Commit Graph

  • 6379a5a9b3 use segmentation-data-generator in example patch-1 wassname 2017-12-20 09:31:57 +08:00
  • b2b895015a Merge branch 'segmentation-data-generator' of https://github.com/farizrahman4u/keras-contrib into patch-1 wassname 2017-12-20 07:07:56 +08:00
  • 82645703b9 Merge branch 'master' into segmentation-data-generator Andrew Hundt 2017-12-19 17:21:00 -05:00
  • 2a87279b84 test_image_segmentation.py pep8 Andrew Hundt 2017-12-19 16:59:59 -05:00
  • 5d09a9368d image_segmentation.py change backend import Andrew Hundt 2017-12-19 16:09:19 -05:00
  • d54f71a7d2 image_segmentation.py pep8 Andrew Hundt 2017-12-19 15:53:48 -05:00
  • 3508fdd714 Merge branch 'segmentation-data-generator' of github.com:farizrahman4u/keras-contrib into segmentation-data-generator Andrew Hundt 2017-12-19 13:09:31 -05:00
  • 3db35ef130 preprocessing __init__.py updated Andrew Hundt 2017-12-19 13:09:22 -05:00
  • 0957603edd test_image_segmentation.py correct import Andrew Hundt 2017-12-19 12:04:23 -05:00
  • 7bc0d8288f test_image_segmentation.py init SegDataGenerator() Andrew Hundt 2017-12-19 11:58:34 -05:00
  • 09f967889c add code source reference Andrew Hundt 2017-12-19 10:23:46 -05:00
  • 5b51b801f0 Merge branch 'segmentation-data-generator' of https://github.com/farizrahman4u/keras-contrib into patch-1 wassname 2017-12-19 15:37:39 +08:00
  • 04173fa8ee SegDataGenerator keras 2.1.x update Andrew Hundt 2017-12-18 22:28:57 -05:00
  • d0a07d9587 Merge branch 'master' of https://github.com/farizrahman4u/keras-contrib into patch-1 wassname 2017-12-19 09:24:50 +08:00
  • cbe961d798 Merge branch 'segmentation-data-generator' of https://github.com/farizrahman4u/keras-contrib into patch-1 wassname 2017-12-19 09:13:34 +08:00
  • 5f6ddf65a1 image_segmentation.py integer division Andrew Hundt 2017-12-18 18:44:20 -05:00
  • b9cbbd61e5 image_segmentation.py remove cv2 Andrew Hundt 2017-12-18 18:26:35 -05:00
  • 647beddedc SegDataGenerator added from https://github.com/aurora95/Keras-FCN Andrew Hundt 2017-12-18 18:24:49 -05:00
  • fa2feb8fc7 Merge pull request #188 from farizrahman4u/travis-scipy-sparse Andrew Hundt 2017-12-07 12:35:15 -05:00
  • a4134f27e5 backend_test.py Andrew Hundt 2017-12-06 19:55:05 -05:00
  • a4cd7fa245 Merge pull request #187 from titu1994/nasnet_weights Andrew Hundt 2017-12-06 16:21:33 -05:00
  • e3a4b9288a Merge pull request #181 from valentindey/master Andrew Hundt 2017-12-06 16:14:28 -05:00
  • 7e69bbab44 Small change to docs Somshubra Majumdar 2017-12-06 15:02:33 -06:00
  • 9433647d08 Update the docs Somshubra Majumdar 2017-12-06 15:01:36 -06:00
  • 5a11356e89 Add reference to where the weights were obtained from Somshubra Majumdar 2017-12-06 14:38:39 -06:00
  • 416bbe026c add lines for pep8 Valentin 2017-12-06 21:36:06 +01:00
  • aeb39bd426 Add full weight support (NASNet Large and NASNet Mobile) Somshubra Majumdar 2017-12-06 14:35:48 -06:00
  • 7c0d465224 Merge remote-tracking branch 'upstream/master' Valentin 2017-12-06 18:39:35 +01:00
  • 0fa1db5e5c Merge pull request #186 from titu1994/nasnet_reg Andrew Hundt 2017-12-05 17:54:22 -05:00
  • b0aa9cc46d Add corrections as required Somshubra Majumdar 2017-12-05 08:31:38 -06:00
  • 23630e9b0a Merge pull request #185 from farizrahman4u/travis Andrew Hundt 2017-12-05 01:17:58 -05:00
  • f842596f72 Correct the application example Somshubra Majumdar 2017-12-04 17:35:10 -06:00
  • 08f9138af7 Correct NASNet implementation for CIFAR mode and correct the cifar example Somshubra Majumdar 2017-12-04 17:15:23 -06:00
  • 6fe004b12e Correctly pass weight decay to all cells Somshubra Majumdar 2017-12-04 13:07:01 -06:00
  • 0fad550fed add regularizers and fix example Somshubra Majumdar 2017-12-04 13:02:21 -06:00
  • 796732c101 Merge remote-tracking branch 'upstream/master' Valentin 2017-12-03 18:30:52 +01:00
  • 1cfbe9f993 .travis.yml remove coveralls Andrew Hundt 2017-12-02 16:41:28 -05:00
  • da7b9b9b2d .travis.yml attempt to fix travis error Andrew Hundt 2017-12-02 14:21:50 -05:00
  • 9f7ab69249 Merge pull request #180 from titu1994/nasnet Andrew Hundt 2017-12-02 13:46:20 -05:00
  • 3ab207d8a5 Fix https://github.com/titu1994/Keras-NASNet/issues/2 Somshubra Majumdar 2017-11-29 08:56:02 -06:00
  • aeebb8c520 Set default weights to None since ImageNet weights arent available yet Somshubra Majumdar 2017-11-23 13:57:21 -06:00
  • 7f549909e4 Merge pull request #174 from infelane/master Andrew Hundt 2017-11-20 20:57:41 -05:00
  • 68acf5e2ee added absolute tolerance to handle comparing zeros lameeus 2017-11-20 11:46:24 +01:00
  • 17ac57d02a Add example script Somshubra Majumdar 2017-11-18 13:52:05 -06:00
  • 09ec44fa80 Improve auxilary head Somshubra Majumdar 2017-11-18 13:38:12 -06:00
  • b48ed724e7 Merge pull request #158 from vfdev-5/densenet-naming Andrew Hundt 2017-11-18 13:44:22 -05:00
  • 04c1acd770 Merge pull request #182 from farizrahman4u/mkl Andrew Hundt 2017-11-18 13:38:34 -05:00
  • 8f0d77448a .travis.yml set MKL_THREADING_LAYER="GNU" Andrew Hundt 2017-11-18 13:17:49 -05:00
  • 052b05ad75 .travis.yml second mkl fix attempt Andrew Hundt 2017-11-18 12:04:21 -05:00
  • 1d5fed1f08 .travis.yml try fixing mkl errors Andrew Hundt 2017-11-18 11:34:13 -05:00
  • 2f95542184 adds tests for swish activation Valentin 2017-11-18 16:54:23 +01:00
  • 942d5c30c1 adds swish activation function Valentin 2017-11-18 16:54:06 +01:00
  • 4a99de40a2 Improve name scope and add CIFAR model Somshubra Majumdar 2017-11-17 15:41:30 -06:00
  • 0d9934b0bf Improve name scope and add CIFAR model Somshubra Majumdar 2017-11-17 13:06:30 -06:00
  • a32ae325a8 Improve name scope usage Somshubra Majumdar 2017-11-17 12:47:06 -06:00
  • 03e649dbca Add name scopes and correct a flaw with assertion Somshubra Majumdar 2017-11-17 12:32:22 -06:00
  • 2cb41dbc03 Add NASNet models Somshubra Majumdar 2017-11-16 22:27:47 -06:00
  • 3f6814ca3c Update GUIDELINES.md Fariz Rahman 2017-11-09 03:27:24 +05:30
  • 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? lameeus 2017-11-07 11:10:39 +01:00
  • 58d1817d4c PEP8 lameeus 2017-11-06 19:29:00 +01:00
  • 9af7625e0f fixed python 2 bug with divison StringIO depends now on python 2/3 as it should PEP8 fixes lameeus 2017-11-06 19:26:50 +01:00
  • f975a52ea9 Second try. For some reason it doesn't print with all weights 0 at server side. lameeus 2017-11-06 18:47:22 +01:00
  • b2520b935d forgot to remove comment lameeus 2017-11-06 18:23:43 +01:00
  • 46f7a46049 Made compatible with python 2.7 and tried to fix weird bug I didn't encounter at own system lameeus 2017-11-06 18:23:13 +01:00
  • 148a4f9de6 Making test compatible with python 2 and hoping an unexpected bug is fixed (didn't get it at my system) lameeus 2017-11-06 18:13:46 +01:00
  • 9238838aae As discussed. The 'warning' is now just a print. lameeus 2017-11-06 17:52:47 +01:00
  • e873325d0b 1) Extended layers to lameeus 2017-11-06 17:50:59 +01:00
  • b4138a7144 Merge branch 'patch-1' of github.com:wassname/keras-contrib into patch-1 wassname 2017-11-04 22:29:10 +08:00
  • df805e32ef update example wassname 2017-11-04 22:28:04 +08:00
  • ed613c8965 Update pascal_voc_jaccard_loss.py Mike Clark 2017-11-04 22:18:19 +08:00
  • 3ac7a732bd ref to paper wassname 2017-11-04 16:41:19 +08:00
  • 0d43a01f54 add example wassname 2017-11-04 16:41:11 +08:00
  • bbb08f42cb Test and PEP8 fixes lameeus 2017-10-27 11:45:43 +02:00
  • 3f206044c8 changed warning method to be adjustable lameeus 2017-10-27 11:25:07 +02:00
  • 5c690d9b0b Merge pull request #176 from farizrahman4u/ahundt-patch-1 Andrew Hundt 2017-10-26 14:18:17 -04:00
  • 984becbe90 test_normalization.py reduce test flakiness Andrew Hundt 2017-10-26 13:43:25 -04:00
  • d299d4c3ad Merge pull request #163 from ahundt/travis Andrew Hundt 2017-10-26 13:06:06 -04:00
  • 601928f6c2 Merge pull request #175 from charlesreid1/master Andrew Hundt 2017-10-26 13:03:17 -04:00
  • 69fdd1d644 Merge pull request #172 from ahundt/normalization_flakiness Andrew Hundt 2017-10-25 09:12:06 -04:00
  • ececa1b7af Made indepent of backend. Now only checks channel priority (as suggested by @arodiss) lameeus 2017-10-25 13:45:52 +02:00
  • 3d710d4984 Simplification (suggested by @arodiss) lameeus 2017-10-25 13:32:01 +02:00
  • d635747da3 Made first test more general + test if bias included + test when using a convolutional layer (Conv2D) lameeus 2017-10-25 12:14:39 +02:00
  • 64ee9acfdb 1) Extended layers to everything that has a ReLU activation 2) Added layer name to the warning 3) Corrected the dead neuron count (should work for both Theano and TF (tested for TF)) 4) Added a print alongside the warning since for me the warning stopped showing after some epochs (might be a bug at my side). lameeus 2017-10-25 12:13:09 +02:00
  • 8493d1f8e9 Merge commit '43c21d9d8d8f3003f0fc075b0c1f51453af8b47a' into normalization_flakiness Andrew Hundt 2017-10-23 19:49:00 -04:00
  • 43c21d9d8d Merge pull request #171 from seeb0h/master Somshubra Majumdar 2017-10-20 09:31:59 -05:00
  • 5af0db7178 Fix bad argument name for require_flatten Sébastieb Bosch 2017-10-20 15:59:03 +02:00
  • 627e45eba5 test_normalization.py reduce test flakiness Andrew Hundt 2017-10-18 11:32:41 -04:00
  • fd73547011 Merge pull request #169 from ahundt/remove_deconv3d Andrew Hundt 2017-10-18 02:10:47 -04:00
  • ad120f95e9 tensorflow_backend.py import dtype Andrew Hundt 2017-10-18 00:49:51 -04:00
  • 80d698c159 tensorflow_backend.py temporarily move _preprocess* functions into keras-contrib Andrew Hundt 2017-10-18 00:34:23 -04:00
  • 0d226a6775 remove more deconv3d code Andrew Hundt 2017-10-17 16:51:53 -04:00
  • a8ec7ad10a remove Deconvoluion3D because it is available in keras upstream https://github.com/fchollet/keras/pull/7161 Andrew Hundt 2017-10-17 16:45:55 -04:00
  • b0a2fbabfd resnet.py initial integration (#148) Andrew Hundt 2017-10-15 21:21:50 -04:00
  • 001ba42938 Merge commit '7dc329a5a3ec678d4c589529726ea5c7a6c458c8' into travis Andrew Hundt 2017-10-15 15:19:16 -04:00
  • 7dc329a5a3 Merge pull request #162 from farizrahman4u/save_load_utils_tf_only Andrew Hundt 2017-10-15 14:23:42 -04:00
  • 70e68c11f5 save_load_utils_test.py make tensorflow only Andrew Hundt 2017-10-15 13:28:43 -04:00
  • 99fe5e3b3a Merge pull request #142 from stygstra/batchrenorm-bugfix Andrew Hundt 2017-10-15 13:13:29 -04:00
  • 38937fdf7c Merge pull request #155 from farizrahman4u/setup Andrew Hundt 2017-10-15 13:04:53 -04:00
  • de768b6048 Merge pull request #161 from farizrahman4u/rm_test_file Andrew Hundt 2017-10-13 09:57:28 -04:00
  • 7577b8780a Fix BatchRenormalization clipping (fixes #127) David Stygstra 2017-08-29 22:44:05 -07:00