wassname/keras-contrib
1
0
Fork 0
mirror of https://github.com/wassname/keras-contrib.git synced 2026-08-06 13:20:08 +08:00
Code Issues Packages Projects Releases Wiki Activity
370 Commits 12 Branches 0 Tags
patch-1
Commit Graph
102 Commits
Author SHA1 Message Date
wassname b2b895015a Merge branch 'segmentation-data-generator' of https://github.com/farizrahman4u/keras-contrib into patch-1 2017-12-20 07:07:56 +08:00
Andrew Hundt 82645703b9 Merge branch 'master' into segmentation-data-generator 2017-12-19 17:21:00 -05:00
Andrew Hundt 2a87279b84 test_image_segmentation.py pep8 2017-12-19 16:59:59 -05:00
Andrew HundtandGitHub 0957603edd test_image_segmentation.py correct import 2017-12-19 12:04:23 -05:00
Andrew HundtandGitHub 7bc0d8288f test_image_segmentation.py init SegDataGenerator() 2017-12-19 11:58:34 -05:00
wassname d0a07d9587 Merge branch 'master' of https://github.com/farizrahman4u/keras-contrib into patch-1 2017-12-19 09:24:50 +08:00
wassname cbe961d798 Merge branch 'segmentation-data-generator' of https://github.com/farizrahman4u/keras-contrib into patch-1 2017-12-19 09:13:34 +08:00
Andrew Hundt 647beddedc SegDataGenerator added from https://github.com/aurora95/Keras-FCN 2017-12-18 18:24:49 -05:00
Andrew HundtandGitHub a4134f27e5 backend_test.py 
remove scipy.sparse as it doesn't appear to be utilized in this file
 2017-12-06 19:55:05 -05:00
Valentin 416bbe026c add lines for pep8 2017-12-06 21:36:06 +01:00
Valentin 796732c101 Merge remote-tracking branch 'upstream/master' 2017-12-03 18:30:52 +01:00
Andrew HundtandGitHub 7f549909e4 Merge pull request #174 from infelane/master 
Fixed and extended Dead ReLU detector callback
 2017-11-20 20:57:41 -05:00
lameeus 68acf5e2ee added absolute tolerance to handle comparing zeros 2017-11-20 11:46:24 +01:00
Valentin 2f95542184 adds tests for swish activation 2017-11-18 16:54:23 +01:00
lameeus a1079c012a I still can't reproduce the bug I get at server side. I commented those lines of code. If someone else could find out what is going wrong? 2017-11-07 11:10:39 +01:00
lameeus 58d1817d4c PEP8 2017-11-06 19:29:00 +01:00
lameeus 9af7625e0f fixed python 2 bug with divison 
StringIO depends now on python 2/3 as it should
PEP8 fixes
 2017-11-06 19:26:50 +01:00
lameeus f975a52ea9 Second try. For some reason it doesn't print with all weights 0 at server side. 2017-11-06 18:47:22 +01:00
lameeus b2520b935d forgot to remove comment 2017-11-06 18:23:43 +01:00
lameeus 46f7a46049 Made compatible with python 2.7 and tried to fix weird bug I didn't encounter at own system 2017-11-06 18:23:13 +01:00
lameeus 148a4f9de6 Making test compatible with python 2 and hoping an unexpected bug is fixed (didn't get it at my system) 2017-11-06 18:13:46 +01:00
lameeus 9238838aae As discussed. The 'warning' is now just a print. 
Test is adjusted to handle printing instead of throwing a warning.
 2017-11-06 17:52:47 +01:00
lameeus e873325d0b 1) Extended layers to 2017-11-06 17:50:59 +01:00
lameeus bbb08f42cb Test and PEP8 fixes 2017-10-27 11:45:43 +02:00
Andrew HundtandGitHub 984becbe90 test_normalization.py reduce test flakiness 2017-10-26 13:43:25 -04:00
lameeus ececa1b7af Made indepent of backend. Now only checks channel priority 
(as suggested by @arodiss)
 2017-10-25 13:45:52 +02:00
lameeus d635747da3 Made first test more general 
+ test if bias included
+ test when using a convolutional layer (Conv2D)
 2017-10-25 12:14:39 +02:00
Andrew HundtandGitHub 627e45eba5 test_normalization.py reduce test flakiness 2017-10-18 11:32:41 -04:00
Andrew Hundt a8ec7ad10a remove Deconvoluion3D because it is available in keras upstream 
https://github.com/fchollet/keras/pull/7161
 2017-10-17 16:45:55 -04:00
Andrew HundtandGitHub 70e68c11f5 save_load_utils_test.py make tensorflow only 
Making this test tensorflow only until https://github.com/farizrahman4u/keras-contrib/issues/157 is resolved.
 2017-10-15 13:28:43 -04:00
Andrew HundtandGitHub 99fe5e3b3a Merge pull request #142 from stygstra/batchrenorm-bugfix 
Fix BatchRenormalization clipping (fixes #127)
 2017-10-15 13:13:29 -04:00
David Stygstra 7577b8780a Fix BatchRenormalization clipping (fixes #127) 
Previouly, a constant value was used for the clipping parameters. This
meant that it stayed at r=1 and d=0 forever, making it essentially
equivalent to regular batch normalization.
 2017-10-11 18:45:35 -07:00
Andrew HundtandGitHub 6ec66fec74 save_load_utils_test.py removes 'model.h5' 2017-10-10 13:17:06 -04:00
wassname 5b4df46547 pep8 2017-09-24 11:36:56 +08:00
Mike ClarkandGitHub 3a3a250a8f Create jaccard_test.py 2017-09-24 09:50:32 +08:00
Somshubra Majumdar 821eff3948 fix normalization tests 2017-09-22 13:14:15 -05:00
Michael Oliver 15133077af Fix test to refer to correct field for momentum weights in optimizer 2017-07-03 13:28:18 -07:00
Gleb SidoraandMichael Oliver 13ec4b4174 Added Dead ReLU detector callback (#115) 2017-06-30 15:08:39 -10:00
Mathieu Marquis BolducandMichael Oliver f0bb5becbb Instance Normalization layer (#101) 
* Instance Normalization layer

* fix Instance Normalization for theano broadcasting rules

* Support rank 1 tensor in Instance Normalization

* address issue with axis parameter in instance normalization and add unit test for per channel correctness

* avoid assert_raises in normalization unit test

* Instance normalization set axis default to None
 2017-06-19 08:57:54 -07:00
Michael Oliver 6a8739b6c7 add ftml_test.py 2017-06-16 13:27:47 -07:00
BenandMichael Oliver ef2daecc58 Move activations, metrics, optimizers and regularizers (#94) 2017-06-15 16:15:59 -07:00
Michael OliverandGitHub 0f78409087 fix sub_pixel_upscaling test 2017-06-09 09:08:17 -07:00
BenandMichael Oliver 4f619fa686 save_all_weights and load_all_weights (#90) 
* save_all_weights and load_all_weights

* doc updates, fix test, remove .gitkeep
 2017-06-07 20:59:48 -07:00
BenandMichael Oliver f0d7fe852d Make callbacks into a directory. (#91) 2017-06-07 20:48:07 -07:00
BenandMichael Oliver ec8010077c Move DSSIMObjective to submodule (#93) 
* Move DSSIMObjective to submodule

* fix losses/dssim_test.py
 2017-06-07 20:47:39 -07:00
Frédéric Branchaud-CharronandMichael Oliver 3ebf304737 fix dssim, add test (#86) 
* fix dssim, add test

* add more tests
 2017-05-20 15:51:31 -07:00
Eric Xihui LinandMichael Oliver d1f4b6ba75 CRF for keras 2.x (#76) 
* implemented CRF

* added test for CRF

* added a chunking example for CRF

* changed to use up2date tensorflow

* added  conll2000 data

* minimize package dependency

* using logsumexp in keras instead
 2017-05-18 10:28:17 -07:00
Michael OliverandGitHub 692e8a5aef Add SReLU activation function (#77) 
* add srelu activation

* update srelu style

* make srelu and pelu work with any initializer

* add missing config line

* make initializations better match paper
 2017-04-24 13:26:27 -07:00
Somshubra MajumdarandMichael Oliver 8ef69698b8 Update BatchRenorm to Keras 2 API (#69) 
* Update BatchRenorm to Keras 2 API

* update tests

* Remove mode 1 test (mode 1 doesnt exist anymore)
 2017-04-15 09:43:55 -07:00
nzwandMichael Oliver 88c60d2459 Style fixes (#71) 2017-04-15 09:41:28 -07:00