b787ae
fixed link
2020-09-10 15:48:54 +02:00
Greg Pauloski
7a4c42501c
[BERT/PyT] Fix dataloader typo
2020-09-09 11:28:31 -05:00
kkudrynski
5d36b4fd2f
Fixing hyperlinks
2020-09-08 16:10:32 +02:00
Sharath T S
8588e9834c
[BERT/PyT] specify GPU for triton ( #666 )
2020-09-01 22:07:36 -07:00
Sharath T S
5cc03caa15
[BERT/PyT] Update pretrained checkpoint links ( #660 )
...
* Update pretrained checkpoint links
* update link
2020-08-21 13:34:05 -07:00
kkudrynski
88864b9291
[BERT/PyT] MRPC and SST-2 support
2020-08-14 16:43:11 +02:00
Sharath T S
e8f87acdb1
Keep wikiextractor version fixed
2020-08-10 11:46:17 -07:00
Sharath T S
1f9226283e
[BERT/PyT] link gitlab -> DLE ( #634 )
2020-08-03 22:02:55 -07:00
Sharath T S
0ef5568a9d
[BERT/PyT] default train:test 9:1 ( #616 )
2020-07-22 14:07:50 -07:00
Sharath T S
3337f72cff
[BERT/PyT] Update DataPrep ( #595 )
...
* Update DataPrep
* Update create_datasets_from_start.sh
* Update README.md
* Update README.md
* Update README.md
2020-07-08 17:19:45 -07:00
Sharath T S
b2763ae273
[BERT/PyT] Update ampere perf params ( #589 )
2020-07-07 12:43:03 -07:00
Sharath T S
18db1c17f5
Update README.md
2020-07-07 02:23:36 -07:00
Przemek Strzelczyk
f8b3a63f81
[BERT/PyT] Updating for Ampere and 20.06 container
2020-07-04 03:12:11 +02:00
Przemek Strzelczyk
96138d5087
[BERT/TF] Updating for Ampere
2020-07-04 01:00:48 +02:00
Przemek Strzelczyk
f838cf3292
[Transformer-XL/PyT] Added Ampere support
2020-06-27 09:57:21 +02:00
Sharath T S
9df464f277
[BERT/PyT] stop and resume, single gpu and timing fixes. ( #509 )
...
* stop and resume, single gpu and timing fixes.
* Update utils.py
* accumulation features check
2020-05-17 12:46:53 -07:00
Sharath T S
3aae0204c3
1. stop and resume
2020-05-16 22:47:53 -07:00
PrzemekS
af32ac2a23
Merge pull request #456 from NVIDIA/sharathts-patch-1
...
Fix to load Google's checkpoint
2020-05-07 11:09:20 +02:00
Sharath T S
4733603577
[BERT/PyT] Fix squad inference corner case ( #462 )
2020-04-20 21:02:18 -07:00
Sharath T S
a9c997cd57
Fix to load Google's checkpoint
2020-04-13 11:34:56 -07:00
Sharath T S
5626846924
[BERT/PyT] Revert from native gelu. Breaks ONNX export. ( #447 )
2020-04-07 10:28:45 -07:00
Przemek Strzelczyk
26c2676104
[BERT/PyT] Triton Inference Server support
2020-04-02 14:39:24 +02:00
Sharath T S
793b92dca7
[BERT/PyT] fp32 and allreduce_post_accumulation compatibility ( #422 )
2020-03-15 23:03:06 -07:00
Przemek Strzelczyk
96ff411ce8
[BERT/PyT] Typo in README
2020-03-05 09:54:16 +01:00
Przemek Strzelczyk
155578a762
[BERT/PyT] New logging and some README updates
2020-02-28 13:21:20 +01:00
PrzemekS
ce73b32068
Merge pull request #392 from NVIDIA/nvpstr/1def26
...
Updating BERT/TF, Transformer-XL and NCF/PyT
2020-02-06 19:47:30 +01:00