Commits · d48cce3aa399db11a4e6fc35a5a896c96a92b641 · Shashank Suhas / seminar-breakout

29 Mar, 2019 1 commit
- Delay the error about duplicated layer registry until it's used (#1125). · d48cce3a
  Yuxin Wu authored Mar 29, 2019
  
  d48cce3a
28 Mar, 2019 2 commits
- still use `tf.get_variable` because `tf.Variable` sometimes takes name_scope... · 5a209dbd
  Yuxin Wu authored Mar 27, 2019
```
still use `tf.get_variable` because `tf.Variable` sometimes takes name_scope rather than variable_scope
```
  5a209dbd
- update docs · a871c622
  Yuxin Wu authored Mar 27, 2019
  
  a871c622
25 Mar, 2019 2 commits
- bump version · f947192a
  Yuxin Wu authored Mar 25, 2019
  
  f947192a
- Translate placeholder's tensor name to op name when contructing TensorSpec (fix #1118) · 61a1d3f2
  Yuxin Wu authored Mar 25, 2019
  
  61a1d3f2
23 Mar, 2019 1 commit
- improve find_library · 2b4ec72e
  Yuxin Wu authored Mar 23, 2019
  
  2b4ec72e
22 Mar, 2019 1 commit
- collect_env_info · 25bf4336
  Yuxin Wu authored Mar 22, 2019
  
  25bf4336
21 Mar, 2019 2 commits
- fix bug in 3d1a30ff · c1f8042d
  Yuxin Wu authored Mar 21, 2019
  
  c1f8042d
- reorganize some conv2d code; bump version · 4888d1ea
  Yuxin Wu authored Mar 20, 2019
  
  4888d1ea
19 Mar, 2019 3 commits
- Move more core code to tf.compat.v1 · 3d1a30ff
  Yuxin Wu authored Mar 19, 2019
  
  3d1a30ff
- Remove the old ModelDesc interface (which was deprecated 1 year ago). · 4057c531
  Yuxin Wu authored Mar 19, 2019
  
  4057c531
- InputDesc -> tf.TensorSpec everywhere · b8a50d72
  Yuxin Wu authored Mar 19, 2019
  
  b8a50d72
18 Mar, 2019 3 commits
- placeholder->TensorSpec refactor · ba679ab1
  Yuxin Wu authored Mar 18, 2019
  
  ba679ab1
- [MaskRCNN] Allow EVAL_PERIOD=0 to disable evaluation (#1112) · 0aa92de5
  Yuxin Wu authored Mar 17, 2019
  
  0aa92de5
- update docs · 167fa123
  Yuxin Wu authored Mar 17, 2019
  
  167fa123
17 Mar, 2019 2 commits
- Workaround dilated conv bugs in tf.layers.Conv2D (#1110). · 353cd04f
  Yuxin Wu authored Mar 17, 2019
```
Another bug in tf.layers .. maybe I should never switch to it.
```
  353cd04f
- Backport TensorSpec; tf=tf.compat.v1 in many files. · 505e28eb
  Yuxin Wu authored Mar 17, 2019
  
  505e28eb
16 Mar, 2019 1 commit
- [MaskRCNN] visualization: add background to text · e4941595
  Yuxin Wu authored Mar 15, 2019
  
  e4941595
13 Mar, 2019 6 commits
- fix some empty op_def issues in old TF · e0cd3cdb
  Yuxin Wu authored Mar 13, 2019
  
  e0cd3cdb
- fix some weird python2 scoping issue · 79c93370
  Yuxin Wu authored Mar 13, 2019
  
  79c93370
- fix last commit · b9ba5bc2
  Yuxin Wu authored Mar 13, 2019
  
  b9ba5bc2
- warn about possibly-blocking initializer · 8f623281
  Yuxin Wu authored Mar 13, 2019
  
  8f623281
- fix bug in Conv2DTranspose For TF1.13 (introduced in 152c4c2c ) · 3862b72c
  Yuxin Wu authored Mar 13, 2019
  
  3862b72c
- [MaskRCNN] improve visualization palette · 55baf374
  Yuxin Wu authored Mar 13, 2019
  
  55baf374
11 Mar, 2019 1 commit
- update docs · f42036ac
  Yuxin Wu authored Mar 11, 2019
  
  f42036ac
06 Mar, 2019 1 commit
- [DQN] make DQN more generic: remove some constants & globals · f417c49f
  Yuxin Wu authored Mar 06, 2019
  
  f417c49f
05 Mar, 2019 2 commits
- Make DQN support states with more dimensions · 0b561b3b
  Yuxin Wu authored Mar 05, 2019
  
  0b561b3b
- Fix bug in resnet; improve logs for #1100 · 87fad54b
  Yuxin Wu authored Mar 04, 2019
  
  87fad54b
03 Mar, 2019 2 commits
- add missing file · ed1030b7
  Yuxin Wu authored Mar 03, 2019
  
  ed1030b7
- Augmentor uses register_at_fork; update docs (fix #1099) · 2878aceb
  Yuxin Wu authored Mar 03, 2019
  
  2878aceb
01 Mar, 2019 1 commit
- update docs; add BN WD in resnet · 964f5d03
  Yuxin Wu authored Mar 01, 2019
  
  964f5d03
25 Feb, 2019 3 commits
- fix FasterRCNN horovod evaluation · 274c7544
  Yuxin Wu authored Feb 25, 2019
  
  274c7544
- update docs & deprecation · dbc50068
  Yuxin Wu authored Feb 24, 2019
  
  dbc50068
- [MaskRCNN] improvements on dataset loading · 69d4e940
  Yuxin Wu authored Feb 24, 2019
  
  69d4e940
22 Feb, 2019 2 commits
- fix FasterRCNN evaluation · 7a19c73f
  Yuxin Wu authored Feb 22, 2019
  
  7a19c73f
- update docs · eb97cf1e
  Yuxin Wu authored Feb 21, 2019
  
  eb97cf1e
21 Feb, 2019 4 commits
- Do not require ALE if using DQN with gym only (fix #1091) · 7b9d9b8c
  Yuxin Wu authored Feb 21, 2019
  
  7b9d9b8c
- Try using cudnn's group conv · e4aca035
  Yuxin Wu authored Feb 20, 2019
  
  e4aca035
- fix typo (fix #1088) · d8d35fb5
  Yuxin Wu authored Feb 20, 2019
  
  d8d35fb5
- Reimplement Conv2DTranspose to avoid Keras bugs. · 152c4c2c
  Yuxin Wu authored Feb 20, 2019
  
  152c4c2c