Add pytorch examples - #1

Merge changes from original

Upstream changes

update master

commit 85f01860095a2d30049cb95b8303d944365498f9
Merge: 23ffbeb3 40e07425
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Jun 17 13:22:55 2019 +0200

    Merge branch 'gh/master' (updates from github)

commit 40e074257fb8670b0284a37c92b9372bb1587354
Merge: 34a01f7b 20f09772
Author: nvpstr <[email protected]>
Date:   Thu Jun 13 17:59:16 2019 +0200

    Merge pull request #65 from JanuszL/fix_RN50_pytorch_DALI

    Fix typo in RN50 PyTorch readme, remove unneeded lines for DALI

commit 34a01f7b7201cffd16528247647b70fc080829c6
Merge: 9df04c29 cd284a62
Author: nvpstr <[email protected]>
Date:   Thu Jun 13 17:35:33 2019 +0200

    Merge pull request #77 from GrzegorzKarchNV/warmup

    updated readme: text.txt->phrase.txt, updated inference results

commit 9df04c29c9168f620f3017069f53f92d2e8716da
Merge: 4aa6f171 a8328ce1
Author: nvpstr <[email protected]>
Date:   Thu Jun 13 17:35:16 2019 +0200

    Merge pull request #79 from nv-kkudrynski/ssd_torchhub_exposure

    ssd exposure via hubconf.py (together with ssd utils)

commit a8328ce1690fe473f3c5ac55811dab6b0f36700d
Author: Krzysztof Kudrynski <[email protected]>
Date:   Wed Jun 12 17:41:57 2019 +0200

    ssd exposure via hubconf.py (together with ssd utils)

commit 23ffbeb3e7a42dfc749b98602efe87fb960a43ce
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Jun 11 16:01:17 2019 +0200

    Adding repos.txt

commit ee7f79238e373fc4bbc3be35770813b56d07667c
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Jun 11 14:17:33 2019 +0200

    few lines in README

commit cd284a62ec3855882fa2f2941beebcfde89947b4
Author: Grzegorz Karch <[email protected]>
Date:   Fri Jun 7 08:29:39 2019 -0700

    changed text to phrase as input to inference in the readme

commit dc7cd1ae5bf3815262bdadc3537646e2b6928c88
Author: Grzegorz Karch <[email protected]>
Date:   Fri Jun 7 08:27:52 2019 -0700

    updated readme

commit 79b9a48bf1662a4f178a1a6a520c79bea23bf41d
Author: gkarch <[email protected]>
Date:   Fri Jun 7 15:10:54 2019 +0200

    added warmup parameter to tacotron2 inference; updated README

commit 4aa6f1716789876066120df5a151f596b8064ffa
Merge: da165c3a 362dfe6b
Author: nvpstr <[email protected]>
Date:   Fri Jun 7 15:11:28 2019 +0200

    Merge pull request #76 from GrzegorzKarchNV/torchhub

    added text_to_sequence in the hubconf model

commit 362dfe6b3b580fa3d2e585d02bcca1f253a7fff5
Author: gkarch <[email protected]>
Date:   Fri Jun 7 14:52:23 2019 +0200

    added text_to_sequnce in the hubconf model

commit f82b52755263b95854de21d548b5c2900bb78d83
Merge: 590a48a7 da165c3a
Author: Przemek Strzelczyk <[email protected]>
Date:   Thu Jun 6 17:45:31 2019 +0200

    Merge branch 'gh/master'

commit da165c3a52cbb48bf0c5f90b8985922edda43585
Merge: bf1b29d7 2f344421
Author: nvpstr <[email protected]>
Date:   Thu Jun 6 16:03:19 2019 +0200

    Merge pull request #73 from GrzegorzKarchNV/master

    updated readme - warmup for inference, updated results for 19.05-py3 container

commit bf1b29d73ed69c48937cd10bc76252c28c17747f
Merge: 698e7756 f3c1cc72
Author: nvpstr <[email protected]>
Date:   Thu Jun 6 16:02:28 2019 +0200

    Merge pull request #63 from khcs/bert/pretrain-fix

    Fix FLAG.save_checkpoint_steps naming inconsistency.

commit 698e7756807a2206a7d7ab9d6824b95fd686b9f4
Merge: c5bbd74f 2774e849
Author: nvpstr <[email protected]>
Date:   Thu Jun 6 16:01:51 2019 +0200

    Merge pull request #72 from jbaczek/master

    Add missing submodule for PyT Transformer

commit c5bbd74f87c8849183d47d81eddb2875d9d7a8df
Merge: 2619f172 e5a0a056
Author: nvpstr <[email protected]>
Date:   Thu Jun 6 15:58:28 2019 +0200

    Merge pull request #62 from NVIDIA/nvpstr-patch-1

    Update README.md

commit 2f34442146dd67e48fa7872942c020ed57ed02e7
Author: gkarch <[email protected]>
Date:   Thu Jun 6 15:27:21 2019 +0200

    update perf numbers

commit 4dd9ecfcfab11c81b4859c2008e6cf729dacf732
Author: gkarch <[email protected]>
Date:   Thu Jun 6 14:16:51 2019 +0200

    updated readme - warmup for inference

commit 2774e849be96654a9b4dc992bf84ff7de477b246
Author: jbaczek <[email protected]>
Date:   Wed Jun 5 16:20:30 2019 +0200

    remove empty dir

commit 743162906cac6783c57f65c37f4ac65a958e5148
Author: jbaczek <[email protected]>
Date:   Wed Jun 5 16:17:24 2019 +0200

    fix submodule path

commit 698516f8f78fd16dd2987f7da15454a31a8620ad
Author: jbaczek <[email protected]>
Date:   Wed Jun 5 16:14:41 2019 +0200

    add missing submodule

commit 2619f172c7670f0f69d0ac8cf64b5f31db01657c
Merge: b7eff3da 867ab68f
Author: nvpstr <[email protected]>
Date:   Wed Jun 5 14:45:27 2019 +0200

    Merge pull request #67 from GrzegorzKarchNV/master

    Update models.py - fix fp16 inference

commit 867ab68fada01b4b8ad9de448729faf6d409af5d
Author: GrzegorzKarchNV <[email protected]>
Date:   Tue Jun 4 15:44:05 2019 +0200

    Update models.py

commit 20f09772f8c741d28373db6f9ddebb5d54735103
Author: Janusz Lisiecki <[email protected]>
Date:   Mon Jun 3 12:03:18 2019 +0200

    Fix typo in RN50 PyTorch readme, remove unneeded lines for DALI

    Signed-off-by: Janusz Lisiecki <[email protected]>

commit f3c1cc7204374951b81e308033034115572c453a
Author: Hoo Chang Shin <[email protected]>
Date:   Thu May 30 11:37:46 2019 -0400

    Fix FLAG.save_checkpoint_steps naming inconsistency.

commit e5a0a056d229175640bbc24e3b99ce836e7c2bb6
Author: nvpstr <[email protected]>
Date:   Thu May 30 13:42:31 2019 +0200

    Update README.md

commit b7eff3da415a3848944c059d139eeed4ac16f447
Merge: 3454305d b25181c9
Author: nvpstr <[email protected]>
Date:   Thu May 30 13:25:58 2019 +0200

    Merge pull request #60 from swethmandava/smandava/bug_fix

    num_gpus pretraining fix

commit b25181c9b8d75ceb84eff2b3f3a648156df25da1
Author: Swetha Mandava <[email protected]>
Date:   Tue May 28 09:52:52 2019 -0700

    num_gpus pretraining fix

commit 3454305d61478afa5116a2892cc99b03fbce8013
Merge: c71a3fe0 5eaebefc
Author: nvpstr <[email protected]>
Date:   Tue May 28 11:19:41 2019 +0200

    Merge pull request #58 from NVIDIA/nvpstr/master

    Updating PyTorch models

commit 5eaebefcd0733f06b122279d84293047e28623cb
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 15:24:14 2019 +0200

    Updating PyTorch models
    * New training scripts & results for RN50 PyT
    * Documentation updates for Trasformer, SSD and MaskRCNN

commit 590a48a7cdb2b9716ff5c2f612536bc1a6e99434
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 05:42:56 2019 -0700

    Update README.md

commit 048ec25a8a0f1db17596142ef00578a304bf88d4
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 05:41:26 2019 -0700

    Update README.md

commit e002d9e69ef62bf98f68d8e3bebcc971ae6e1728
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 05:39:44 2019 -0700

    Update README.md

commit 89f91de43c09cb24ff8003c94075f4874f9e8830
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 14:37:50 2019 +0200

    Updating SSD PyT

commit d5e64e5dd7bd8125fbdfc1335a3f3a0f556151a7
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon May 27 14:30:30 2019 +0200

    Updating RN50 PyT

commit b10122c6837563584616f079367be807173a4938
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 02:16:17 2019 +0200

    Updating Transformer PyT

commit 612067beba1b6ddc7d798fd986209927e02c26cf
Merge: 2f251b95 c71a3fe0
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 01:45:53 2019 +0200

    Merge branch 'gh/master'

commit c71a3fe05b05f736cb108186d8698c085b6bdeb0
Merge: 12c7bc23 3306ea2a
Author: nvpstr <[email protected]>
Date:   Sat May 25 01:43:24 2019 +0200

    Merge pull request #52 from sharathts/master

    Fix README

commit 3306ea2ab93d497152a0464a7e4d11e00114578a
Merge: 604b64e5 12c7bc23
Author: nvpstr <[email protected]>
Date:   Sat May 25 01:43:11 2019 +0200

    Merge branch 'master' into master

commit 12c7bc23020554c8380b788964d02bcf2c594820
Merge: 9712cf82 e46eed4c
Author: nvpstr <[email protected]>
Date:   Sat May 25 01:37:31 2019 +0200

    Merge pull request #45 from eric-haibin-lin/patch-1

    Fix a few typos, and add gluonnlp implementation link

commit 9712cf8281699e6d5256887ab1c21d656347a00a
Merge: 94472422 8662db35
Author: nvpstr <[email protected]>
Date:   Sat May 25 01:37:07 2019 +0200

    Merge pull request #54 from vinhngx/master

    adding resume capability for wget in case data downloads stop early

commit 94472422fa03852a19c9552c9d8c37d10ed7dcc4
Merge: 2d33a722 d2bc3da0
Author: nvpstr <[email protected]>
Date:   Sat May 25 01:36:46 2019 +0200

    Merge pull request #55 from NVIDIA/nvpstr/master

    Adding UNet(medical)/TF & updating other TF models

commit d2bc3da0a185c2cdadb0b837456b272bcbc742dd
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 01:23:11 2019 +0200

    Changes in TF models:
    * added UNet for medical image segmentation
    * added TF-AMP support for RN50
    * small updates for other models (READMEs, benchmark & testing scripts)

commit 2f251b9575b147760594ffdd2b04e6cbdac7638d
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 01:09:42 2019 +0200

    Small fixes for NCF TF

commit e0db822e411e0fcb42260b3cc908623e322ff703
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 01:03:55 2019 +0200

    Changing script name in GNMT TF

commit 62d90b3b571f9c309a5c521e1f09541df65e9d27
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:57:36 2019 +0200

    Updating main README

commit 90eff9fa9c3ae242f33e16841ae774c0ce4b9ba3
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:54:52 2019 +0200

    Updating Dockerfile in Unet industrial

commit 5b0ef2d1fd82d713c884641b303ad7b12f91a427
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:54:09 2019 +0200

    Updating UNet industrial for 19.05

commit 8936de91e5b769398f9d158eb233177bec9e4be8
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:40:24 2019 +0200

    Changelog updated for SSD, RN50, UNet(med) in TF

commit 168d8d2cfdcffe39283c941ab98d612e7f0aaf39
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:29:54 2019 +0200

    Adding UNet for medical image segmentation in 19.05

commit 61d3a4e06d1c054c38eff7547919e9f0b51a1070
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat May 25 00:19:23 2019 +0200

    Updating SSD TF to 19.05

commit 429a0066b9fe3529cf633a8b75c869477b83f219
Author: Przemek Strzelczyk <[email protected]>
Date:   Fri May 24 23:15:52 2019 +0200

    Changelog updated for RN50 TF

commit a3831e5699ca3a40fcbe2f2a878ec49a1a5b9cd0
Author: Przemek Strzelczyk <[email protected]>
Date:   Fri May 24 23:08:04 2019 +0200

    Updating RN50 TF to 19.05

commit b0c675239e519822287df48a90ead2146731af7a
Merge: 218e0691 2d33a722
Author: Przemek Strzelczyk <[email protected]>
Date:   Fri May 24 22:11:19 2019 +0200

    Merging changes from github back

commit 8662db35a7d1e8b9a25c653e6b5e1e49fa3050f6
Author: Vinh Nguyen <[email protected]>
Date:   Fri May 24 14:47:43 2019 +1000

    fix readme: direct link for Apex. Fix download script location

commit eee04908100efd2f59c7114857c1f28a7a91c46c
Author: Vinh Nguyen <[email protected]>
Date:   Fri May 24 14:41:23 2019 +1000

    adding resume capability for wget in case downloads stop early

commit 604b64e55fcb69c6e134fc2d24fc3908337bdf64
Author: sharatht <[email protected]>
Date:   Thu May 23 12:29:46 2019 -0700

    resolve merge conflict readme

commit ac3e9ddb5188ecf50bd30ccd532268272b5debe5
Merge: f2bc6ae8 2d33a722
Author: Sharath T S <[email protected]>
Date:   Thu May 23 12:26:15 2019 -0700

    Merge branch 'master' into master

commit f2bc6ae8ab16f8f8adbc4567e2ec79656f6dea3d
Author: sharatht <[email protected]>
Date:   Thu May 23 12:21:00 2019 -0700

    Fix README

commit 2d33a72240cb8e359e199712f506f170e1d158b5
Merge: bc64ab1a 161a4ea1
Author: nvpstr <[email protected]>
Date:   Thu May 23 16:17:40 2019 +0200

    Merge pull request #48 from nv-kkudrynski/torchhub_exposure

    hub exposure fix and improvements

commit 161a4ea16544997c7f44e5f3f8f689057cd2239e
Author: Krzysztof Kudrynski <[email protected]>
Date:   Tue May 21 15:10:38 2019 +0200

    caching of pre-trained weights added to entrypoints

commit 14fe91fad5c95dc9458047a595338877e4546bf5
Author: Krzysztof Kudrynski <[email protected]>
Date:   Tue May 21 13:32:36 2019 +0200

    added missing nested-modules-discovery from global entrypoint like torch.hub

commit bc64ab1ac585ae5227e30c00441058939b002d94
Merge: 3cf8d6c3 5ebe18f9
Author: nvpstr <[email protected]>
Date:   Mon May 20 14:00:12 2019 +0200

    Merge pull request #46 from zhiqwang/master

    Fix typo

commit 3cf8d6c3c568b13219564784275aaa67735b4ed2
Merge: 205058fd c5951be7
Author: nvpstr <[email protected]>
Date:   Mon May 20 13:59:49 2019 +0200

    Merge pull request #47 from swethmandava/master

    fix mismatch in cl arguments for BERT

commit c5951be72749d54f887d3a2d305d646377aba9b3
Author: Swetha Mandava <[email protected]>
Date:   Sun May 19 10:10:58 2019 -0700

    fix mismatch in cl arguments

commit 5ebe18f983f75526e650bb172a43b1ca9515ff6b
Author: zhiqiang <[email protected]>
Date:   Sun May 19 18:45:53 2019 +0800

    Fix typo

commit e46eed4c573d7c87bff5e625a8af2c2735566163
Author: Haibin Lin <[email protected]>
Date:   Sat May 18 22:41:42 2019 -0700

    Update README.md

commit 205058fd7c1941cd7d6a86a4eaf417bdc32730cb
Merge: f1911365 49acab48
Author: nvpstr <[email protected]>
Date:   Wed May 15 18:26:14 2019 +0200

    Merge pull request #43 from nv-kkudrynski/torchhub_exposure

    torch.hub entrypoints for tacotron2, waveglow and ncf

commit f19113657982ab0bbe39e9bf9ef605ac1af3ffb9
Merge: c4a42e4a fcaa639b
Author: nvpstr <[email protected]>
Date:   Wed May 15 15:52:22 2019 +0200

    Merge pull request #36 from khoa-ho/master

    Add current models to README

commit c4a42e4acb96356655c832b59874abd2c9635156
Merge: a9d0554e b0d25926
Author: nvpstr <[email protected]>
Date:   Wed May 15 15:48:14 2019 +0200

    Merge pull request #42 from khoa-ho/patch-1

    Typo/stylistic fix for BERT README title

commit 49acab486b70522831f1d7868f3e620eadbba9b3
Author: Krzysztof Kudrynski <[email protected]>
Date:   Wed May 15 14:16:46 2019 +0200

    torch.hub entrypoints for tacotron2, waveglow and ncf

commit b0d259264e437b1f158371ddfb00d17106ad7a0f
Author: Khoa Ho <[email protected]>
Date:   Tue May 14 10:30:27 2019 -0700

    Typo/stylistic fix for README title

commit a9d0554e5d6ac2e81077411ffc1f83eb7365a720
Merge: 531a570c 55517528
Author: nvpstr <[email protected]>
Date:   Mon May 13 17:32:30 2019 +0200

    Merge pull request #30 from swethmandava/smandava/readme_edit

    changing perf numbers to steady state throughput

commit fcaa639bd94737fa594c3d8803642611704df99d
Author: Khoa Ho <[email protected]>
Date:   Fri May 3 15:29:25 2019 -0700

    Add current models to README

commit 5551752822fa4abe7f08defbd9b6c62aa75577ac
Author: Swetha Mandava <[email protected]>
Date:   Thu Apr 18 11:51:46 2019 -0700

    Update readme error

commit 83fceae509f216366a4c6f80911793e742b52870
Author: Swetha Mandava <[email protected]>
Date:   Wed Apr 17 14:16:40 2019 -0700

    changing perf numbers to steady state throughput

commit 531a570c5cc1705041ca69f3841bdb437022309b
Merge: 51d3dceb 2409559c
Author: nvpstr <[email protected]>
Date:   Mon Apr 8 18:19:11 2019 +0200

    Merge pull request #25 from skierat/patch-1

    Update Dockerfile

commit 51d3dceb3fefdf44d1b073c4625dc7bf549410b7
Merge: 2acfcb26 ec9eaf73
Author: nvpstr <[email protected]>
Date:   Mon Apr 8 18:18:39 2019 +0200

    Merge pull request #24 from swethmandava/smandava/readme_edit

    fix readme

commit ec9eaf732a4b6358b740b95cb4c0bfec0c7bd06a
Author: Swetha Mandava <[email protected]>
Date:   Thu Apr 4 13:14:01 2019 -0700

    removing fast_math, etc from README

commit 42c6a147f2f551ec6b858d924915e254f128f446
Author: Swetha Mandava <[email protected]>
Date:   Mon Apr 1 11:30:36 2019 -0700

    extra argument removed

commit 2409559c12e4e21bf01493ecd657c43d633bfc15
Author: skierat <[email protected]>
Date:   Mon Apr 1 14:34:29 2019 +0200

    Update Dockerfile

commit bb0332437b4320d121da9ee851cff84288250a48
Author: Swetha Mandava <[email protected]>
Date:   Thu Mar 28 16:10:54 2019 -0700

    fix readme

commit 2acfcb263a5189937c2183450b55b740c6801f5c
Merge: 041c3524 558917b3
Author: nvpstr <[email protected]>
Date:   Wed Mar 27 17:52:09 2019 +0100

    Removing obsolete code (BERT)

    Nvpstr/fix

commit 558917b3c29aa4458b281c236936bea89d9342e2
Merge: 93fa95e3 041c3524
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Mar 27 17:47:46 2019 +0100

    Additional cleanup (BERT)

commit 93fa95e3c78a7c5490cef20aa26cd0068dc5eabc
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Mar 27 17:34:23 2019 +0100

    Removing obsolete code (BERT)

commit 218e06919f31ffc1e646f89681f676432cdd1873
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Mar 27 17:16:33 2019 +0100

    Merged bert_tf 8cebfc0377b72d37562762d341ef79d9caf72c38

commit 041c35247b3ca917510922ecd8ee9e04f3a72473
Author: Michał Krzysztof Filipiuk <[email protected]>
Date:   Sat Mar 23 15:27:15 2019 +0100

    Fixed indentation

commit 783f5facde9f24394d8368a1fff4e7f23d838568
Merge: 0e66c6da dfbce5d1
Author: nvpstr <[email protected]>
Date:   Fri Mar 22 16:47:21 2019 +0100

    Merge pull request #19 from swethmandava/master

    removing multinode scripts

commit dfbce5d17880d56d0a9dc24fb55c56047d5334a0
Author: Swetha Mandava <[email protected]>
Date:   Thu Mar 21 23:02:48 2019 -0700

    removing multinode scripts

commit 0e66c6dabb8b4c90bd637e27aeb4e67722ca95fc
Merge: 21024d91 4b09b347
Author: nvpstr <[email protected]>
Date:   Tue Mar 19 17:14:07 2019 +0100

    Merge pull request #16 from NVIDIA/nvpstr/master

    BERT README update

commit 4b09b347a764de1aebb12af13ec1a2354acc6647
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 17:12:43 2019 +0100

    BERT README update

commit c0887ff7e1879e75a6193f1c32f518e6d68764f0
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 16:59:43 2019 +0100

    BERT README update

commit dfd3b4114fc70301535d23e562616896eafe2493
Merge: a550334e 21024d91
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 16:52:48 2019 +0100

    Merge branch 'master' of https://github.com/NVIDIA/DeepLearningExamples

commit 21024d9126376f14a0de29d85e1687b53ba4c003
Merge: a68a2a0e 3d90b6db
Author: nvpstr <[email protected]>
Date:   Tue Mar 19 16:52:05 2019 +0100

    Merge pull request #15 from NVIDIA/nvpstr/master

    Updating SSD and BERT.

commit a68a2a0e729214a5b6833f90a098ae2c0f984426
Merge: 8a3d34e8 f0ec1d40
Author: nvpstr <[email protected]>
Date:   Tue Mar 19 16:50:20 2019 +0100

    Merge pull request #14 from DEKHTIARJonathan/patch-1

    [UNet-TF  Industrial] - Readme Fix

commit a550334e7da4747e50f494173be7929f443b9fd8
Merge: 3d90b6db 7b142a12
Author: DEKHTIARJonathan <[email protected]>
Date:   Tue Mar 19 16:39:14 2019 +0200

    Merge branch 'prep'

commit 7b142a12b92b16085c271568e52b91b445cb18a3
Author: DEKHTIARJonathan <[email protected]>
Date:   Tue Mar 19 16:38:44 2019 +0200

    UNet TF - Readme Typo Fix

commit f0ec1d409211d7f4e0fac6b1533763b94d34c859
Author: Jonathan DEKHTIAR <[email protected]>
Date:   Tue Mar 19 16:36:25 2019 +0200

    [UNet-TF  Industrial] - Readme Fix

commit 7892cbace510ac143582b3fc5a23d3046e745892
Merge: 24db4a2d 3d90b6db
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 15:17:02 2019 +0100

    Merge branch 'master' into prep

commit 3d90b6dbf03242803b374756447bacbcc62220ef
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 15:16:07 2019 +0100

    Updating SSD and BERT.

commit 24db4a2d4a175ba775f23589a3efc23a9f04fd2d
Merge: c5562e04 01609217
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 15:13:27 2019 +0100

    Merge branch 'merge_fix' into prep

commit 016092173331d982369933aa31effbf2460530c7
Merge: 11c52f81 8a3d34e8
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 15:12:10 2019 +0100

    Merge branch 'master' into merge_fix

commit 11c52f819af1ea38dcb9cd62ce208251ca7b8dc4
Merge: 87cec99b 7a8544e0
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 15:09:25 2019 +0100

    Merge commit '7a8544e0a24df6c0fec06c044e5901de63b50ccd' into merge_fix

commit c5562e04da8d064a01258fb0b2108013f2a9e8d4
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 14:12:54 2019 +0100

    SSD scripts with better dataset paths - compliant with dbcluster.

commit a36ea48f02a0c9275fb3e5aed690b46f994edb2f
Author: Przemek Strzelczyk <[email protected]>
Date:   Tue Mar 19 11:19:12 2019 +0100

    BERT update

commit 8a3d34e8a48c2f33485e629703f7f1d6f3637ce3
Merge: b1ae8dd4 f1661d8c
Author: nvpstr <[email protected]>
Date:   Tue Mar 19 11:02:28 2019 +0100

    Merge pull request #13 from sharath-ts/master

    Fix README and Dockerfile

commit f1661d8ca41ff289cb28dc6f8d51868908fdf0e6
Author: sharatht <[email protected]>
Date:   Mon Mar 18 18:06:48 2019 -0700

    Fix README and Dockerfile

commit b1ae8dd47c0e1474741431ad7f0d7a14d6bd6210
Merge: 713d9541 7a8544e0
Author: nvpstr <[email protected]>
Date:   Mon Mar 18 20:57:20 2019 +0100

    Merge pull request #12 from NVIDIA/nvpstr/master

    Adding 9 new models (6 in TensorFlow and 3 in PyTorch)

commit 7a8544e0a24df6c0fec06c044e5901de63b50ccd
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 19:45:05 2019 +0100

    Adding 9 new models: ResNet50/TF, SSD/TF, BERT/TF, NCF/TF, UNet/TF, GNMT/TF, SSD/PyT, MaskRCNN/PyT, Tacotron2+Waveglow/PyT

commit f0c8479eb32c728f209d05af95eb60806746249a
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 19:40:01 2019 +0100

    Hidding unnecessary info

commit 87cec99b020f40ddfa39fc9d0ee536b26dcc08d8
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:32:58 2019 -0700

    Update README.md

commit 2e03c932018b6f62aac01f6e1824135b2a3a09a3
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:32:46 2019 -0700

    Update README.md

commit d7f098d060f34ca75e1923b147c4a919da07b7f9
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:32:37 2019 -0700

    Update README.md

commit 3116b35c504acd05a9fdc159aa91a686188c89cf
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:28:13 2019 -0700

    Update README.md

commit a15916ec621f62b7f362d065ab5159e5b32cd06b
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:27:31 2019 -0700

    Update README.md

commit 578bdf998c094a5ef2068c8e9e8b81580a410317
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:26:38 2019 -0700

    Update README.md

commit c0879118ef6248963802ca5ab56ae34667ee5852
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:25:07 2019 -0700

    Update README.md

commit f814cc535f0505e99474d06dcd4c7d2a8059fb36
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:21:11 2019 -0700

    Update README.md

commit da8981a43059c0074e325e2e55c383f7548848ed
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:17:29 2019 -0700

    Update README.md

commit f2b00c9f3c832f4f2631f265b31b890ad1160420
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:16:16 2019 -0700

    Update README.md

commit 89dbdf89fe2825ca1d8c7e3521daa57bc9e9c0ea
Author: DEKHTIARJonathan <[email protected]>
Date:   Mon Mar 18 20:08:30 2019 +0200

    UNet-TF - Indus: Dockerfile email edit

commit e918e290d475da56220ae8496d97ccf22e7e97ad
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:07:54 2019 -0700

    Update README.md

commit 5e5b6d475ea24f26bb2eb499f8605221382b9c17
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 11:06:59 2019 -0700

    Update README.md

commit f46c264449483fbd731b514d0b52e9274ff7c18c
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 10:51:29 2019 -0700

    Update ssd320_double_4gpus.config

commit 967abbdc9e43d5a9743b9a5e69d5e7ba6a1a0017
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 10:51:01 2019 -0700

    Update ssd320_double_1gpus.config

commit 058523e7a15fb05388c27aa1e0fb5ee7fdd6739e
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 10:50:32 2019 -0700

    Update ssd320_bench.config

commit 5a8ffa2dfb6a999e7c49660fdd08ecd88240829d
Author: Szymon Migacz <[email protected]>
Date:   Mon Mar 18 18:29:04 2019 +0100

    Changed base container paths to publicly available 'nvcr.io/nvidia/...'

commit 2cd1f7ff745449bd083f54468f72a16ee0a0b92e
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 10:23:26 2019 -0700

    Update Dockerfile

commit 76bd5627428d556acbfe39c8b95e662e09afa949
Author: Stephen Jia <[email protected]>
Date:   Mon Mar 18 10:01:55 2019 -0700

    NCF-TF Add option to enable XLA

commit 305775fed4b25c16cd323c4fefa90e5f293c8f39
Author: Grzegorz Karch <[email protected]>
Date:   Mon Mar 18 09:25:32 2019 -0700

    Update README.md

commit 6a1bbb63ed2d5d858134abdb0e9660d90b2084f4
Author: DEKHTIARJonathan <[email protected]>
Date:   Mon Mar 18 18:02:25 2019 +0200

    UNet-TF Indus - Resync Readme Updates

commit 879fae794e9c792b3f4a45dda5190d834b3ba60e
Author: Pawel Soltysiak <[email protected]>
Date:   Mon Mar 18 16:33:25 2019 +0100

    Updated TF BERT readme

commit de524e20596c611f417097efea2fb33933d5e9dc
Author: Slawomir Kierat <[email protected]>
Date:   Mon Mar 18 08:18:04 2019 -0700

    Update README.md

commit c792e176aad784cf0b8b3a0df9f2975f959882f5
Author: Slawomir Kierat <[email protected]>
Date:   Mon Mar 18 08:13:06 2019 -0700

    Update README.md

commit f537dffda4abc1497b53ede66c819d4e37082413
Author: Slawomir Kierat <[email protected]>
Date:   Mon Mar 18 07:41:34 2019 -0700

    Update README.md

commit 0125752e87a865fa5b25c91f3f5f8258a5b272f0
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 15:39:11 2019 +0100

    SSD: Tensorflow README updated

commit dc5eb25a36c40a8b8c7ce90354399d22500cf6ac
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 15:36:13 2019 +0100

    SSD README updated

commit 65a469a61b2f419f24fc84da3cfd4048aaa866de
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 15:32:53 2019 +0100

    SSD: NOTICE added

commit 467b0713577e307979b7e85834691a62e26763e6
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 15:13:20 2019 +0100

    NCF TF - added NOTICE; logger headers updated

commit e16aaaa5da3b1736d5cc13ce38ae7004e158338e
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 14:52:50 2019 +0100

    GNMT - added NOTICE and updated README

commit 83d19f8f4a956285cea1fe728f42390d85adc7e2
Author: Grzegorz Karch <[email protected]>
Date:   Mon Mar 18 06:50:22 2019 -0700

    Update README.md

commit 0ffaa09295642227801cbb387ddd7d185f3e9ddc
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 14:48:59 2019 +0100

    MaskRCNN - added NOTICE and updated README

commit 6e5293e7364bfb568e53291b51c6a9beba669766
Merge: c9b62743 50bfa270
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 14:38:11 2019 +0100

    Merge commit '50bfa270fdc9b6b15289f4aa13a852e31884822d' as 'TensorFlow/Segmentation/UNet_Industrial'

commit 50bfa270fdc9b6b15289f4aa13a852e31884822d
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 14:38:11 2019 +0100

    Squashed 'TensorFlow/Segmentation/UNet_Industrial/' content from commit a70a057

    git-subtree-dir: TensorFlow/Segmentation/UNet_Industrial
    git-subtree-split: a70a057bb7252ca3c3c55e74a9f33483810a6bf4

commit c9b62743d470fd2b1f305856ae9459fd1afaaeb0
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 14:37:48 2019 +0100

    reintroducing UNet

commit 1d059367d553d49b1f4ac399eaf568ae6830f4ae
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 09:50:34 2019 +0100

    Squashed commit of the following:

    commit 2802f5dccc05bdcb21a143a2d244b8b157e335e0
    Author: Przemek Strzelczyk <[email protected]>
    Date:   Mon Mar 18 01:49:12 2019 -0700

        Update README.md

commit d9e84db08e33e54e0b22917bbe60dbbd6967df80
Merge: d72a687a 9a1a3fbf
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 09:45:21 2019 +0100

    Merge branch 'gtc_2019' of https://gitlab-master.nvidia.com/dl/JoC/bert_tf into prep

commit 9a1a3fbfc6ae5c06ee0aec9e2e89c1086abc5202
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 23:41:23 2019 -0700

    modifying path from data_generators to data in readme

commit e75773d8944c57e10da075ccc90a222a5205e070
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 22:44:56 2019 -0700

    adding version number = 19.03

commit 750b8143323c8ac7bfc95cfb177747657bd82f0c
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 22:36:25 2019 -0700

    changing image/sec to sentence/sec

commit eec9ad92a07a32693132cdfbb531eaf92c4b98fc
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 22:26:09 2019 -0700

    Removing speedup information

commit 800396adcc1f24a06577c358326aae6d08934b6c
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 22:10:43 2019 -0700

    fix typo

commit 7189dd55826f918aec1f84d7d31bfd5e43845a4c
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 17 22:08:03 2019 -0700

    adding new read me and public container

commit d72a687a7642602e361b659b1b9371028494fd09
Merge: abe8c025 37b1c879
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 00:59:01 2019 +0100

    Merge branch 'prep' of ssh://gitlab-master.nvidia.com:12051/dl/JoC/examples into prep

commit abe8c02557de8cbabd94545fafa60b136a75260b
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 00:35:40 2019 +0100

    GNMT PyT update (fix in encoder/decoder)

commit 37b1c87954060adda46a0d5e6009771098a3bea6
Author: Przemek Strzelczyk <[email protected]>
Date:   Sun Mar 17 16:14:10 2019 -0700

    Update README.md - clone command

commit db6829f7c65b5866183b590c36fbcaa0de756c0b
Author: Przemek Strzelczyk <[email protected]>
Date:   Mon Mar 18 00:09:04 2019 +0100

    Updating RN50TF

commit 3c3203b117cbc68602d70a468654036ed646497c
Author: Przemek Strzelczyk <[email protected]>
Date:   Sun Mar 17 22:58:17 2019 +0100

    Dockerfile hotfix

commit 9bf7e04b2bf21f17f48195cfdec9b4ec7491830f
Author: Przemek Strzelczyk <[email protected]>
Date:   Sun Mar 17 22:45:46 2019 +0100

    Adding BERT TF, Tacotron2+Waveglow and cleanup

commit 9799d9a42531193b3b4b5bbb67b9112042e02580
Author: Chris Forster <[email protected]>
Date:   Sun Mar 17 10:09:11 2019 -0700

    adding sentence_segmentation_nltk.py from cforster/wiki

commit d18d71cc439b56f1320b2e70d6d4b5ab658a15cb
Author: Chris Forster <[email protected]>
Date:   Sat Mar 16 13:07:18 2019 -0700

    converting relative paths to absolute

commit 740807f8b28922e7090a5bf90e7e82e4ce96fa0e
Author: Chris Forster <[email protected]>
Date:   Sat Mar 16 12:34:36 2019 -0700

    changing relative paths in download helper to absolute

commit d78f811920f9e57be7469e5635d06894dfe6c310
Author: Chris Forster <[email protected]>
Date:   Fri Mar 15 17:07:40 2019 -0700

    silencing output from wikiextractor

commit 256134b23cb796aa36bd5fba94bd8c1c04d69f2d
Merge: ebb7375b a41ed59b
Author: Thor Johnsen <[email protected]>
Date:   Fri Mar 15 16:35:10 2019 -0700

    Clean-up

commit ebb7375be5f69a21195e15ff0704a3eca98095b0
Author: Thor Johnsen <[email protected]>
Date:   Fri Mar 15 16:30:39 2019 -0700

    Clean up

commit 221eea272c2e7ec5f878139b1b67ec8ff7c5b025
Author: Przemek Strzelczyk <[email protected]>
Date:   Sat Mar 16 00:15:53 2019 +0100

    Adding new models

commit a41ed59b8e69f6b8991cbf4b8894aaf78c86a18c
Merge: 4f81c23b 0e01bdd8
Author: Chris Forster <[email protected]>
Date:   Fri Mar 15 15:58:45 2019 -0700

    Merge branch 'gtc_2019' of ssh://gitlab-master.nvidia.com:12051/dl/JoC/bert_tf into gtc_2019

commit 4f81c23b73eeb059332c94487e6dfd03977ebc93
Author: Chris Forster <[email protected]>
Date:   Fri Mar 15 15:58:40 2019 -0700

    fixing path issue

commit 0e01bdd860bf07a981786b98ea12fc38c28d6b69
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 15 14:59:43 2019 -0700

    changing launch file to suit pretraining

commit eb164940ad74c2a6453099de58ce344f970adc7b
Author: Thor Johnsen <[email protected]>
Date:   Fri Mar 15 12:29:05 2019 -0700

    Fix interactive pretraining script

commit 713d95411a10b5516a4e2562a68bbdaad1eaea1b
Merge: b0509634 7c10b43d
Author: nvpstr <[email protected]>
Date:   Fri Mar 15 17:31:47 2019 +0100

    Merge pull request #11 from tgrel/master

    Update NCF-PyTorch README

commit 413d9936c8099523bd8a2d278e6eefa6bca7d55d
Author: Thor Johnsen <[email protected]>
Date:   Thu Mar 14 20:07:53 2019 -0700

    Pretraining scripts for JoC

commit f905c118a3dd50dc26ce6745e8f05be8c959a9cc
Author: Chris Forster <[email protected]>
Date:   Thu Mar 14 16:35:17 2019 -0700

    Data script updates, adding eval set generation

commit debe0af125ba878cda168652e27835d83c909e1d
Author: Swetha Mandava <[email protected]>
Date:   Thu Mar 14 11:28:39 2019 -0700

    data script fixes, default hparams for pretraining

commit b2709f83643de1efd00031181f709b5ee6ab67ef
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 21:38:47 2019 -0700

    adding installs to support data download

commit beae6ef703048e3bba712d6eb9346554c827d972
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 21:16:05 2019 -0700

    nvidian container for qa, pretraining script

commit 0f4d8a8b6c2280d2775a58e616fe9b558217386e
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 19:00:04 2019 -0700

    data download from within container

commit 344574b7d5f4713b4b486ea12cedcb2f6141c869
Merge: 968b3568 a1a9a3a3
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 18:14:39 2019 -0700

    merging conflicts

commit 968b3568f1834d73cbf11cbb4da853bb58c24afb
Author: Thor Johnsen <[email protected]>
Date:   Wed Mar 13 19:43:50 2019 -0500

    Don't multiply learning rate by hvd.size(), it confuses users

commit a1a9a3a3eaf7e02dcc49def9ab3209c789564f32
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 17:29:59 2019 -0700

    Update Dockerfile

commit cc6d1cc6d6f8b93085aa51d2969a059c2c4398fc
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 17:07:29 2019 -0700

    clean up

commit d4017139c4b6a914ed7823c9dbbb7434426878bc
Author: Thor Johnsen <[email protected]>
Date:   Wed Mar 13 19:01:31 2019 -0500

    Fix bug in way learning rate is computed

commit 4a795fe20bca71d21b71c1ba663013c9cb034ea4
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 16:58:33 2019 -0700

    fix data script

commit f2b86436b8df25f0237d7cb6457ddf467d05688d
Author: Swetha Mandava <[email protected]>
Date:   Wed Mar 13 16:53:56 2019 -0700

    cleanup + adding data scripts

commit 5f55072542298ea6dbc044b885ee82cd0d681032
Author: Thor Johnsen <[email protected]>
Date:   Wed Mar 13 18:53:37 2019 -0500

    Add --amp support with model level loss scale optimizer

commit f33174c3fd1c9b8544cf9216a8425106036e1ba9
Author: Swetha Mandava <[email protected]>
Date:   Tue Mar 12 13:17:26 2019 -0700

    squad datadownload script

commit 423c89b8f94403c6cb88f318cc48ee47163eb492
Author: Swetha Mandava <[email protected]>
Date:   Mon Mar 11 20:43:27 2019 -0700

    moving to 19.03, adding loss scaler, joc scripts

commit c54110993ebf855344a0b9b8990cc447dc47edfe
Merge: 323eb377 1a866120
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 16:09:02 2019 -0800

    Merge branch 'revert-323eb377' into 'gtc_2019'

    Revert "Merge branch 'smandava/devel' into 'gtc_2019'"

    See merge request dl/JoC/bert_tf!7

commit 1a866120b1c02ef00394e6f1d93cd6ffe25172d5
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 16:08:48 2019 -0800

    Revert "Merge branch 'smandava/devel' into 'gtc_2019'"

    This reverts commit 323eb3770dea6921e771464bea7885a6c0870762

commit c8e67bf57aaf905cbe895d92329e3e5faab99d80
Merge: 8c37e461 19a194a3
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 16:02:03 2019 -0800

    Merge branch 'smandava/devel' of ssh://gitlab-master.nvidia.com:12051/dl/JoC/bert_tf into smandava/devel

commit 8c37e46117dc003bb6baccac32497f925e376c5b
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 15:58:53 2019 -0800

    adding training hooks for estimator stats

commit 19a194a3c26dbd2e8f8764493fd42c5d96ad80ff
Merge: 88524a16 323eb377
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 15:55:36 2019 -0800

    Merge branch 'gtc_2019' into 'smandava/devel'

    Merging gtc_2019 into dev

    See merge request dl/JoC/bert_tf!6

commit 323eb3770dea6921e771464bea7885a6c0870762
Merge: cf819aee 88524a16
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 8 15:55:26 2019 -0800

    Merge branch 'smandava/devel' into 'gtc_2019'

    # Conflicts:
    #   Dockerfile
    #   modeling.py
    #   optimization.py
    #   run_pretraining.py
    #   run_squad.py

commit cf819aee59bfc37a41ca4cfcf134c3b43c35e9f9
Author: Thor Johnsen <[email protected]>
Date:   Thu Mar 7 17:49:09 2019 -0800

    Add support for AMP+FM

commit bb57786f268392163cc24f9c7c6c0069a52ad1c4
Author: Thor Johnsen <[email protected]>
Date:   Tue Mar 5 14:04:34 2019 -0800

    Bug fix in optimizer for fastmath

commit 88524a16a5a430a32132118b0b6b97c57dd70063
Author: Swetha Mandava <[email protected]>
Date:   Mon Mar 4 17:23:27 2019 -0800

    run_squad on cpu

commit f5fa9f73ab9d685be817a5f36b2998c30b1de02d
Author: Thor Johnsen <[email protected]>
Date:   Mon Mar 4 19:30:10 2019 +0000

    Bug fix, change fasthmath to fastmath

commit 38451d2ea59379af692b744376086adf567643df
Author: Swetha Mandava <[email protected]>
Date:   Mon Mar 4 10:05:13 2019 -0800

    adding optional hvd

commit 0a8e6f0f784e55f823c520b53e2a9f96de4f865f
Author: Thor Johnsen <[email protected]>
Date:   Mon Mar 4 17:54:28 2019 +0000

    Revert to fp32 allreduce

commit 2dd11fa7b601956f293f0765dbbd8bda96b0993f
Author: Thor Johnsen <[email protected]>
Date:   Mon Mar 4 09:12:34 2019 -0800

    Bug fix for fastmath

commit 11934ca85ed8bc7755b9ad5f5199b2aaaa0e7c3b
Author: Swetha Mandava <[email protected]>
Date:   Mon Mar 4 01:17:38 2019 -0800

    fix num_gpus

commit 4e5f9ab213e05f98038f9233b8c2ea1bc4f4bb61
Merge: cf8b3210 af84120a
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 3 22:57:55 2019 -0800

    adding xla, num_gpu options to script

commit cf8b32109b4d9446a2d8e5c3ccda9c306331e21b
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 3 22:56:24 2019 -0800

    adding xla, num_gpu options to script

commit af84120a90897704d55b307c2e62543f69cf746e
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 3 22:43:03 2019 -0800

    adding xla, num_gpu variables

commit 774909b12d3145850ac27ce7ee40fc8c9d13480c
Author: Swetha Mandava <[email protected]>
Date:   Sun Mar 3 21:54:05 2019 -0800

    dividing tf.train_record between gpus

commit 0899d306be74a0f6b33b6542a1471c1b712c3638
Author: Thor Johnsen <[email protected]>
Date:   Sat Mar 2 20:44:59 2019 -0800

    Set environment vars for fastmath

commit 97c2dcd3821d7883c4e9623d323bb30e610a0417
Author: Thor Johnsen <[email protected]>
Date:   Sat Mar 2 20:20:45 2019 -0800

    Add --fastmath

commit 75e0715d44deaab26d4d8cc06d4782bd37389a02
Author: Swetha Mandava <[email protected]>
Date:   Sat Mar 2 10:59:13 2019 -0800

    ignoring corrupt record in dataset

commit 3277cf1ce19cf4eaf8720f0087eb429227e38fea
Author: Swetha Mandava <[email protected]>
Date:   Fri Mar 1 14:30:06 2019 -0800

    Adding horobod support to squad

commit d2fc37cf025b0a4cb7b002ff2596e5362117d5ef
Merge: 5e1836bf d5d18888
Author: agray <[email protected]>
Date:   Thu Feb 28 17:05:00 2019 -0800

    Merge branch 'gtc_2019' of https://gitlab-master.nvidia.com/dl/JoC/bert_tf into gtc_2019

commit 5e1836bfe11dbe1eadf60a8801545630035e98a1
Author: agray <[email protected]>
Date:   Thu Feb 28 17:04:39 2019 -0800

    fixed a few typos, add np 1 to run_pretraining.sh -one pid all gpus- likely change this later

commit d5d18888a39b4b80a8173206a239395b0710ad83
Author: chris <[email protected]>
Date:   Thu Feb 28 13:50:39 2019 -0800

    removing some debugging printing

commit 04f31b9fedcfd5cf3c9d14ef16c7e8ecea1410d6
Author: chris <[email protected]>
Date:   Thu Feb 28 13:39:10 2019 -0800

    updating Dockerfile

commit 3aea0ba1e7bb271eaf21a3483cd8f6236a4cbeaf
Author: chris <[email protected]>
Date:   Thu Feb 28 13:25:53 2019 -0800

    Fixing gpu printout in run.sub for TF (previously pyt)

commit f185939ae0e214159a519d662b77e35d8801036d
Author: chris <[email protected]>
Date:   Thu Feb 28 12:46:21 2019 -0800

    adding launchers for convenience

commit fbd36d1b83df57966eafc2a50a2f8630bc3137ca
Author: chris <[email protected]>
Date:   Thu Feb 28 12:41:02 2019 -0800

    Adding remaining scripts

commit ebfa8b1a79619b9875fd28681166d2ecdb291863
Author: chris <[email protected]>
Date:   Thu Feb 28 12:39:24 2019 -0800

    Initial commit to add prom scripts

commit 4a6894379cab9200054c1903e4ef5cba120ac3d6
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 27 19:23:51 2019 -0800

    Custom dockerfile for gtc 2019

commit 9cbf96f3174ef975018f5c30e3a4fb153308e776
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 27 19:12:19 2019 -0800

    Dockerfile to build custom container for GTC 2019

commit cccab2e79ebc01cd835e16c92a12916513c23690
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 27 19:00:28 2019 -0800

    Add vocab, config file for BERT large

commit 090e5be2ed5270daf1f452a994884800f98c7fb5
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 27 18:18:46 2019 -0800

    Filter out none grads before handing them over to optimizer

commit c15647210993c4a592eb68503464eed5540caf58
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 27 23:16:29 2019 +0000

    Add fp16, xla support to classifer, squad tasks

commit 93bb076ae0a7a5c248887b8c9c648d00069f6776
Author: Thor Johnsen <[email protected]>
Date:   Tue Feb 26 17:40:31 2019 -0800

    Minor bug fix for FP32

commit a86b8864e9b16282f6fe9bb71eb335d84c9ee272
Author: Swetha Mandava <[email protected]>
Date:   Tue Feb 26 11:57:53 2019 -0800

    scripts for amp and fast_math with automatic loss scaling

commit 7de51c020a813bc3453ad0ebf101760d63163cc9
Author: Swetha Mandava <[email protected]>
Date:   Sat Feb 23 18:59:07 2019 -0800

    modifying scripts for new container

commit 1e0e672a58b2ef969812c7434b9d60576fde27cf
Author: Thor Johnsen <[email protected]>
Date:   Thu Feb 21 23:49:03 2019 +0000

    Add dynamic loss scaling for mixed precision

commit 7c10b43d3ecf79ed4ddc04535c955d16fd3a06f1
Author: Tomasz Grel <[email protected]>
Date:   Fri Feb 15 17:29:12 2019 +0100

    update NCF README

commit 71a09b63cace143598b142bdbc91987c6992a7a9
Author: Swetha Mandava <[email protected]>
Date:   Wed Feb 20 17:48:45 2019 -0800

    adding fast_math and amp support to squad

commit b05096344433b098b5c29d8fd246195ec05b6389
Merge: b99d45b3 8efbea40
Author: Szymon Migacz <[email protected]>
Date:   Thu Feb 14 12:40:30 2019 +0100

    GNMT update

commit 8efbea403f06c0033c566e23d04302226b536286
Author: Szymon Migacz <[email protected]>
Date:   Thu Feb 14 12:40:30 2019 +0100

    Squashed 'PyTorch/Translation/GNMT/' changes from 4dc145a..51a90b1

    51a90b1 Feb 14, 2019 update

    git-subtree-dir: PyTorch/Translation/GNMT
    git-subtree-split: 51a90b1667e7c1d45bd68bf719f8bf5d4c4521e3

commit b99d45b3f53a7904bd334b1757ba2f0731ae98ff
Author: Szymon Migacz <[email protected]>
Date:   Thu Feb 14 12:39:20 2019 +0100

    Revert "Updating GNMT"

    This reverts commit 072523d5b21444a97b82876bf4c6d1d3e36ba255.

commit 8f16e4d4bf15c58f8382e8aba074970124e56f79
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 13 12:45:00 2019 -0800

    Relying on upstream having fixed layer norm epsilon for fp16

commit 02102dba5529d74095111e7387a151817a0fdce8
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 13 13:16:26 2019 -0600

    FP16 convergence

commit 62bf94756a40c1e26c5914c0f40d1686a4467dc3
Author: Thor Johnsen <[email protected]>
Date:   Wed Jan 16 17:26:08 2019 -0800

    gpu_environment.py not needed anymore

commit f9e6f6bf6186847fae577a824e76946cd671ea44
Author: Thor Johnsen <[email protected]>
Date:   Wed Jan 16 17:24:01 2019 -0800

    Change epsilon from 1e-12 to 1e-4 to fit FP16 tensors

commit 25d135168a36e5d341b8c4f6b2b0cbe4b98bdf1a
Author: Thor Johnsen <[email protected]>
Date:   Wed Jan 16 17:20:44 2019 -0800

    Update README.md with instructions for new GPU options:

commit 4bfc69263a433541b10286333e947a815eba079a
Author: Thor Johnsen <[email protected]>
Date:   Wed Jan 16 17:16:58 2019 -0800

    Multi-gpu support, XLA patches

commit 61b0cb69809a93ae95c1066c1e89599d8c7fba7c
Author: Thor Johnsen <[email protected]>
Date:   Thu Dec 20 21:29:49 2018 +0000

    Less intrusive changes

commit b425e39807c3af28b4e5bcdf67e59892525eb7b0
Author: Thor Johnsen <[email protected]>
Date:   Mon Dec 17 09:10:40 2018 -0800

    Fix small bug

commit c0fdc71b8d7f2d84a86d7e560ed19382fa01fbe1
Author: Thor Johnsen <[email protected]>
Date:   Thu Dec 13 21:28:17 2018 -0800

    XLA the whole model

commit b72423b044c2da553606224394ceb54252404302
Author: Thor Johnsen <[email protected]>
Date:   Tue Dec 11 19:34:41 2018 -0800

    Set loss_scale to 1.0 when use_fp16 is False

commit b4a912ce920f708b8041c4a56bae415e15427bd8
Author: Thor Johnsen <[email protected]>
Date:   Tue Dec 11 19:25:11 2018 -0800

    Add support for fp16 math and xla optimizations

commit ffbda2a1aafe530525212d13194cc84d92ed0313
Merge: 0a0ea64a 3084a390
Author: Jacob Devlin <[email protected]>
Date:   Tue Feb 12 14:17:23 2019 -0800

    Merge pull request #431 from dalequark/colab-tfhub

    Changed colab link

commit 3084a3908c486085cbbabfc68f0a16febeaaa59f
Author: dalequark <[email protected]>
Date:   Tue Feb 12 14:09:10 2019 -0800

    Changed colab link

    Fixed the colab notebook link.

commit 0a0ea64a3ac1f43ed27d75278b9578708f9febcf
Merge: bee6030e 46ec4d8d
Author: Jacob Devlin <[email protected]>
Date:   Tue Feb 12 14:06:08 2019 -0800

    Merge pull request #430 from dalequark/colab-tfhub

    Added bert tfhub example on colab

commit 46ec4d8d8f1a35fc5bf9cd5568f6db6e4ee0ca37
Author: Dale Markowitz <[email protected]>
Date:   Tue Feb 12 13:42:37 2019 -0800

    Update README.md

    Added colab link in the readme

commit 55cb665e4016bdf9b0e98efed09ecbb5a1d8f70d
Author: Dale Markowitz <[email protected]>
Date:   Tue Feb 12 13:30:36 2019 -0800

    Ready-to-run colab tutorial on using BERT with tf hub on GPUS

commit 802057d73dfd41dc77af8fea1f3210f69c1de10a
Author: Thor Johnsen <[email protected]>
Date:   Tue Feb 12 10:00:09 2019 -0800

    Add MLM Loss and NSP Loss to report

commit bee6030e31e42a9394ac567da170a89a98d2062f
Author: Jacob Devlin <[email protected]>
Date:   Thu Feb 7 13:16:36 2019 -0800

    Adding TF Hub support

commit 4ce8413b21073eb62e5fc452caa48aabb5301825
Author: Thor Johnsen <[email protected]>
Date:   Wed Feb 6 13:18:44 2019 -0800

    Towards convergence for fp16

commit 072523d5b21444a97b82876bf4c6d1d3e36ba255
Author: nvpstr <[email protected]>
Date:   Fri Jan 25 16:20:05 2019 +0100

    Updating GNMT

commit a1aff31c42129e308db455016701db2433d5f752
Author: nvpstr <[email protected]>
Date:   Fri Jan 25 16:02:48 2019 +0100

    Adding training improvements.

commit a8e1d1cd83ec7b40895f12b07bbaeaa322197df7
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Jan 23 17:14:51 2019 +0100

    Adding ResNet50v1.5 to MxNet/Classification

commit edc3486e6fe62a79046ee7c746f442c95e92eed4
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Jan 23 17:08:35 2019 +0100

    Small fix to README

commit d755a84080f23b34aa5c0c95755c7060a2c9a863
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Jan 23 17:03:45 2019 +0100

    Adding Transformer to PyTorch/Translation

commit 585973766fa474bb7113646e563c4da5b7aa2918
Author: Przemek Strzelczyk <[email protected]>
Date:   Wed Jan 23 16:59:07 2019 +0100

    Adding NCF to PyTorch/Recommendation

commit f39e881b169b9d53bea03d2d341b31707a6c052b
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 15:52:39 2018 -0800

    Padding examples for TPU eval/predictions and checking case match

commit b8ba348c4e91f886d0cd8724fd9f2a4edf382101
Merge: c74caa32 284945c2
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:40:44 2018 -0800

    Merge pull request #266 from msramalho/patch-1

    Tokenization code simplification

commit c74caa32955485050ce0d0f79043052648e01ffc
Merge: 95beaf39 0bf3c6e4
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:36:05 2018 -0800

    Merge pull request #214 from hsm207/patch-1

    Fix  BERT acronym to be consistent with paper

commit 95beaf3935129a4aebed61b5eb4afeb19cceb642
Merge: 7f51d420 00bea233
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:34:51 2018 -0800

    Merge pull request #209 from leoxzhao/typofix

    Fixed some typos in comments.

commit 7f51d420e5bb33b3e89db962f588b074dffbf118
Merge: 85f453a6 933848ac
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:29:23 2018 -0800

    Merge pull request #270 from ywkim/typo

    Fix typo in optimization.py

commit 85f453a6ebbb1354f5a54d15fcee779e2acdfb63
Merge: 4be9723c 738b648a
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:29:11 2018 -0800

    Merge pull request #252 from tianxin1860/master

    fix typo

commit 4be9723cf0d95d02fa2182a0dec0e68bad7ed99f
Merge: 4a47cc2d 094e9de0
Author: Jacob Devlin <[email protected]>
Date:   Tue Dec 18 10:25:07 2018 -0800

    Merge pull request #274 from soloice/master

    FIX: Python3 compatibility.

commit 094e9de02dcc9ba782ac4702d08f54eb0e64e84e
Author: soloice <[email protected]>
Date:   Sun Dec 16 01:50:37 2018 +0800

    FIX: Python3 compatibility.

commit 933848ac2786a918209861d7eadf108530288ce6
Author: Youngwook Kim <[email protected]>
Date:   Sat Dec 15 07:47:32 2018 +0900

    Fix typo in optimization.py

commit 284945c2ea518cb7d83b2d8023ea3acec630f238
Author: Miguel Sozinho Ramalho <[email protected]>
Date:   Thu Dec 13 15:15:33 2018 +0000

    Tokenization code simplification

commit 738b648a4519e589997e88b78eb41dfdae5aaa68
Author: TianXin <[email protected]>
Date:   Tue Dec 11 17:26:53 2018 +0800

    fix typo

commit 0bf3c6e4f5c719df5c0088870fa596f41b7267ea
Author: hsm207 <[email protected]>
Date:   Sun Dec 2 13:46:16 2018 +0800

    Fix  BERT acronym to be consistent with paper

commit 00bea233aa32def2620b640cfcc7e9082cc980bc
Author: Leo Zhao <[email protected]>
Date:   Fri Nov 30 23:12:41 2018 -0500

    Fixed some typos in comments.

commit 4a47cc2da23dcb4ab1bf5c08085910b6fd94a4cf
Merge: d4526127 ad88ff1b
Author: Jacob Devlin <[email protected]>
Date:   Tue Nov 27 17:26:49 2018 -0800

    Merge pull request #170 from imcaspar/patch-1

    Fix typo in README.md

commit d45261273d2dc27e0d25b1cf9e3adb312545a236
Author: Jacob Devlin <[email protected]>
Date:   Tue Nov 27 17:22:44 2018 -0800

    Putting nampledtuple in global scope

commit ad88ff1b140d5695c5c7233543c62f489dd8cec3
Author: Caspar ZHANG <[email protected]>
Date:   Mon Nov 26 11:43:01 2018 +0800

    Fix typo in README.md

commit 68d50765f613bb24a8884cb579a50ee3b2c48471
Author: Jacob Devlin <[email protected]>
Date:   Sat Nov 24 16:52:26 2018 -0800

    Adding BERT-Large cased

commit 32ac4763a58a050c7853b36458eeb34544b4f64d
Merge: a9ba4b8d f76bd0f7
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 23 19:11:17 2018 -0800

    Merge pull request #167 from eric-haibin-lin/patch-2

    Fix typo

commit a9ba4b8d7704c1ae18d1b28c56c0430d41407eb1
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 23 17:11:27 2018 -0800

    Clarifying readme about Chinese space tokenization

commit 332a68723c34062b8f58e5fec3e430db4563320a
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 23 17:02:39 2018 -0800

    Adding new multilingual model

commit f76bd0f72402121a8ab6d5742d1a92213deae741
Author: Haibin Lin <[email protected]>
Date:   Fri Nov 23 16:20:48 2018 -0800

    Fix typo

commit 1cd50d7aa7d26638e05eb5b8a282b30cc92c6eb6
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 23 16:18:08 2018 -0800

    Updating formatting

commit 216d92a89eb3f896b3a2b22c3f453fdea1a3b75d
Merge: 4d4e26dc b32ebca9
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 21 15:11:02 2018 -0800

    Merge pull request #158 from zhaoyongke/master

    Fix squad 2.0 dev data link

commit b32ebca969e99f3cd0b2f1d59cd1adf2d883b14a
Author: 卜居 <[email protected]>
Date:   Thu Nov 22 00:29:44 2018 +0800

    Fix squad 2.0 dev data link

commit 4d4e26dc24bf78d849104e646dba422e3adba458
Merge: c8757ce5 d9ec1447
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 16 11:29:55 2018 -0800

    Merge pull request #137 from JasonJPu/master

    Fix squad 2.0 command to use 2.0 train and dev files

commit d9ec1447d45e91ab8906211c770f801dee28dfa6
Author: JasonJPu <[email protected]>
Date:   Fri Nov 16 11:21:31 2018 -0800

    Fix squad 2.0 command

commit c8757ce50b5fda15f17cf5010ca8ed95c8c45ba1
Merge: 60454702 7bebe59c
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 16 08:35:54 2018 -0800

    Merge pull request #132 from pengli09/fix-boundary-checking-bug

    fix boundary checking bug

commit 7bebe59c8b4b2583ff8b3662cb9f1cd73f163a77
Author: Peng Li <[email protected]>
Date:   Fri Nov 16 20:10:01 2018 +0800

    fix boundary checking bug

commit 60454702590a6c69bd45c5d4258c7e17b8a3e1da
Author: Jacob Devlin <[email protected]>
Date:   Thu Nov 15 15:39:59 2018 -0800

    Adding SQuAD 2.0 support

commit 9d81f96d6b9f9fb6aa97c2298e1e77cfe5899e09
Merge: 35deae77 075828de
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 14 09:33:02 2018 -0800

    Merge pull request #120 from qwfy/master

    Define initialized_variable_names even if init_checkpoint is None

commit 075828de113360001f9368c7ca95984fa3c5c6f2
Author: Incomplete <[email protected]>
Date:   Wed Nov 14 15:40:40 2018 +0800

    Define initialized_variable_names even if init_checkpoint is None

commit 35deae77e6b6c668cc8050853d16b4046193cbf0
Merge: 3c67c1de ff7d05af
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 9 14:32:35 2018 -0800

    Merge pull request #93 from artemisart/fix-test

    fix tokenization_test

commit ff7d05aff5c81682fdc8afd6fc5ce6a43fd5c57f
Author: Mathis Chenuet <[email protected]>
Date:   Fri Nov 9 17:11:03 2018 +0100

    fix test

commit 3c67c1de0b3b5e0c9e43e7cd8653f05ec2f844ce
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 9 08:00:41 2018 -0800

    Running through pyformat to meet Google code standards

commit aefad127db7dbd8c1bb1b47aa6d530912fbb442f
Merge: 870e56f2 2d7a34b6
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 9 07:24:46 2018 -0800

    Merge pull request #79 from BogdanDidenko/backward_convert

    Add method for converting ids to tokens.

commit 2d7a34b6a30c2603174884a49882ed881f8a608f
Author: Bogdan Didenko <[email protected]>
Date:   Fri Nov 9 11:39:11 2018 +0200

    return API method "convert_tokens_to_ids(vocab, tokens)" and add  "convert_ids_to_tokens(inv_vocab, ids)"

commit 870e56f203f53d19235c1bc94814ecfea1af9975
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 16:06:27 2018 -0800

    Fixing comment

commit 1da14bd918c3c94b1a50958b42de470ad39428f3
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 16:04:32 2018 -0800

    Fixing bug introducing in classification

commit ea8737bdd9dd24c0260910b8d8f5a0fb5789fa34
Author: Bogdan Didenko <[email protected]>
Date:   Wed Nov 7 23:41:34 2018 +0200

    Add method for converting ids to tokens.

commit 803d02590a62e861020b482ed878aec928f67fad
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 12:16:25 2018 -0800

    Update README to clarify feature extraction

commit 32e04fedd6e2e55b0fbd0120eadd20fa908a9c28
Merge: b47db614 b545f6f2
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 10:51:03 2018 -0800

    Merge pull request #68 from cbockman/patch-1

    fix to model variable scope

commit b47db614538fe22f122470de21af9dd2da7993e7
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 10:49:50 2018 -0800

    Updating chainer FAQ

commit 8fbff635d3fc860f1adac343148d60fdfdd96833
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 10:40:40 2018 -0800

    Fixing CoLA predict mode

commit 2930b5f5222b3193ae32428def4a4429d0c23912
Merge: a6fbc509 08509ea5
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 10:33:46 2018 -0800

    Merge pull request #76 from zrcahyx/master

    fix label_map key error in function `convert_single_example` in `run_classifier.py` when doing prediction

commit a6fbc50983f5e318480601631e534edfd0072377
Author: Jacob Devlin <[email protected]>
Date:   Wed Nov 7 10:30:56 2018 -0800

    Fixing python2 logging for extract_features.py

commit 08509ea5f5bf67fe6793a5322a195d951efc6de0
Author: Ruchen Zhang <[email protected]>
Date:   Wed Nov 7 16:31:49 2018 +0800

    fix predict example label map error

commit a21d4848ec33eca7d53dd68710f04c4a4cc4be50
Merge: b443ca91 ee24fa50
Author: Jacob Devlin <[email protected]>
Date:   Tue Nov 6 16:12:08 2018 -0800

    Merge pull request #69 from aijunbai/fix

    fixed predict mode APIs miss-usages

commit ee24fa50e3ad271ca6c38945956808cdf3e6e5b3
Author: Aijun Bai <[email protected]>
Date:   Tue Nov 6 16:00:23 2018 -0800

    fixed predict mode APIs miss-usages

commit b545f6f280614121df85556c54782badf3007cfa
Author: cbockman <[email protected]>
Date:   Tue Nov 6 14:28:49 2018 -0800

    Update modeling.py

commit a22de7ec24ba873f699e989533122f36b4f6f693
Author: cbockman <[email protected]>
Date:   Tue Nov 6 14:25:42 2018 -0800

    fix to model variable scope

commit b443ca91c6d1b3766f1b21e3ecf83303aa144c84
Merge: 99baaf74 98268e3f
Author: Jacob Devlin <[email protected]>
Date:   Tue Nov 6 12:53:26 2018 -0800

    Merge pull request #56 from abhishekraok/support_predictor

    Add predictor for fine tuned classifier

commit 98268e3f232089c0a6ee607a6e2f31c32372380b
Author: Abhishek Rao <[email protected]>
Date:   Tue Nov 6 10:47:05 2018 -0800

    Rename inference mode inputs from predict to test

commit 99baaf7405e5a3b504e7bdf0e0e19d3ab5980a75
Author: Jacob Devlin <[email protected]>
Date:   Tue Nov 6 10:43:12 2018 -0800

    Fixing typo filed -> file

commit 41d6111d343d8108815786adb6cae594dbc6380a
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 14:19:21 2018 -0800

    Match indentation of the project

commit 9126c8b8b565a62beeb5063a9c223aeee5b59f05
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 14:16:59 2018 -0800

    Fix comments and remove unused variable

commit 970f9340ad8fba3df4ceaab6c19864768217de6f
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 14:11:46 2018 -0800

    Fix function keyword for EstimatorSpec

commit 0c1b415757ab4b09f4c1009eef06ed242b4b0297
Merge: 1724840c 5e10be4c
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 13:35:39 2018 -0800

    Merge branch 'support_predictor' of https://github.com/abhishekraok/bert into support_predictor

commit 1724840c624a8693c754e902ca0c72722cc96f40
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 13:35:31 2018 -0800

    Change predict output to tsv

commit 5e10be4cd23468946a66aba31ef0e2ef77b21331
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 13:34:49 2018 -0800

    Add info about predict mode

commit e13c1f3459cc254f7abbabfc5a286a3304d573e4
Author: Jacob Devlin <[email protected]>
Date:   Mon Nov 5 12:19:16 2018 -0800

    Fixing typo in function name and updating README

commit 2f82f216478e8b90de2040f02a18d5501c79d34e
Merge: ee615511 f10ab0b9
Author: Jacob Devlin <[email protected]>
Date:   Mon Nov 5 11:19:05 2018 -0800

    Merge pull request #52 from rodgzilla/typo_correction

    typo correction.

commit 6901202b62890187beb78696463081d65d9b1a3c
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 10:52:59 2018 -0800

    Add get_predict_examples functions to data processors

commit c472857e31caad33c8c98f0ae8a2af142caf9889
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 09:56:32 2018 -0800

    Add prediction function

commit ee6155110f170c1d31e09ed46a4e468a7680f063
Author: Jacob Devlin <[email protected]>
Date:   Mon Nov 5 09:47:37 2018 -0800

    Updating run_classifier.py to be backwards compatible

commit e6b57e13388594c1db02cf8a6ccea0237e10298c
Merge: bd515960 04f9e5d1
Author: Abhishek Rao <[email protected]>
Date:   Mon Nov 5 09:28:53 2018 -0800

    Merge branch 'master' of https://github.com/abhishekraok/bert

commit f10ab0b9fee8aa524d23a02af9a830c0636f6041
Author: Grégory Châtel <[email protected]>
Date:   Mon Nov 5 15:35:02 2018 +0100

    typo correction.

commit bd515960d8f1db4f8a819356af25ad4d82627bd2
Author: Jacob Devlin <[email protected]>
Date:   Sun Nov 4 17:56:28 2018 -0800

    Clarifying multilingual README

commit c175dd0e655bc227f6fe02ac87ef7d78f33ce678
Author: Jacob Devlin <[email protected]>
Date:   Sun Nov 4 15:24:18 2018 -0800

    Fixing typo

commit 66a81a655f3aa463b269c284331f492471322abf
Author: Jacob Devlin <[email protected]>
Date:   Sun Nov 4 12:58:26 2018 -0800

    Fixing multilingual table

commit f18bd94b8fee9bda3c293e0932d100add35b780c
Author: Jacob Devlin <[email protected]>
Date:   Sun Nov 4 12:48:46 2018 -0800

    Adding support for multilingual models

commit a4dc5daf72568628f27d1d920db976ace50d3a8d
Author: Jacob Devlin <[email protected]>
Date:   Fri Nov 2 15:51:56 2018 -0700

    Improving run_squad.py flag validation

commit 04f9e5d15d85f85a8deff9f53442b8a54abab5ea
Merge: 9c8798fb 4e2e06ab
Author: Abhishek Rao <[email protected]>
Date:   Fri Nov 2 15:50:52 2018 -0700

    Merge pull request #1 from google-rese…

Enable the option for FusedLAMB and FusedAdam.

Add pytorch examples

Upstream changes

update master

Update the codes