Commits · 1567c2dc88883a4876a0ef6ad6e1014f6211d9da · Shashank Suhas / seminar-breakout

15 Sep, 2018 1 commit
- Let LocallyShuffleData not depend on __len__ · 1567c2dc
  Yuxin Wu authored Sep 15, 2018
  
  1567c2dc
13 Sep, 2018 1 commit
- _init(locals()) set everything except for self and privates. fix #893 · 68b8a7b7
  Yuxin Wu authored Sep 13, 2018
  
  68b8a7b7
11 Sep, 2018 3 commits
- fix bug in last commit. · 4a917d6e
  Yuxin Wu authored Sep 11, 2018
  
  4a917d6e
- before_train option in PeriodicTrigger · 2cf9ad75
  Yuxin Wu authored Sep 11, 2018
  
  2cf9ad75
- fix SaverRestoreRelaxed · ba965954
  Yuxin Wu authored Sep 11, 2018
  
  ba965954
09 Sep, 2018 1 commit
- fix setup.cfg (fix #889) · 2aefaf78
  Yuxin Wu authored Sep 09, 2018
  
  2aefaf78
04 Sep, 2018 3 commits
- make the pix2pix support multi gpu trainning (#884) · 8250786f
  yogurfrul authored Sep 05, 2018
```
* make the pix2pix support multi gpu trainning

* Update Image2Image.py

* Update Image2Image.py

* Update Image2Image.py
```
  8250786f
- fix wrong use of len(data) in examples · b8c7b6f4
  Yuxin Wu authored Sep 04, 2018
  
  b8c7b6f4
- update docs; bump version · 193f6056
  Yuxin Wu authored Sep 04, 2018
  
  193f6056
01 Sep, 2018 2 commits
- fix build · 65c0b326
  Yuxin Wu authored Sep 01, 2018
  
  65c0b326
- update docs; Mask R-CNN horovod mode eval only on master machine · f63e0ee4
  Yuxin Wu authored Sep 01, 2018
  
  f63e0ee4
31 Aug, 2018 3 commits

make dataflow idiomatic python container objects (fix #869) (#872) · 7b8728f9

Patrick Wieschollek authored Aug 31, 2018

* make dataflow idiomatic python container objects (fix #869)

* use ABCMeta and fix tqdm

* python2 compatible version

* fix

* some more replacement

* rollback changes of _reset_df_and_get_size

* fix unrelated pbar issue

* revert the __len__ changes on InputSource

7b8728f9

update docs · 3476cb43
Yuxin Wu authored Aug 31, 2018

3476cb43

use svg (#881) · 45d63caf

Patrick Wieschollek authored Aug 31, 2018

* use svg

I can't look at it anymore.

* Update README.md

use gitter brand color #ed1965 or #46bc99

45d63caf

30 Aug, 2018 5 commits
- fix a missing chief_only flag · fd19f4e2
  Yuxin Wu authored Aug 30, 2018
  
  fd19f4e2
- fix horovod trainer broadcast stage again · f257f0e0
  Yuxin Wu authored Aug 30, 2018
  
  f257f0e0
- update docs · 8fec1bfb
  Yuxin Wu authored Aug 29, 2018
  
  8fec1bfb
- fix the broadcast stage of horovod trainer. · 8f8f9179
  Yuxin Wu authored Aug 29, 2018
  
  8f8f9179
- update docs and new model · 822997c7
  Yuxin Wu authored Aug 29, 2018
  
  822997c7
27 Aug, 2018 5 commits
- [MaskRCNN] parallel evaluation in horovod mode · 65bbd28a
  Yuxin Wu authored Aug 27, 2018
  
  65bbd28a
- Add missing file from last commit · 79711478
  Yuxin Wu authored Aug 27, 2018
  
  79711478
- Use msgpack instead of pyarrow for "serialization to disk". · 805c44d5
  Yuxin Wu authored Aug 27, 2018
  
  805c44d5
- update docs · 708e07b0
  Yuxin Wu authored Aug 26, 2018
  
  708e07b0
- [MaskRCNN] add Cascade-RCNN · 443cb84d
  Yuxin Wu authored Aug 26, 2018
  
  443cb84d
26 Aug, 2018 3 commits
- fix bug in last commit · f7dfb86a
  Yuxin Wu authored Aug 26, 2018
  
  f7dfb86a
- [MaskRCNN] handle empty forground in frcnn head. · 438aef79
  Yuxin Wu authored Aug 25, 2018
  
  438aef79
- [MaskRCNN] BoxProposals struct to manage proposals; rename probs->scores · 6041a1a4
  Yuxin Wu authored Aug 25, 2018
  
  6041a1a4
25 Aug, 2018 1 commit
- [MaskRCNN] move fastrcnn outputs out of head. support class-agnostic regression · cf97218c
  Yuxin Wu authored Aug 25, 2018
  
  cf97218c
24 Aug, 2018 2 commits
- add flownet2 inference examples (#853) · 0d36de5f
  Patrick Wieschollek authored Aug 24, 2018
```
* add flownet2-S inference example

* fix readme

* flake8 fix

* fix flownet-s, flownet-c scaling

* uuu

* Small simplifications

* Add centercrop in inference. Update readme
```
  0d36de5f
- Deprecate "BilinearUpSample" and move it to examples. · 01245d68
  Yuxin Wu authored Aug 24, 2018
  
  01245d68
23 Aug, 2018 3 commits
- [MaskRCNN] train-time scale augmentation · 843d44e9
  Yuxin Wu authored Aug 21, 2018
  
  843d44e9
- [MaskRCNN] score was improved by 8f4a27f0 . rerun evaluation. · 2efc98f7
  Yuxin Wu authored Aug 23, 2018
  
  2efc98f7
- [MaskRCNN] use FastRCNNHead to manage head inputs&outputs · 99ba595d
  Yuxin Wu authored Aug 22, 2018
  
  99ba595d
22 Aug, 2018 6 commits
- update docs · 02eb02e2
  Yuxin Wu authored Aug 22, 2018
  
  02eb02e2
- Update issue templates · 5df194d9
  Yuxin Wu authored Aug 22, 2018
  
  5df194d9
- Update issue templates · 0f82735f
  Yuxin Wu authored Aug 22, 2018
  
  0f82735f
- [MaskRCNN] use dict as input · 49e04ffa
  Yuxin Wu authored Aug 22, 2018
  
  49e04ffa
- Deprecate warning for old modeldesc interface. · 787be08e
  Yuxin Wu authored Aug 22, 2018
  
  787be08e
- Make FeedInput & QueueInput support dict-based dataflow (#768) · 3700a803
  Yuxin Wu authored Aug 21, 2018
  
  3700a803
21 Aug, 2018 1 commit
- [DQN] fix the layout when channel>1 · 209da29e
  Yuxin Wu authored Aug 21, 2018
  
  209da29e